^^ O^.. ^^^> ^\..v BT HUGH L0NGB0U3NE OALLENDAB, M.A., FJB.S., AMD JOHN THOMAS NICOLSON, B.So. WITH AN ABSTRACT OF THE DISCUSSION UPON THE PAPER. EDITED BT J. H. T. TUDSBEBY, D.Sc., M. Inst. C.B., BKOaSTAMT. By pennission of the Oouncil. BxoerptMinutesofProoeedingsofThelnstitutionofCivilEnginoers. VoL oxMXL Session 1897-98. Part i. L(>NDON: 9nblt>[)(li is ^t Snstittttion, OEEAT GEORGE STREET, WESTMINSTER, S.W. [TEMaBAMS, "iHSTITOIIOir, LOKDOV." TlLBPBOn, "WwinilllBTIil, 81."] 1898. ITht right qfPiMie»lir E, the electro- ition, one of the 0., was found >T0ltS. plied chiefly to bure at a known the simpler for- jree, namely : — it observation of orresponding to as deduced from e dotted curve, iriations of tem- ProooodiiiKi.] lAW or ooHoraniATiOM or stiAic. peratnre observed at a depth of O'Ol inch in the cover, at a speed of 100 revolutions per minute. The observations are shown at {x). The cycle shows a range of only 4 • 3T., corresponding to an absorp- MOL Fig. 5. . u Mtf ^K . Hnru 1 / '» •^ > / ^Ti. *»*w /. ' \ "\ ^ -\ ■"WlWiW r^ X r'^ g - — 'V ^ p^ --,' ' z s 1..^ '*««> a / v?^ fi 1 I / ^1!. IE / X >J \ tflrf 4 « — i TIMI W 1 IN 1 M < IXTIB 1 >|BRI 1 vou. • to tS 10 M «0 4S SO ITION fnOA BACK SNO O*' STROKE Speed, 100 rcvolntloni par minute.'' Mktal Cycle at ^ Inch Oovkb. Fig. 6. 40*' m mm^mmmmmmmm nm« 8 OALLENOAB AND NIOOLSON ON THE [Hlnutfli of In taking the observations for the wall-cycle, shown in Fig. 6, the differential method was adopted for the first time, which aooonnts for the greater smoothness of the carve. In this method the thermo-electric circuit consisted simply of a small portion of the cast iron of the cylinder and of two iron wires making contact at different depths in the metal, at points which are at nearly the same mean temperature Only a small difference of temperature, for whioh the thermo-electric method is admirably suited, is therefore observed. By this method, nearly all the troubles that otherwise arise from the slow changes 400 Fig. 7. n - — lid § i mwk'^ VK) Wm\ W('f\ 1 k MIMN TfMR Of WAUJI IN OOVKII / 1 \" \.* / 1 \ \ 1 f 1 \j t^\ ^ t VIq^. tKi 7 ' — ' *v ;—T -TWM «^ / y If \ ^L. nAJL-. \— '^*-.^ ^— — - SSJ^f tod ( * p-^ - , 1 » «S BO U O k 10 IS to ta 30 3S 40 4« SO TIMB IN SIXTIETHS OP A REVOLUTION niOM BACH CNO OF STROKE Speed, 7.1-4 revolutloni per minute. Hktal Gtolk at ^ iMon Cotxb. of conditions always in progress are avoided, since a gradual change of temperature will affect the metal at a depth of ^ inch to the same extent as the surface of the metal ; whereas the rapid cyclical changes are practically evanescent at the greater depth. For the remainder of the experiments the differential method was always used for the wall-cycles. The cycle shown in Fig. 7 is an illustration of an observation taken about a month later, at a depth of 0-039 inch, and a speed of 73-4 revolutions per minute. The shaded area of the steam curve, above the line representing the mean temperature of the cover, may be called the " Con- densation Area," as it appears to determine the amount of con- densation taking place. For the cycles shown in Figs. 5 and 6, the test-wires were pressed against the metal of the cover. For <• [Mlnutfli of >wn in J^'j;. 6, time, which ited simply of d of two iron »tal, at points Only a small eotric method lethod, nearly bIqw changes ID Z 4.0 h U z 3 ■> t 6 a ■c III K> 4« so OF STROKE loe a gradual 3pth of ^ inch areas the rapid greater depth, al method was in Fig. 7 is an iter, at a depth linute. iC representing ed the "Con- mount of con- Figa. 5 and 6, le cover. For (f 4 ':" ■•vt^ IftmMdingt.] LAW OF OOKDENBATION or 8TEAH. V Fig. 7 and subsequent observations, the apparatus had been entirely rearranged. The holes had been bored and moaisure ^ 3 (A 10 2 \ ^ \ III K 0. i \ \ S ■ 1 -j I .05 -15 -25 35 -45 SS ■< 36 ■■ rs as I 16 f -f unnuin u* s r h u n e Speed, T3-4 revolution* per minute •, cut-oir O-aoo. Inuioatob Diaobah. junctions were fitted along the side of the cylinder to observe the wall-cycles and the distribution of temperature along the barrel- surface. The arrangement of the side junctions is shown in Figs. 4, Plate 6. To facilitate making and changing connec- tions, the majority of the test-wires were connected to mercury- cups in a slip of wood fixed to the cylinder inside the cast-iron lagging. At the back end of the stroke, and at 4 inches, 6 inches and 12 inches along the side from it, pairs of junctions were inserted, one bored to leave 0-04 inch of metal, and the other il inch at each point. The remainder of the junctions, at 2 inches, 8 inches, 10 inches, 14 inches and 16 inches respectively, were '' ' ''ft. l-^'rriM'j^-» a *»A: «w 4rt *'^*Ti»^Cfi%.U.»«f•ifrtvt*aSh^^?i*^ .Tt^ ^uaa fi fc ii iV i ' [Minutes of m wires could ' temperature. }elow on a line loles, 2 inches it distances of ir the insertion ating the side tures. arves in Fig. 9 an temperatuie 14 -ST. lower iituated at this D. ns zvrs3: i^SL'- * SUicjiya,; IS ID Z o 10 9 ■> d si u 4« OP STROKK Bjdjc «• }sed to the r.ame 9 cover and side, i being 44 revo- )n8 W3re, cover, 1 particular cycle lout the trials. e surface of the ion is shaded as the temperature pression, shortly am shortly after The fall due to Procoedings.] LAW OF 001 SENSATION OF STEAM. II re-evaporation is nearly as rapid as the rise. In drawing the lower boundary of the condensation area for this cycle, the probable surface temperature is taken instead of the mean ; but the difference of area is small. The magnitude of the condensa- tion area would be little affected by taking the mean temperature instead of the surface temperature, but the cycle curve would not then correspond so well, allowing for lag, with the probable rate and epoch of oondensation. i : v^~;> -r!- ' . . r ; Side-Wall Cycles beyond Cut off. — Fig. 10 gives an example, on a temperature scale two and a half times as large, of cycles observed on junctions Nos. 3 and 4, at 4 inches and 6 inches along the cylinder from the back end. The lower observations at the points Fig. 10. 46 so SS O B 10 IS CO CS 40 35 40 45 SO TIME IN SIXTIBTHS OP A REVOLUTION PROM BACk END OP STROKE Speed, 4S-6 rerolutions per minute. Metal Gtclks in Nob. S and 4 Smz. 0, 10, 15 and 20, for No. 3, were taken after the observation of cycle No. 4, and illustrate the order of accuracy attainable in the measurement of these cycles. It will be noticed that the tem- perature begins to rise at each point before steam reaches it, if the piston is assumed to fit accurately. This may be explained partly by the friction of the hot piston, partly by a probable small leakage of steam, and partly by the fact that the piston-rings are about ^ inch from the face of the piston. In estimating thu condensa- tion areas for these junctions, ^ inch is allowed for the imperfect fit of the piston ; that is to say, it is assumed that the steam reaches the junction when the face of the piston is ^ inch behind the correspond- ing point of the stroke. The thickness of metal ut these junctions ! '"4 12 OAIiliENDAB AND MOOLSOM ON THE [Minutes of was • 037 inch and • 039 inch reipectively, and is probably aocarate to Tri^. The speed was 46 • 6 revolutions per minute. Effect of Later Cut-off. — A few trials only were run at a later cut-oflf— three at one- third, and two at one-half. The data of these trials were not in general sufficiently complete to afford a satisfactory basis of comparison. With a later cut-off the range of the wall-cycles observed on the cover did not differ materially from those previously recorded. The longer steam contact was compensated by a higher wall temperature. On the side wall, near the back end of the cylinder, the rise of temperature was nearly 20° F. at one-half cut-off as compared with one-fifth. The range of the cycle was reduced from 11*0° F. to 9-2° F., at a depth of 0*037 inch in ti 'i metal, and a speed of 49 revolutions per minute. At 4 inches and 6 inches along the cylinder the ranges of the wall-cycles, at the same depth and cut-off, were increased to 7*2° F. and 5-0° F. respectively, as the full-pressure steam reached these points of the wall. The tem[)erature of the middle of the cylinder was raised nearly 30° F. as compared with one-fifth cut-off, but the comparison could not be made quite satis- factorily owing to slightly different conditions of steam pressure. The total condensation, including the later portions of the stroke, was probably increased somewhat, but at the same time the oylirder feed was more than doubled. Temperature Distribution and Steady Heat Flow. Outward Temperature Gradients. — Careful measurements of the temperature gradients were made in various parts of the cylinder with a view to deduce the steady flow of heat. From the mean of several observations in the thickness of the cover, which was pro- tected by an air-jacket, a probable gradient of • 66° F. per inch was deduced, a value which corresponds fairly well with the probable external loss. At the points 4 inches and 6 inches along the side, the temperature of the inner surface was 2*4° and 1 • 1° F. respectively lower than that of the outer surface. This curious and at first sight paradoxical result, means that the cylinder wall at these points was losing heat to the steam. Temperature Gradient* on Barrel Surface. — The longitudinal dis- tribution of temperature for one-fifth cut-off, single-acting, is shown in Fig. 11, together with a corresponding diagram of steam temperatures deduced from the average indicator-diagram, with a mean speed of 45 -6 revolutions per minute. On several occasions special experiments showed that a change '•= [Minutes of bly accurate 1 at a later 7he data of to afford a ;he range of materially sontaot waa ) side wall, arature was -fifth. The 2° F., at a revolutions sylinder the lut-off, were ull-pressure iture of the apared with quite satis- m pressure. f the stroke, le time the Flow. lents of the the cylinder the mean of ich was pro- ' r. per inch )11 with the inches along and 1-1° F. rhis curious iylinder wall itudinal dis- le-acting, is ram of steam ^am, with a lat a change ProaedingB.] LAW OF CONDENSATION OP STEAM. 18 of speed produced little change in the distribution of temperature. The mean temperature at the centre of the cover in the same trials is shown by the mark (-), at 306° F,, opposite the section of the cover. Effect of Longitudinal Conduction, and Piston Convection. — The greater part of the condensation on the barrel- admission surface in a small single-acting engine is probably due to the lowering of temperature caused by conduction and convection along the cylinder. In this set of observations the maximum gradient of 9 "3° F. per inch occurs at a point a little after cut-off, and corresponds to a loss of heat by the admission surface of 11*2 Fig. 11. Sbotion of Cvundkb and Oovkb, bhowino Position of Thebuo- JUNCTIONB AND ThERMOHETEBS. thermal units per minute. The cyclical convection of heat by the piston is a factor which may be of some importance in a single- acting engine. It will depend on the surface of the piston, on tho closeness of the contact, and on the difference of temperature between the ends of the cylinder. In the present case the mean difference of temperature was 45° F. at one-fifth cut-off, and nearly 66° F. at one-half cut-off. The curved surface of the piston was approximately 1 square foot. The effect of piston convection may l>e traced in the form of the cycles observed at one of the side junctions. In a double-acting engine, wliore both ends of the tmrm'mif'i'^mffl'HFi 14 OALLENDAB AND NIOOLSON ON THE [Minntes of cylinder are at nearly the same temperature, the effect of piston convection would be of very little importance. In a high-speed single-acting engine like the Willans engine, it is probable that piston convection might be the most important factor in cooling the barrel-admission surface, and causing initial condensation. Abstraction of Heat by the Condensation of Wet Steam. — For a distance of nearly 6 inches along the side, while the engine was running at one-fifth Cut-off, the external surface was nearly 1° F. on the average hotter than the internal. Heat, supplied by conduction, was being abstracted from the inner surface at a rate of at least 5 thermal units per square foot per minute. On stopping the engine the temperature along this belt immediately began to rise, and continued rising for some minutes. The heat abstracted by evaporation is greater than that supplied by condensation over this part of the surface. The probable explanation is to be found in the wetness of the steam due to adiabatio expansion. (See p. 41.) Conductivity and Specific Heat of Cast Ieon, The electrical resistance of cast iron was found to be nearly ten times greater than that of wrought iron. Considerable difference Scale ♦ inch = 1 foot. Apparatus for DEXKBMraiNO the Tbebmal Conductivitt of Cast Iron. in the thermal conductivity was therefore to be expected. Observa- tions on the cylinder showed that the conductivity of cast iron was probably some 30 per cent, less than the value generally assumed. It was therefore desirable to attempt a special determination of this important constant by the most accurate methods. With this object the apparatus shown in Fig. lU was designed. The metal used was a 4-inch bar of iron, cast from the same ladle as the cylinder cover. Two independent methods were employed — (1) The calori- metric method, in which the quuatity of heat transmitted is "wiwiipi<.°-'?NCK«.''*' **^ °*^ MO" siieO Speed, 42 revolutions per minute s range, 20" at surface. Tkmpbbatueb Dkfth Cubves; Cabt-Ieon Cylindbb. the cyclical absorption of heat by the metal of a cast-iron cylinder at a speed of 42 revolutions per minute, for a simple harmonic variation of surface temperature with a range of 20°. The full curves show the simultaneous values of the temperature at ■^) % miimmmml'ff ■9»- ■WPP THE [MinutcR of of the Bar. —The 1. for a plate 1 foot ven in Table I, are 36-6. It will he Per cent. Total carbon 808 Graphite . . 2-86 Silicon . . 289 Phugphoms . 1*05 Manganese . 0-85 Sulphur . . 022 > Total 7-89 1 Fig. 13 illustrate "^^^^ mSa ntkJ p. "'T! ftSEl 0.5O kce. iteo }ast-iron cylinder simple harmonio } of 20°. The ) temperature at Proceedingg.] LAW OF OOMDBNSATIOM OF BTBAM. 17 m e diflferent depths for four typical points of the cycle. The heat absorbed or rejected by the metal between any two points of the cycle is proportional to the area included between the correspond- ing temperature-depth oui* .es. The dotted boundary curves have the equation t= ± e~"", and show the rate of diminution of the range of temperature t with increase of depth a. The index coefficient « is given by the formula m = \/ — p-, where n is the number of revolutions per minute. In the case figured, = 10- 5. The wave-length of the temperature-oscillation is inch. At this depth the retardation amounts to one complete period, and the range is reduced to less than one five-hundredth part of its value at the surface. The wave-length in each case may thus conveniently be regarded as the practical limit of penetration of the heat-waves. Numerical Values for Coat Iron. — The following Table has been calculated for a oast-iron cylinder, with a simple-harmonic cycle of 10° F. range at the surface, at various speeds, assuming the z* values i = 5*4, c =4*5, - = 1*20, in the units given. The values c for any other range are directly proportional to the range. For cal- culating the various columns the numerical formulas used are : — _ 6-28 Index coefficient m = 1'618 ^n. Wave4ength = ~^' Temperature range at 0-040 inch depth = 10° x e" '"*''". Heat absorbed in thermal units Fahr. per square foot per cycle 3-18 ■ ■ '. J.., .-,•■ -.....•.•.- = 10° X . ■ -: •• m Table II. — Gyolioal Hkat Absobftion at Difrbrbnt Spxrds ik Oast Ibon. BeTolatluns Index Wave- LcDgth (Penetntlon). Temperatnre Ktnge tx 0-04U incb. Hett Absorbed in Tbemul Units Minute. Coefflcleni. per .Square Foot. ti. m. Inch. i'er 10" Surlkoe. Per Ilevolntion. Per Minute. 25 8-09 0-777 7-25 8-93 98 40 10-24 0-618 6-63 8-11 124 50 11-44 0-549 6-32 2-78 139 60 12-54 0-601 6-05 2-53 152 70 18-54 0-464 5-82 2-85 164 80 14-47 0-484 6-60 2-20 176 90 15-85 0-409 6-41 207 187 100 16-18 0-388 5-24 1-97 197 150 19-82 0-317 4-52 1-61 242 200 22-88 0-275 4 00 1-40 280 SOO 28-02 224 3-26 114 342 400 32-36 194 2-75 0-98 392 500 86-18 0-174 2-85 0-88 440 [THK INST. C.K. VOL. CXXXI.] •mf wmm ■<■¥< 18 OALIiBNDAR AND NIOOLSON ON THE [MinntM of The aboTo Table is used in oalcnlatiDg the results of the wall-oyoles. Effect of the Form of the Cycle. — The formnlas given, refer to a simple-harmonio wave, which is propagated without change of I' Id |J21 g \'1 :s \ m» r 4 *T i '.T is yf ' .-V^'T*- ''2 W 1 f y %^*i 1 1 ^ V / v> i A 1 f/ Wr MEAI T«M •. 9r '*"■ o* *-l ■■ w*i ._y^-< ■i^j "-"i _.-. >* — . _ — - ■•~M \-2 -■••■ -f — •••■ •—■>. ..... :? — — -6 4« so S5 O S 10 IS 20 TIME SCALE IN SIXTIETHS OF CYCLE 25 30 35 «0 45 SO Triangula- Cycle roii a ranqi or 4 5° at ^ inch nc MnAL. form. Cycles of any other form must be analyzed by the Fourier method into their simple-harmonic components. In order to com- pare the results of assuming an entirely arbitrary and peculiar 30 3S 40 45 so Simplk-Habmonio Cyolb fob a banqb of 4-5° at ^ inch. "VS 50 55 O 5 10 IS 20 2S TIME IN SIXTIETHS OF A CYCLE cycle, instead of the simple-harmonic, the coefficients of the first twelve terms of the series, representing the triangular cycle shown in Fig. 14, were calculated by the Fourier method. This series, when plotted, gives a curve. Fig. 14, rising to a maximum j^cMWHiMaNuii [MinntMof (wnlts of the iven, refer to out change of !9* 8* >. OF lAiX. o- ' — -• ProoowHngi.] LAW OF 0ONDBN8ATI0N OF BTBAM, 19 U) 4S SO a Metal. by the Fourier I order to oom- y and peculiar 40 4S SO r ^ INCH. jnts of the first triangular cycle r method. This to a masimuin of 10" at the point 0, and giving a very nearly constant temperature of —1° for Jths of the cycle. The corresponding curve of heat- absorption, and the curve of temperature at a depth of * 040 inch, may be compared with the simple-harmonic cycle, Fig. 15. The temperature range of Fig. 14 at a depth of 0*040 inch is found to be 4 • 6°, corresponding to a surface range of 1 1°, at a speed of seventy-seven revolutions in cast-iron. For the same range at a depth of • 040 inch, the simple-harmonic surface-range is 8° only. The surface-ranges differ considerably, but the quantities of heat absorbed are nearly equal. Tablb III.— Rbsclts or Obsbrvbd Wall-CSolks. II -vii XVIIa XVIU xviiS XVIII XVIII XIX XIX XXa XXa XXa XXh XXh XXb la Tempentnre Bulges. F -2 I 11 ai a t u i I 100 102 46 77 ma. Ins. OOIOC 0100 ! 0040 C 087 8 48 3 8 6 1 no o 61 1 00 318 295 4-5 0-88 819 296 : 9-4 a-75 813 287 ' 18-9 425 317 277 7»-4 70 4 70-4 07 45 6 45-6 43 8 43-8 47-7 47-7 47 7 81-7 81-7 817 00.S9C 089 C 0013 0039 I 0039 8 6 0-037 8 4 0039 I 0-037 8 039 ' 0-037 8 037 8 4 0-039 ! 0037 8 , 0-037 ■ a 4 4 6-8 4 6-7 5 6-6 8 3 6-2 3-5 6 4 4-7 7-2 4-9 7-5 13-5 20-2 4-6 7-2 11 17-2 4 3 6-7 3-4 60 8 8 14-2 8 5 6 600 1-60 0-67 1-45 0-61 4 28 0-64 5 00 70 824 301 835 810 835 810 329 305 829 251 827 262 329 305 328 291 331 307 331 298 831 265 331 306 330 292 331 264 2-58 2 64 2 61 1-90 2-43 3-45 4-20 9 13 3-80 7-65 8-31 2-32 4 60 0-61 0-60 0-59 66 64 0-60 53 0-66 0-64 0-64 0-58 66 0-67 65 The above Table includes all the observations of wall-cycles, for which complete data were available. In the third column C stands for cover, and S for side. For the latter junctions, the distance in inches along the side id also given. In trial X, the cycle at 3, inch, was taken against one of the junctions in the covt.r. The maximum card temperature is deduced from an indicator- diagram taken in the middle of the oyole. The mean wall tem- perature is that at the middle of the wall at the position of the junction, from observations taken before and after the cycle. These temperatures are probably right to - 5° F. c 2 HI 20 OALLENDAB AND NI0OL80M ON THE [MinntM of ;*■ Belation between Temperature, Speed and Condentation. — A oom- pariHon of the temperature-ranges at the surface of the metrl, with the steam-oyoles and the mean wall temperatures, appears to demonstrate that, even at the lowest speed of these trials, and making allowance for the form of the surface cycle, the time does not srffioe for the steam to raise the temperature of the surface of the wall more than a small part of the way up to its own tempera- ture. The largest surface-range of 20° P., observed in Trial XIX at forty-four revolutions, with a condensation lasting for nearly I stjcond, would raise the surface to 301° P. only, the temperature of the steam being 328° P. during the latter half of the interval. It is hardly possible to suppose this resistance to the passage of the heat from the steam to the metal, to be due to the existence of a surface film of oil or water. Assuming such a film to havu a conductivity only one-hundredth of that of iron, it would require a thickness of at least 0-020 inch to produce the observed eflFects, if the surface of the film itself were instantly raised to the steam temperature. The cylinder was frequently examined immediately after a run, but the interior was invariably found clean and nearly free from grease. The grease film was certainly less than one- thousandth of an inch in thickness. With regard to water, the case is perhaps stronger. The maximum observed absorption of 6 T.U. per square foot, would correspond to the condensation of a film about one- thousandth of an inch thick. If the resistance is due to a water film, the evaporation must be incomplete, and a film must remain from stroke to stroke. Farther, this must also take place so uniformly and consistently as to give perfectly regular resistauoe in all parts of the cylinder, and throughout the trials. The steadiness and consistency of the readings are perhaps the best proof, but there is other strong evidence, that such a film was not present. The obvious inference from the observations is that the rate of condensation of steam is physically limited, and it is necessary to assume a provisional law of condensation in order to compare the results. As a simple and workable hypothesis, and for other reasons, the rate of condensation of steam on a metal surface was asuumed to be proportional to tne difierenoe of temperature, and to be independent of the pressure. This assumption would make the amount of condensation taking place on any part of the walls, proportional per cycle to the average excess of temperature of the steam, multiplied by the time during which the temperature is above that of the walls. This product is found by measuring the condensation area on the cycle diagram included between the curves T [MlnutM of [lion. — A oom- of the metrl, res, appears to jse trials, and the time doea the surface of own tempera- in Trial XIX ing for nearly le temperatare f the interval, the passage of 3 the existence film to havti a would require i)eerved effects, d to the steam )d immediately Ban and nearly less than one- l to water, the [ absorption of udeusation of a le resistance is omplete, and a this must also give perfectly ^nd throughout le readings are evidence, that hat the rate of ; is necessary to to compare the and for other 3tal surface was perature, and to would make the ■t of the walls, iperature of the temperature is r measuring the bween the curves PmoeodiiigH.] LAW Or OOMDBMSATION OV BTB&lf. 21 representing the steam and the wall-surface cycles. It is con- veniently measured by the product of the time in seconds into the average differeno<- of temperature in degrees. If this area on the cycle-diagram is measured in degrees of temperature, and in sixtieths of a revolution, the result gives the condenstttion area in degree-seconds per minute. This result divided by the revolu- tions per minute gives, farther, the condensation area in degree- seconds per cycle. These condensation areas have been measured for eaoh of the observed wall-nyoles, and are given in the last column but one of Table III. The last column gives the ratio of the heat Q absorbed per cycle to the oondensatiou area A in each case. The approximate constancy of this ratio would appear to indicate that the hypothesis is at least a first approximation to the truth. Method of Eatimating the Total Condentation at any Epoch. — It is evident from Table III that the condensation area is fairly pro- portional to the amount of condensation taking place at any point over the range covered by the experiments. It is further assumed, if a vertical line is drawn on the diagram.of the cycle correspond- ing to the position of the piston at any epoch, the condensation areas measured to that line may be taken as proportional to the condensation which has taken place at each point of the surface up to the epoch considered. It is possible in this manner to arrive at a fair estimate of the amount of condensation at or shortly after cut-off, before re-evaporation has commenced. There is, further, evidence that the ro-evaporation follows the same law, and may be in many cases treated in a similar manner. Numerical Application of Method of Condensation Areas. — To make an estimate of the condensation in Trials XVI-XX, at 0*250 of the stroke, shortly after cut-o.., and at 0-700 of the stroke, shortly before release, the mean wall temperatures are assumed to be independent of the speed, provided that there is no wire-drawing or other change of conditions. Hence the condensation areas, measured on the cycle -diagram in degrees of temperature and in sixtieths of a revolution, are the same or nearly so at different speeds. It is sufficient, therefore, for this estimate to take the area so measured and multiply it by 0* 61, the mean of the values of the ratio v in Table III, in order to deduce the cyclical heat- A absorption for the parts of the surface considered. The following Table gives the total heat-absorption on the clearance surfaces. The portions of the barrel surface subse- quently exposed by the motion of the piston, require to Le M -.S' !■'? ="■■■1 22 OALLINDAR AND NI0OL8ON OH THK [Minatoi of treated somewhat diflferently. The temperatnre of the pUton w»« apeoially determined by means of a platinum thermomoter inserted through a hole in the piston-rod. Tabi B IV.— Cyclioai, Hbat-Aiwoiiition roH CLCAiuifca SuarAon. OniKltn- H**» ArM MMn Miloii Arf* AlMorlifd la Port'oM uf the Harfacs CoiuldcnHl of Tiimp«r- rwr ■'quiire Thfrmnl Surfic*. •liira. Mluuu. Unit K»hr. piT MInut*. Sqawt Fact. •F. oSwnmla. Cover, face 10 -6 inahea diameter . 0-60 80S 185 68 „ •ide 80 0-70 805 185 79 Pigton, fneo 10 -A inohci diamoter . 60 205 800 110 „ Bide 0-9 Oil 295 800 20 nnrrel, «ide 80 0-71 297 285 128 (^ouuterbore 5 , 12 291 880 28 PorU aud valve ...... 090 805 185 102 8umi and Minni . . 8-74 801 281 680 At 0-260 of tho stroke, 3 inches of the l)arrel surface have been exposed by the motion of the piston. Estimating the total contri- bution of this portion of the surface at 55 T.U.F. per minute up to 0'250 of the stroke, there is a total heat-absorption of 586 T.U.F. per minute at 0-250 of tho stroke, which oorrpsponds approximatoly under the conditions of the trials, to the conden- Fation of 0- 66 of a pound of steam per minute. The condensation per cycle at this point for any of the trials considered, may be obtained by dividing this rdsult by the revolutions per minute. It will be observed that the clearance surfaces contribute about 90 per cent, of the total condensation. Edimate of Re- evaporation during Expantion. — It is not possible to make an estimate of re-evaporation with the same degree of probability as condensatior, but some idea may be gained, by a similar method, of the amount of re-evaporation that has taken place before release. A probable excess of re-evaporation over condensation equivalent to nearly 300 T.U.F. per minute is found, and the quantity of heat rejected by the metal at release and during the exhaust period, generally called the exhaust waste, may be estimated at 260 T.U.F. per minute, under the conditions of the present trials. Steam Cycles. i ^;": The Steam Thermometer. — The thermometers, which were suffi- ciently sensitive to follow the changes of temperature of tho steam - -.^KlKiit^ate ^j ih i a ifai Mth Vf iBWigi ■waL--ija^ij-^i^i!£feijs;^.i.;^^.s^^ ■ [Minntoi of le piHton WM loter inserted HrRTAOH. iiiltn- II i\rf» "qimr* •I iw nut*. Abaorlwd In Tlimniil Unit Kibr. pr MInuM. oindi. 85 68 HA 79 ino 110 100 20 !8« 123 IMO 28 I8S 102 231 S80 aoe have been e total oontri- F. per minute b-abHorption of oh oorrosponds to the oonden- e condensation dered, may be 18 per minate. intribute about is not possible jame degree of 9 gained, by a that has taken aporation over linute is found, at release and sxhaust waste, ' the conditions lioh were suffl- ro of the steam '-'$ ''»'. PlOC««AB AND NIOOLBON ON THE [Minutes of indicator. It soon, however, rises above the indicated curve, and, with the exception of a sudden drop at release, due to adiaba'io expansion, continually approaches the temperature of the '\^all8 during the exhaust period. The actual superheating of the steam during compression must have been greater than that shown by the thermometer for two reasons. The method of observation gives the mean temperature during a certain interval of contact, generally ^ revolution, and cannot therefore reproduce a very short and sharp maximum. Secondly, although the thermometer was certainly very sensitive, the lag must have been appreciable on a rise of 100° F. taking place in 0*1 second. It is remark- able that the effect of radiation from the cool surrounding walls is not more noticeable. To test the effect of pure radiation, as compared with that of convection, on these thermometers, a special experiment was made, when it was found that the rate of loss of heat by pure radiation for this very fine wire, at tempera- tures between 200° F. and 350° F., was between fifty and one hundred times less than that due to convection. The possible error due to direct radiation from the surrounding walls does not, therefore, amount to more than 1° or 2°, and the thermometer is really indicating the temperature of the steam around it. The peculiar characteristics of the platinum curve in Fig. 17 were verified on several occasions, and with different settings of the valve, the results observed in every case being similar. Piston Steam Thermometer. — To observe the temperature of the main bodv of the steam at a distance from the walls, a similar thermometer was attached to the piston in the manner shown in Fig. 16. The thermometer projected from the piston for a distance of about 3 inches, and was received at the back end of the stroke in a tube 1 inch in diameter in the centre of the cover. The indi- cutions of this thermometer at different speeds and at different settings of the valve, were in remarkably close agreement with the card. Systematic differences, however, were always observed, which, from their consistency and from the great number of obser- vations, cannot be attributed either to errors of the indicator or of the thermometer. The curve in Fig. 18 is drawn from the card taken simultaneously with the observations of the platinum thermometer. The differences observed are almost too small to be shown. They are much greater, however, than the uncertainty of the observations themselves. The temperatures at admission and release were much steadier than in tha cover. At most other points of the stroke it was possible to take readings to 0*1 degree, and the extreme variations often did not amount to ■ i! ft'; A ,*.^,iiit»*VMii»:f^t^^e6yisiii> ■ I [Minutes of sited curve, and, iue to adiaba'io re of the walls ag of the steam that shown by of observation rval of contact, produce a very lie thermometer een appreciable It is remark- rrounding walls pure radiation, /hcrmometers, a that the rato of ire, ct tempera- n fifty and one The possible ; walls does not, thermometer is iround it. The n Fig. 17 were settings of the ilar. iperature of the walls, a similar lanner shown in >n for a distance id of the stroke )ver. The indi- and at different agreement with Jways observed, number of obser- le indicator or of Q from the card f the platinum it too small to be the uncertainty res at admission jover. At most ike readings to 1 not amount to ProooedingB.] LAW OP CONDENSATION OF STBAU. US more than 0-5 degree for several minutes. The superheating shown during compression was small; the greatest amount was shown in the middle of the admission period. The centre of the steam appears to have been at this epoch generally 4° P. or 5° F. above the indicator. Fig. 18 shows the smallest value recorded, namely, 2° F. Other examples will be found in Fig. 7 and Fig. 10. Throughout the expansion curve, the readings of the platinum thermometer were between 2° F. and 3° F. lower than the indi- cator. At the end of the stroke, and for part of the exhaust period, the temperature fell to between 207° F. aud 208° F., but recovered to 212° F., or close to the barometric temperature, before the end of exhaufc^. This may have been due to a real lowering of pressure, Fig. 18. >ali!WMU(m>»'~ 400' - 5 i S PLATI HOMT IBRM( IKTBI m n S^of!.! xmi LSISf ff o issd -^ > ^ '■ 1 > b= Mti NWAI UTl* •INC 1v^m aqji 4 300* 1 -f Nt: / N r. r — !?, &2SO J N ■ mit TWTT '.Vfel ZiJi rr- 5 c GxiBTB 8 owing pressui differei . . of the large indicat Testi possibl adjuste e,ud be it is hs would and th Ml OXL tSPS .-ms ^MIDDlir24«i! F — /- V ^^ r- ■^^ « SO S5 ME IN SIXTIETHS OFA f HO^VISa AQBKEMBNT to the rapidity of ■e slightly below 1 ice of pressure is steam during ex] Etnd regular to 1 or ; lag of tbo th ng the Indicators.- e under the oon( d ; they were dai ick-lash. At the irdly possible tha be required to ex 9 platinum therm icvoumoH'njoifBAc/tMDorsT^SKE**' ** "" BETWEEN InDIOATOB AND PLATINUM ThEBMOMETBB. condensation in the surface condenser at a hat of the atmosphere. The corresponding only 1 lb. The lowering of temperature >ansion appears, on consideration, to be too 30 explained by any error or lag of the Brmometor would have the opposite effect. —The indicators were tested as nearly as iitions of the trials, and were carefully ly oiled and cleaned, and tested for friction comparatively low speeds of the trials, t they should have so considerable a lag as plain the difference between the indicator IOmeter. 26 OALLENDAB AND NI0OL8ON ON THB [HinntMof Teat of Scnsttiveneag. — A test of the sensitiveness of the platinum thermometer, in which the engine waa run with air instead of steam in the cylinder, with a view to determine the probable amount of lag, is illustrated in Fig. 19. The lag could not have been greater than 2° with the temperature of the air rising at the rate of 100° per second, but it is not possible to obtain any form of indicator sufiBoiently sensitive and acoorate to perform this test satisfactorily. Superheating due to Wiredrawing.— Fig. 20 gives an illustration of a diflferent kind of steam-cycle, taken during the measurement Fig. 19. 45 50 5S 6 & 10 IS ZO 2B 30 3S 40 45 50 time in sixtieths of a revolution from backend of stroke Tempehaturk Cyclk with Aib in Cylindkb. of a valve leak, and showing a very unexpected amount of superheating. The temperature of the steam leaking into the cylinder under these conditions, as me&sured by the platinum thermometer on the piston, is shown by the dotted curve. The full curve below shows the temperatures deduced from the indi- cator on the assumption that the steam was saturated. Conclusions. — From the steam-cycle observed in the hole in the cover, it appears that, even within -^ inch fjpm the walls, the temperature of the steam is greatly afifected by adiabatic com- pression and expansion, but that during comparatively quiescent periods of the cycle, such as the exhaust, the steam close to the walls is heated nearly to the wall temperature. In a single- f ■a»*ftw^t»(ft^,*M5.'iE*^'»^^v,'-**^ i^'-&f,«\^**j»>-«^ «i&^v.w:a g^*i*t.;^«.»ai«ife««y<«^r*amifttaay^:jc;A.ffig^i^^ [Minuteaof the platinum ir instead of the probable aid not have rising at the 1 any form of >rm this test a illustration measurement Proceedings.] LAW 6» OONDENBATION OF 8TBAM. 27 )d amount of king into the the platinum I curve. The rom the indi- i. le hole in the the walls, the idiabatic com- vely quiescent [u close to the In a single- acting non-condensing engine, with a moderate degree of compres- sion, water is not likely to collect in the corners and recesses of the clearance; and the clearance contents, consisting chiefly of superheated steam, cannot be regarded as a primary cause of con- densation of the admission steam. The fact that the platinum thermometer, after falling below the indicator at cut-off, crosses it again at a temperature slightly below that of the walls, and then rises considerably above it, shows that re-evaporation from the cover is probably complete some time before release, and that evaporation from a highly heated wall is probably a process of a rapid and explosive character. From the piston steam thermometer, the temperatures deduced for the indicated pressures Fig. 20. aX so SSO S l0l5 2O2S3O35*0*550 TIME m SUCTIETMS OF A REVOUiTION FROII BACK (MD OF STROKE. DUOBAH IIXrSTBATING SUPBBHBATING DUE TO WlBB-DRAWINO. seem to represent very fairly the average state of the main body of the steam, but the steam is probacy slightly superheated during compression and admission, and slightly supersaturated during expansion and exhaust. The eflfeots observed are probably too large to be explained by lag of the indicator. The superheating of the steam during admission may be partly explained by the further compression of the already superheated cushion steam, which in the present case forr 3d one-fifth of the cylinder con- tents. It is also partly due to the kinetic energy of the inrush. In any case it is evidence that the steam supply from the boiler was fairly dry. Ur ier the conditions of the trials it is impossible that the steam supply could have been superheated. In fact, MilBiiMHWMiirtiiMM ■«4^ 28 OALLENDAB AND MI00L80N ON TKB [HinniM of thermometers in the steam-pipe and steam-ohest indicated the normal temperature. The results of tests with a number of different thermometers showed the temperature of the steam during expansion, and during the early part of the exhaust, to be lower than that deduced from the indicator. Dynamical Equilibrium of Expanding Steam. — Apart from the statical condition of equilibrium between steam and suspended water-drops, which depends on the size of the drops and on the value of the surface tension, there is also a dynamical condition in the case of rapidly expanding steam, which has not, so far as the Authors know, been previously noticed. When steam is expanding adiabatioally,it requires, as is well known, a condensation equivalent to nearly one thermal unit per pound per degree Fahrenheit of fall, to keep it up to the saturation temperature. This condensation must take place chibdy on the surface of drops already formed. The temperature of the drops can be maintained only by continual evaporation. Unless the steam condensed is at a lower temperature than the drops, there can be no balance of condensation, and the temperature of the drop cannot fall. The lowering of steam- temperature required will evidently be proportional directly to the rate of condensation, and inversely to the surface exposed by the drops. Since the drops are at once foci of condensation and foci of heat, there must be powerful obstructive influences at work, and the lc««rering of temperature of the steam may therefore be considerable. In the absence of more certain indications these obstructive influences may be assumed similar in magnitude to those which limit the rate of condensation of steam on a metal surface. It is possible to make a numerical estimate of the order of the lowering of temperature. At n revolutions per minute, the initial rate of fall during expansion, in the experiments at one- fifth cut-off, may be taken as about 10 « °F. per second, and the required balance of condensation as 8 n thermal units per pound per second. If the wetness of the steam is 1 per cent., and the average diameter of the drops 0-000024 inch, the surface exposed per pound of steam would be nearly 480 square feet, and the lowering of the steam ter-.perature at 100 revolutions would be about 2*7° F. The lowering required increases in direct pro- portion both to the speed and the diameter of the drops. With nearly dry steam at high speeds the initial lowering may be considerable, especially if the drops are large and few. If it be admitted that the temperature of rapidly expanding steam may fall considerably below its saturation temperature, it affords a possible explanation of a certain loss of efficiency in high-speed ^"ygf !.>.»i j riAt- ' i'iiilS!U4*B [Minutes of idioated the number of iteam during to be lower f rt from the 1 suspended , and on the condition in so far as the is expanding m equivalent nheit of fall, sondensation lady formed, by continual temperature ion, and the ig of steam- l directly to I exposed by ensation and influences at lay therefore cations these aagnitude to 1 on a metal of the order r minute, the lents at one- ond, and the its per pound ent., and the rface exposed ieet, and the ns would be i direct pro- irops. With ring may be iw. If it be ; steam may it affords a n high-speed Prooeedinp).] LAW OF CONDENSATION OF BTBAH. n-. engines. From the behaviour of other vapours in a supersaturated condition, the pressure, if the steam remained dry, would be much more reduced for a given expansion of volume than it would be if the steam had time to maintain itself by condensation at its saturation temperature. The statical condition mentioned affords an explanation of the unwillingness of steam to condense otherwise than on dust nuclei, or on drops of water already formed. It will be observed, however, that in order to account for a lowering of 2 or 3 degrees, the drops would have to be of so minute a size as to be invisible in the most powerful microscope. The linear dimensions of the drops actually occurring in steam-engine practice are probably between one hundred to one thousand times greater. For the dynamical condition, on the contrary, the larger the drops the less the surface they expose, and the greater the fall of temperature required. A diameter of j o-q-sI) inch would mean a fall of about 50° F. below the saturation temperature at 400 revolutions per minute and wetness 1 per cent. In the absence of accurate knowledge of the properties of steam under these conditions, it is not possible to say exactly how much missing steam such a fall of temperature would account for. It would probably be between 5 per cent, and 10 per cent., according to the extent of the drop surface. The subsequent recovery of temperature, as more drops were formed and condensation proceeded, would simulate the effect of re-evaporation. The initial fall of temperature at a high speed, however considerable, would probably be of very short duration. It is not improbable that steam in this supersaturated condition may tend to condense more readily on any surfaces exposed to it, which happen to be below the saturation temperature, than would be the case with ordinary wet saturated steam. The temperature of the steam itself inay have some influence, as well as that of the surface on which it is deposited. For the same degree of adiabatic expansion, the heat abstracted from the walls would be the same, whether the steam condensed is dry and supersaturated, or wet and saturated, provided that in the latter case the suspended moisture ' is deposited along with the steam. The amount of condensation, however, might be greater in the case of the supersaturated steam, if the temperature of the steam itself has any influence. Supposing the steam is many degrees below its saturation temperature during rapid expansion, no thermometer, however sensitive, could indicate the whole extent of the phenomenon. The rapid motion of the steam and the piston might tend to cool it, but the condensation on its surface would tend to keep it near mm 30 OALLENDi^n AND NI0OL8OM OH THE [Minutna of the sataratiou temperature corresponding to the pressure. The lag of the thermometer would also make the reading too high. If, on the other hand, it is inoonoeivable that the thermometor should indicate anything but the saturation temperature, the differences observed must be due to a real difference of pressure, owing to the rapid vortical motion of the steam, between the centre and the circumference of the cylinder contents. Such differences must exist, and have often been regarded as important. It is probable that further experiments of this nature might throw some light on the question. Valve AND Piston Leakage. ■ ^^. l,.C»t^j " The Slide Valve. — In estimating the amount of condensation in the cylinder by comparing the measured feed per revolution with the steam indicated by the diagrams, the valve and piston leak is generally assumed to be negligible. The effect of leakage. Figs. 21. Ibl however, is in many ways so similar to that of condensation, that one may readily be mistaken for the other ; and no estimate of con- densation deduced from diagram and feed measurements can have any claim to consideration, unless the state of the valves and piston in respect of leakage is simultaneously investigated. In addition to trying the stationary test for leakage, which is very easily applied, the leakage was measured as accurately as possible under the actual conditions of running. The stationary test was found to be of little oi no value. Preliminary Leakage Testa, — To measure the leakage into and' out of the cylinder, with the valve in motion under the conditions of running, but just not admitting steam directly, the piston was blocked and cards were taken, the barrel motion being obtained from the valve spindle, and the engine being driven by a motor. Figs. 21, a and b, are sample indicator-diagrams taken at the back end of the cylinder, in the above manner. The first, a, showing a maximum of 20*3 lbs., was taken on July 29th, diagram b on [Minut<» of ssBure. The ig too high, thermometor erature, the of pressure, between the «nts. Such s important, ature might densation in olution with piston leak of leakage, nsation, that imate of oon- nts can have esand piston In addition very easily jssible under ist was found ige into and' lie conditions e piston was ling obtained L by a motor. a at the back t, a, showing liagram b on ProoewUngs.] I^W OF CONDENSATION OF 8TKAH. 31 August 29tb, after the valve had been very carefully scraped and refitted; diagram c is one taken at the crank end on the same date. The following were the conditions of the test : — 1 Indl- ATerage Preirate. Maxi- mum Tempera- ture. Volume of Cylinder. Revolu- tliina per Minute. »»'•• dl«Krain 1 (Type). 1 a»uge. Card!. Maximum. Leak. July 29 . . i o August 29 . b August 29 . ; e 970 81 810 19-2 4-87 SB e 801 816 Feet. 0-860 0-618 o-ocs 41-8 75-7 75-7 IM. per Knur. 88-6 806 8 6 Fig. 22. A smaller indicated pressure in the second case was nearly compensated by the higher speed and greater volume, so that the resulting leak deduced is not very far from being proportional to the difference of pressure under which the leak took place. It would appear that the leakage is not merely a question of such minute differences of fit as those corrected in the scraping. Direct Exhauit Lea/fcagfc— Preliminary trials showed the direct leakage of steam from the steam-chest into the exhaust to be by far the largest and most im- portant. In order to measure this leakage as nearly as pos- sible under the conditions of running, both the steam ports were blocked with lead, and the valve was driven by an electric motor, the piston being disconnected. The following are the results of two experi- ments made with the same valve setting as for the later series of trials at one-fifth cut-off. In the first trial, 1 1 2 lbs. were condensed in 25 • 1 7 minutes, at a gauge pressure of 91 lbs. per square inch, and the rate of leak appeared to increase slightly as the oil-film was gradually dissipated. The effect of the dissipation of the oil-film is well illustrated in the second trial under the same conditions, for which the results of the separate weighings of the feed have been plotted in the curve shown in Fig. 22. In this trial 317 '6 lbs. were oondensed . pressure per hour should be proportional to the perimeter j) c^ the port, divided by the mean overlap I. nature of the oil-film, which should be the same for the same type Thus k = — ,— , where is a coefficient depending on the V [Miiiutet of The initial value, agrees , the leakage ) of pressare. di£ferenoe of the "rate of ind difference bis particular ber Jfc = 3-00 J within the this leakage be most per- investigate Dade on the mgine. The 1 a pressure sxhaust pipe, dth pressure le valves are and guiding a high order. T steam-tight of transpira- oe chiefly in 1, considering itent both for leakage of a 8, depending proportional neter of the irfaces. The loving valve, be following »ure through kness in each ir should be by the mean mding on the le same type ,-y'f- V ProoeedingL] LiW OF OONDBNSATION OF STBAK. «t of valve. It might possibly bo nearly the same for valves of different types, if it depends on son^e physical property of a mixture of oil and water. '''■'' '■'■■''■' 'w 5- -if: H-^:- Tabli v.— Oomparisov or Batm or LiAEAoa Vtire oonildarad. PeriiMter of Port, OvrrUp (mMii), Rttlo. P Obv-rved tUtouf Lwk,*.' Dcdnood Volua of Balanced, Robb .... InobM. 72 IlMlhM. 0-6 144 8-00 021 Unbalanced H.H.P. . . . L.P. . . . . » 1, > • . . 80 6S 65 1-5 10 10 20 6S 0-88 1-20 1-88 0019 0019 0-021 The leakage coefficients observed with the different valves, are at leant of the same order of magnitude. If so, we are justified in oonoluding that the leakage probably takes place in the form of water, and is proportional to the difference of pressure. It would also appear probable that such leakage is the normal state of things with a moving valve, and that the excessive leakage observed with the balanced valve, is not a defect peculiar to this type, but is simply a consequence of its comparatively large size. At the high speed for which this engine was designed, the leakage would not be a very serious matter. If the leakage were simply due to accidental tilting motions, or to bad fitting, it is difficult to see why the value of the coefficient should be of the same order for valves of such different types. If, on the other hand, the leakage is not purely accidental, but follows a regular law, it is a matter of such practical importance as to be well worthy of further investigation. Leakiige into Cylinder after Cut-off. — Prom the experiments described on p. 31, it is possible to obtain an independent verifi- cation of the value of the coefficient C. With I = j^ inch, and p = 20 inches, the values are C = • 022, and = • 020 respectively. By means of the time integral of the expression ^^ ' -- \ the leakage taking place into the cylinder during expansion between the points 0-25 and 0-70 of the stroke, for trials XVI.-XX., amounts approximately to 6 • lbs. per hour. The smallness of this result is due to the fact that the difference of pressure is small just after cut-off when the overlap is least. Piston-Leak under Conditions of "Running. — Under running con- [thE IN8T. as. VOL. CXXXI.] D mmmmmmmmmialmiimm :••% !-■* 84 OALLBMDAB AND MICOLSOM ON THB [M inatoa of ditioDB, the steam leaking past the piston was foand to amonnt to 15 '3 lbs. per hour under a mean pressure of 33 lbs. The pistor- leak taking place during expansion between 0*25 and 0'70 of ihe TABLX VI.— COMPAMIOM OF OaBM AHD FllD WITH WaIX-OTOUS. single- Acting Non-Condensing Trlali. Out-o«f bI ■ 200 of itToke. Cleamnoe, 10 per oeut. ; Releaae, 0'7S0; Expan«ioni (cut-off to relesie) 2-83. 1. No. of trial . . . 2. Duration in minutes 8. Mean revs, per minute 4. Mean gauge preuure XIX |xvin 370 68-0 48 8 45-7 87-9 89-2 5. Grow feed per rev. . 6. Leakage correction . 7. Corrected feed per rev 8. Calculated ouBhion\ Hteam . . 9. Total weight of Juid I expanding in oy-> Under . . . .) XXa 650 47-7 94-4 14221 -1437 1004-0976 04180461 0107! 0104 0525-0565 Mmiw. 1.2,3 46-7 90-8 1488 0990 0493 0108 . XVIIo XVI 79-0 760 70-4 ;78 4 98-1 ,92 0462 0596 10. Indicated weight at\ 0-250 . . . ./ 11. Indicated weight at 0-700 .... 12. IncreHie of indicated^ weight . . . .j 13. Adiabatio condenB-> ation . . . ./ 0407|-0414 0466 -0456 0059 -0042 0019,-0020 14. Indicated evapor-) atiou ... .J 15. Calculated evapor- 1 atiou ... ./ 16. Indicated ooudena ation 17. Calculated condens- ation .J • • • J ,„ 1' Water Uper cent, of ^'•j present \ feed (7) f per cent. ( fluid (9) 19 1 at J per cent, of [ 0-250 J\ " 20. Indicated HP. . . 21. Lbs. per I.UP. hour 22. Condensation per\ i.ttP. hour . ./ 0078-0062 0076 0073 0487 0488 0051 0021 0118-0151 0148 "14-2 28-8 22-5 82 7 26-8 4-10 26-8 7-6 4-84 89-1 9-5 0072 0070 0667 XXo 85-0 81-7 04-2 10941 -1086 0697|-0627 0897-0409 0099-0098 0496 0418 0460 -0507 1000 0676 Hetiu 4,1,6 78* 7 95 1 0424 0100 0524 0894 0436 0042 0020 0159 0136 32-8 26-7 4 78 29-5 9-5 0069 0078 0146 0142 31-7 25-8 4-48 28-6 9-1 0062 0048 0078 0092 20-0 15 7 7-02 23-8 4-8 0042 0019 0408 0454 0046 0020 0407 0506 XVII6 250 97 96 0856 0494 0862 0105 0467 0061 0046 0118 0089 27 7 22-8 6 67 271 7-5 0066 0041 0062 0046 0116 0080 0426 0019 0062 0085 27-8 22-1 24-8 19 6 7-71 26-9 7-3 0099 0088 7-00 25-7 6-8 0074 0067 20-4 15-9 8-81 28-8 4-8 [MinatMof to amount to The piston - 10*70 of the .l-Gtouh. of stroke. >releaM)2-8a 6 Me ins 7 XXe SO 1-7 1-2 78" 7 95- 1 XVlIb 250 97 96 1000 0576 •• •0856 ■0494 0424 •0407 ■0862 0100 • • •0105 0524 •0505 •0467 0408 • • •0898 0454 •• •0426 0046 .. •0083 0020 ■ ■ •0019 0066 •0062 •0052 0041 •0046 ■0035 0116 •0099 •0074 0080 •0088 •0067 J7-3 248 204 J2-1 19 6 15-9 7-71 26-9 700 25-7 8-81 238 7-8 6-8 4-8 ProoMdiDgi.] LAW or OOlTDBnATION OF fflBAll. 18 •troke, under the conditions of trials XVI-XX, amonnte to only 2 ■ 5 lb*, per honr, the greater part of the leak ooonrring daring admiwion. Oompariton of Indicated and Oaleulated Oondmuation. — The reenlta given in Table VI, with the exception of the first fonr and the last five linee, are expreaeed in terme of the weight of steam and water per cycle in lbs. The differencea of the weights in lines 9 and 10 are given in line 16. They depend entirely on the rate of leakage assumed for applying the very large leakage correction given in line 6. Oonsidering the magnitude of this correction, it is most remark- able that the weights in line 16 should have so great a degree of consistency. A variation of only 8 per cent, in the rate of leak would be sufficient to explain the largest discrepancy from the mean. The order of consistency shown under such differences of speed and pressure, and probably also of lubrication, from trial to trial, is perhaps the strongest proof that could be given that the phenomenon of valve leakage is subject to regular laws, and deserves much more attention than it has hitherto received. Con- sidering the range of speed covered by the trials, there is strong evidence that the rate of leakage is nearly independent of the speed of reciprocation of the valve. The evidence that it is simply proportional to the pressure is much less oo'aolusive. Line 17 gives the total condensation at 0*250 of the stroke, calculated Arom the results of the observed wall-cycles for three low-speed trials. This assumes that the cyclical condensation per minute is independent of the speed provided that the temperature conditions remain unchanged. It is evident, on comparing the numbers in lines 16 and 17, that the indicated condensation agrees as closely as can be expected with this view. There is perhaps a slight indication of a greater rate of condensation at the higher speeds, but this is partly accounted for by the hi£;hf«r mean pressure and temperature corresponding to these trials. TIio results in line 17 were calculated to correspond to a gauge-pressure of 00 lbs. The Nature and Effects of Valve Leakage. — The foregoing experi- ments would make it appear probable that a moving valve, however well fitted, is subject to a regular leakage of a peculiar type, which has not been previously suspected. The leakage appears to take place in the following manner. So long as the valve is stationary, the oil-film may suffice to make a perfectly tight joint, but as soon as it begins to move, the oil-film becomes broken up and partly dissipated. Water is being continually condensed on the colder y i d2 r „ .r ^mm 86 OALLnrDAR AND NI(X)L80N ON TBI [M inotM of parts of the Burfaoe expoMd by the motion of the TaWe. Thii water works its way through, and breaks up the oil-film under the combined infiuenoe of the pressure und the motion. The oontinnal re-evaporation taking place in the exhaust tends to keep the vaWe and the bearing-surfaces of the seat cool, and to maintain the leaking fluid in the state of water. The exhaust steam from the cylinder has the same tendency. The coefficients of viscosity of steam and water at the temperatures which occur in a steam- engine are not accurately known. But whereat that of steam increases with rise of temperature, that of water diminishes very rapidly. It is not improbable that the quantity of water which can leak through a given crack under a given difference of pressure, may be from twenty to fifty times gpreater than the quantity of steam which can leak under similar conditions. This agrees with well-known facts in regard to leakage, and explains how it is that the leakage in the form of water is so great. A few simple experiments were made with regard to the transpira- tion of water and steam under the conditions in question, and the !i«kage in the form of water was more than twenty times as great, the water being at a temperature below boiling point. The motion both of the water and the steam, owing to the high velocity, was certainly turbulent or eddying, which would have the effect of greatly increasing the resistance as compared with that due to viscosity, if the motion were steady. For the case of steady motion, comparative tests were made of the relative values of the viscosity of water, cold and hot. The moasurements were not sufficiently accurate to give the law of the variation of the viscosity with temperature above 212°; but it appeared that the viscosity at 212° F. was only one quarter of that at 62° F., and that it con- tinued to diminish very rapidly. Under the actual conditions of the valve-leak .experiments, the water leak t^ more likely to have been between forty or fifty times the s.eam leak. An explana- tion is thus furnished of a possible form of leakage, indirectly due to condensation and re-evaporation — so many times greater than the steam leakage, which, alone, engineers have been in the habit of contemplating, that it might well claim attention on its own merits, apart from the very limited number of valves on which it has hitherto been possible to make direct experiments. ; ''#;-.: The analysis ot a large number of observations, in addition to the few made by the Authors, leads to the conclusion that all valves leak more or less when in motion, and that in many oa8<» the greater part of the missing quantity is to be attributed to leakage of this description. Whatever the precise manner in [MtnntM of T»We. Thit Im ander the 'he oontinnul >ep the Talve nainUin the mm from the ▼isoosity of in a steam- lat of ateara ainisheti very water which difference of ter than the itiona. Thia and explains ao great. A he transpira- ^uestion, and snty times as ; point. The ligh velocity, ive the effect ;h that due to teady motion, ' the viscosity )t snffioiently 'iscosity with ) viscosity at I that it con- conditions of likely to have An ezplana- ndirectly due greater than a in the habit n on its own >s on which it a. in addition to ision that all n many oas'W attributed to le manner in Prooewling*.] LAW Or OONDIMBATIOM Or VIBAM. 81 which the leak takes place, it appears to be nearly pro|)Qrtional to the difference of pressure, and to be in most oases independent of the speed. In any cms, it appears probable that the leakage is connected in some way with the condensation taking place on the valve surfacas. If so, it way evidently be greatly reduced, if not entirely cured, by jacketing, or otherwise heating the valve-seat, to minimise the condensation. These views have an important bearing on the design of valves. For low-speed engines, separate steam- and exhaust-valves should poHsess advantages over the ordinary slide-valve. The sn|>eriority of the oom^iound engine would also appear to be partly due to the groat reduction of pobbible leakage. Gknkral Conolusiomb. On the Bate of Condetuation of Steam. — The most imjiortant general conclusion to be derived from the experiments is that the rate of condensation of steam on a metallic surface is limited, and is proportionvl to the difference of temperature. The small increase observed in the ratio --, *>. 21, with increase of speed, A may be most naturally explained as due to the diminished thick- ness of the water-film deposited, and the smaller range of the metal cycle. The slightly smaller values observed on the cover as compared with the sides, though possibly due to slight differences in soldering or in the form of the cycle, may also be attributed to the state of the snrface. Taking these factors into consideration, the probable rate of condensation of steam on a clean and dry metal surface is found to be 0'74 thermal units per second per square foot per 1° F. difference of temperature at 300° F. Expressed in more familiar quantities, the rate above given would correspond in a surface condenser to the condensation of 27 lbs. of dry steam per square foot per hour for a difference of temperature of 10° F. between the steam and the surface. • Condensation in Term$ of Temperature Distribution, — Assuming that the amount of condensation is limited by the rate of condensa- tion of steam given by the above law, the problem of estimating the amount of condensation taking place in any given engine with any given steam cycle, is reduced to the (oompiiratively) simple problem of determining the temperature distribution on the cylinder walls while the engine is running. Given the temperature distribution, the condensation is inferred by measuring the ooadeusation areas on the cycle diagram. '•*><■ mmmamm mmm m A1-" 88 OALLENDAS AND KI0OL8ON ON THE [HinntM of Limit of Cyclical Oondetuation for any given Cycle. — That it shonld be necessary to observe the temperature distribution in any caffe, in order to be able to deduce the cyolioal condensation, may appear at first sight a somewhat disappointing result. The form, how- ever, of the law of condensation and re-evaporation as deduced from the experiments, leads directly to a limiting value of the cyclical condensation, and gives a result of strikingsimplioity, which is undoubtedly applicable to a large number of important cases. If the conditions of eztemal and internal heat loss are supposed to be such that tho mean temperature of the clearance surface, on which the greatoi' part of the initial condensation takes place, is reduced to the m'jan of the steam cycle, it is plain that the con- densation and evaporation areas on the cycle diagram will be equal. If thfi temperature of the clearance surface falls below this point, evaporation will be incomplete. Water will then accumulate in the cylinder until a balance is attained by the mechanical removal of the e&oess. It is obvious that all steam condensed on the surface, and then mechanically rejected in the form, of water, repieseiits the communication of a quantity of heat to the walls, tquivalent to the total heat of the steam condensed, and thus rejected, reckoned from the temperatnre at which the water is thrown off from the walls. The quantity of heat thus communi- cated per lb. of water rejected, may be between twenty and fifty times greater than that communicated by the condensation and subsequent re-evaporation of an equal quantity of steam. If the engine starts cold, and the surfaces are gradually heated by the action of the steam, it is clear, from the same considera- tions, that the rise of temperatnre up to this point, so long as water is being mechanically rejected, will be extremely rapid. The mean of the steam cycle is, therefore, \,ii the provisional law, a natural minimum of temperature for the wall surface, corre- sponding to a maximum limit of condensation for any giveii cycle. In order to deduce the limiting value of the condensation per square foot per hour for any given cycle, it is simply necessary to draw the cycle diagram corresponding to the indicator-diagram, and to rule across it the line representing the time average of the ateam temperature, as shown, for instance, in Fig. 18. The area above this line is the maximum condensation area corresponding to this particular cycle. The maximum value of the condensa- tion, measured in thermal units per hour, is forty-five times this area measured in degrees F. and sixtieths of a cycle. As a general rule, the condensation must be less than this limit, because the temperature range of the surface of the metal, which - n-, ,\m t m. ' i ' !.fv^i:M!i^ »!,. a i ^ I Ji^ ' ;:iHJ!^'^! \ ■ipailMiil [MinatM of TliatitBhoald in any oai: %nd walls at any point of the cycle. The quantity of heat giv?':. to the walls by the condensation and re-o^"poration of dry steani, is given by the formula, Q = * 30 (<* — t"), per pound. A more elaborate or appa- rently exact formula, is useless, because the valuo of the constant 0-30 in this expression, being the value of the change in the total heat of steam per 1°, is one of tho most uncertain elements in the whole theory of the steam-engine. According to the experi- ments of Begnault and Griffiths, the mean value of this constant between 30° and 100° G. should be 0*40, but it appears not improbable that itu value diminishes with rise of temperature. It has often been pointed out that, as a result of the compara- tively small rate of increase of the total heat of steam with rise of temperature, a relatively small loss of heat, or a slight change of conditions, is competent to account for a considerab^ change in the initial condensation. The balance is extremely delicate, and is very easily turned. The farther possibility, that there should be a limit of condensation, results from the form of the law of condensation, and could not have been foreseen so lon^ ^ ■ ■Mfafettf^ttU st i m u rM* i»)i>fi! iijjiiwMi.mil ^-fm^ .■..ri.-i.:txn..:.:^i.'KMfiif3.iimitl»,l'. 40 OAIitEMDAB AND NIC0L80N ON THK [MinnteB of as the rale of oondenaation was regarded as infinite. When this limit is reached, the conditions as regards increase of condensation are extremely stable. If the limiting range i — <" is 60° F., which is not uncommon in compound engines, a loss of 15 T.U. is sufficient to account for each pound of initial condensation and re-evaporation at this range. But if at this point the rate of loss of heat is suddenly doubled, the initial condensation will be increased by less than 2 per cent. If, on the other hand, the rate of loss of heat were reduced to one-quarter, the initial conden- sation and the range of temperature between condensation and re-evaporation would each be reduced to one-half of their limiting values. It is also interesting to observe that the form of the law of condensation would make the limiting value of the condensation in any cylinder depend chiefly on the temperature range in that cylinder. Of all the results which have been empbioally established with regard to cylinder condensation, this result has always been regarded as the most certain. To express the result more accurately, according to this law, the limiting value of the condensation, when measured in pounds per hour, should vary as the area included between the steam-temperature cycle curve and the line representing the mean temperature of the steam-cycle. The limiting value should also increase slightly with increase of speed, because the temperature range of the metal surface is reduced, and the effective condensation area is thereby increased. Correction for the Metal Surface Cycle. — It is clear that no simple formula can be constructed to take account of all the possible varieties of cycle. The correction is not large, and might be neglected if it were not that it varies with the speed, and with the point of the cycle considered. In the majority of oases which occur in practice, the maximum point of the wall-surface tem- perature cycle is found to coincide very nearly with the point of ci ofif, and the point at which the steam-cycle curve crosses the wall-surface curve is generally very near the point of release. For these two points of the cycle, in the case of limiting condensa- tion, a sufficiently approximate correction can be applied for the effects of variation of speed by the following simple method. The condensation area is measured from the line of mean wall- temperature, and up to the line of cut-off or release as required. The area so measured is then reduced in the proportion jr. ;- I for the point of cut-off, and in the proportion -, ~ . l (3 -h V") *^ ^ (3 -H V") for release. The correction at the point of cut-off varies from 26 per oent. at » = twenty -five revolutions per minute to 9 per cent, at four [MinnteB of . When this oondensation <" is 60° F.» IS of 15 T.U. oondeusation loint the rate ensatioD will ler hand, the litial oonden- ansation and beir limiting a of the law condensation , knge in that empirically . is result has IS the result ' value of the ould vary as le curve and steam-oycle. I increase of 1 surface is increased. at no simple the possible 1 might be nd with the , jases which urface tem- the point of crosses the of release, g oondensa- iied for the lie method, mean wall- is required. proportion (3 + Vn) from 26 per lent, at four Proceedings] lj»,w 0» OONDBNBATION OF BTBAM. 41 hundred revolutions per minute, if the steam-oycle remains the same. This formula assumes a cast-iron cylinder. The main effect of making the correction is to increase the value of the oondensation constant from 0*61 T.U. to 0-74 T.U. per degree- second per square foot, as deduced from Table III. It also has the effect of making the observations agree rather better with the calculations at higher speeds, both in Tables, pp. 19 and 34. Effect of Initial Wetnen of the Steam.— In view of the observations on the effect of the oondensation of wet steam, it is interesting to make an estimate of the possible increase of oondensation thereby produced. It is, however, necessary to make a few assumptious, which are probably not in all oases justifiable. Let X be the dryness fraction of the steam condensed at a temperature t', and let it be assumed that the proportion of sus- pended moisture (1 — te) is all deposited on the walls together with the oondenseid steam. Let it be further assumed that the whole of this water is re-evaporated non-explosively at an average temperature <". The mean temperature of the wall-surface, t% will be somewhere between these two extremes, and will be modified by conduction from the neighbouring parts of the cylinder. To simplify the conditions the point considered may be situated on the clearance surface near the middle of the cover, or the cylinder to be so large and thin that the effect of conduction may be neglected. The other losses may also be supposed negligible in comparison with that due to re-evaporation. If L' L" are the latent heats of vaporization at the temperatures «' <", L" - L' = 0-70 (I* - <"). The heat supplied by oondensation per pound of steam condensed is L'. The weight of water deposited by the condensation of 1 pound of steam is by hvpothesis -, and (f — f) supplies heat to the amount = ' in cooling from t' to t". The ce heat abstracted by re-evaporation of this water is — . The balance of heat abstracted is therefore L"-^ + t" — L', which may be written in the form ^' ^^ -») _0'30(t!- Q ^^^^ ^^^ X X temperature of the wall is maintained by external agency, it will therefore continue to fall until (f—f) = — ^-^k"" • ^°' instance, the effect of 6 per cent, wetness is to lower the wall temperature until the range i' - t" is 160° F. mf n 4i OALLEMDAB AMD MIOOLBON OM THB [Minutes of Effect of Water m the Cylinder. — After the aconmti nation of water in the cylinder has oommenoed, it is not qnite so cl« at whether the effect of the water would be to increase or dimini'>h the total con- densation. It appears probable there would not be a great change. In a working cylinder, the water could not accumulate to any considerable thickness, except in special pockets. If the water were present in suflSoient quantity to be thrown into spray, and thoroughly mixed up with the steam, so as to expose a large surface to its action, the water so broken up would almost certainly be carried out of the cylinder with the steam; in proportion to the minuteness of its subdivision. The film left on the surface would, therefore, probably be very thin, and would not seriously affect the result, either in the direction of increase or diminution. To test this supposition, the law of condensation was applied to calculate the mean wall temperatures and the amounts of condensa- tion observed by Donkin in his " Revealer " experiments, in those oases in which water was probably present. The results thus calculated agreed with the observations within the limits of error of the measurements. Cn- a great change, itnalate to any If the water into spray, and expose a large ilmost certainly roportion to the ) surface would, seriously aJfeot or diminution, was applied to ats of condensa- iments, in those he results thus I limits of error from the above on in which the ited separately, also because the ihese conditions tion is complete, actually in this iometer in some surface, and to an of the sieam at or near the Kmdensation has ly calculated by r diagram of the le application of [rface. In oases ) not given, but estimate of the r, double-acting, 9 range by four IS jacketed with vould hardly be ; nearly twice as Tabu YIL— iLLmnunoir or Lmrrnra OoirDnrsAnoir. Trial of " ViUe do Douvret." » CJomponnd, 2,977 LHP. ; Feed per hour, 61,800 lbs. ; BiaTolnttoni per minute, 86 -8. I. Cylinder oonaideiad | 5. Point of stroke 8. Percentage of feed miaiing .... ... 4. Miacing quantity lbs. per hour 0. Mean of ateam tempentnie oyole .... ° F. 6. Bange „ „ ,,....„ 7. Bange of pretHura . . . . lbs. per iqnMe inch 8. Cleaianoe surface square feet 9. Condensation area ° seconds 10. „ ooneoted for metal cycle .... 11. Thermal units absorbed per square toot per minute . 12. Clearance condensation lbs. per hour 18. Banel surface exposed square feet 14. Mean condensation area, corrected 15. Barrel condensation lbs. per hour 16. Total „ » » » 17. Bemainder to be aoconnted for great as Uie condensation limit. Leakage of the high-pressure valves or piston appears probable. The remainder, line 17, in the case of the low-pressure cylinder, corresponds fairly well with the probable wetness of the steam due to adiabatio expansion com- bined with partial drying due to friction, re-evaporation and other causes. The examination of a number of Guch cases leads to the following general conclusions. In the high-pressure cylinder, it is not generally likely, owing to the probable dryness of the initial steam, and also to the greater probability of partially explosive evaporation, that the condensation should reach its limiting value. Nevertheless, in the great majority of oases, the missing quantity at ont-off in the high-pressure cylinder is much greater, often many times greater, than can ba aoconnted for on the supposition of limiting condemation, according to this method of analysis. To aooount for this, the oondenaation must either be proportional to the density of the steam, or Urn greater part of the missing quantity represents leakage. The former sappaaitum c mrf l ia ti with the Authors' experiments i^Jid those of Donkin, but is in agreement with those of some other observers. The latter supposi- tion would generally require rates of valve and piston leakage similar to those observed, and would appear the most natural explanation in the light of these experiments. > Proceedings of iae Institution of Mechanical Engineers, 1892, p. 108. •%.'ri '■^•.i:^:rjf^T^<^WJ^ 41 OAUiENDAB AND MIOOLBON ON THE [Minntea of In the low-pressure oylinaer of an nnjaoketed engpine, the con- densation may frequently have its limiting value, ovring to the initial wetness of the steam. It is also much more important than the leakage for two reasons. The condensation is greater, because the initial surface exposed is much larger. The leakage is less, because the di£ference of pressure on the valves and piston is much smaller. It is also evident that the perimeter of the ports and piston, upon which the leakage mainly depends, varies directly as the linear dimensions, whereas the surface exposed for condensation varies as the square of the diameter. Case of PartM Condensation or Complete Be-evaporation. — It is evident, from the smallness of the results obtained in the small single-acting engine at low speeds, and from many similar results obtained by other observers, that, in the case of the simple engine, when working at a moderate ratio of expansion, the initial con- densation is often very far below its limiting value. He-evapora- tion is probably complete on the clearance surfaces, either at release, or at a very early period in the exhaust, and the walls are probably dry for most of the return stroke. The case of complete re-evaporation, or of partial condensation, as it may be called to distinguish it from the special case of limiting con- densation, does not admit of the same simplicity of treatment as the limiting case. The temperature conditions are evidently far less stable, and the amount of cyclical condensation, which depends on the balance of heat loss and supply, is liable to be a£feoted to a much greater extent by the variations of the con- ditions of running, and by differences of type and arrangement in different engines. It would, therefore, be unsafe to attempt to apply the results deduced from any one engine, under special conditions, to any other engine. At the same time it is possible that some light may be thrown on a very complicated problem by the careful consideration of the results observed in a particular case. The amount of cyclical condensation would appear not to be greatly affected by a moderate variation of speed. Some increase ° is to be expected, both on account of the diminished mnge of the metal cycle and also on account of the greater convective action of the exhaust steam; but the former cause partly tends to compensate itself by producing a higher wall-temperature, and the latter depends so much on the initial pressure and on other conditions that it cannot be satisfactorily represented by a formula. Effect of Varying the Conditions of Bunning. — Assuming as a first approximation that the cyclical condensation is independent of the %.- [Minutes of ig^ne, the oon- , owing to the ore important ion is greater, The leakage TOb and piston imeter of the lepends, varies oe exposed for toration. — It is I in the small similar results simple engine, he initial oon- Ke-evapora- oes, either at and the walls The case of i, as it may be limiting con- f treatment as i evidently far isation, which is liable to be DS of the oun- rrangement in to attempt to under special it is possible cated problem in a particular tpear not to be Some increase ' d nvnge of the Lveotive action .rtly tends to uperature, and I and on other resented by a ming as a first pendent of the ProoeedingtJ LAW OF OOMDBiraATION OV BTBAV. m speed, it may most conveniently be expressed either in thermal units per minute or in pounds of water per hour. (a) Variation of Cut-off. — From observations made at one-half, one-third, and one-fifth cut-off, it is inferred that, when the engine , is working single-acting, non-condensing, the cyclical condensation measured per minute at cut-off is to a first approximation indepen- dent of the ratio of expansion. If the condensation is measured as a percentage of the indicated steam at cut-off (excluding cushion steam), this result is equivalent to the statement that the percentage condensed increases nearly in direct proportion to the ratio of expansion, defined as being the ratio of the volume occupied by the feed steam at cut-off to the volume of the cylin- der. It is not convenient to measure the condensation as a percentage, either of the whole cylinder contents or of the whole cylinder feed, because these involve cushion steam and leakage. In a double-acting engine, the condensation on the barrel surface is necessarily less, and the temperature of the clearance surfaces higher and less variable. The change of each term would therefore be less, and it would seem probable that a similar compensation would occur, leaving the condensation at cut-off unaltered by change in the ratio of expansion. The formula is very attractive on account of its simplicity, which is the first desideratum in a formula of this kind intended to cover roughly a variety of conditions. Its applicability to the case of the double- acting engine is not to be suggested were it not that it appears to represent very fairly many of the most reliable results. In a large class of engine trials, the effect of varying the ratio of expansion r, on the observed percentage t of the cylinder feed condensed at cut-off, is closely represented by the semi-empirical formula of Thurston, z = a ij r, where a is a constant depending on the other conditions. The numerical value of a for the engine would be 15. According to the Authors' formula z = -rr-— , where e has the ° 100 + cr numerical value 10 in the present case. It is remarkable that these two formulas, which are at first sight so totally dissimilar, should give results not differing by more than 2' 5 per cent. throughout the whole range, from three to twenty expansions. From <^^hree to one expansions, the Authors' foriuula would appear to be preferable, as that of Thurston generally gives results which are too high as compared with experiment. (b) Double- veraua Single-Acting. — It is possible to draw general conclusions from a consideration of the effect on the distribu- tion of temperature. In a double-acting trial there would be I -a 46 OAIiLINBAB AND mOOLSON ON TBM [MinniM of pnotioally no effect of oonveotion of heat by the piston. The gradient of longitadinal conduction would be halved at an early cut-off, and would practically disappear at a late cut-off. The effect would be to raise the temperature of the barrel portion of the admission surface very materially. The condensation on the piston would also be reduced, probably to less than one-half. Against these reductions the effect of the piston-rod at the crank end, and of conduction of heat to the framework of the engine, are to be set, which would tend to lower the temperature at that end as compared with the back end of the cylinder. It is evident that the condensation reckoned as a percentage of the steam would be considerably reduced. For the engine under review, the reduction is estimated at 80 per cent.; that is to say, the total condensation per minute, instead of being doubled, would be increased by about 40 per cent as an outside estimate. In the case of large engines, the effect of conduction being negligible, the percentage saving by double action would be less. (e) Variation of JniUai Preuwe. — The initial pressure was not sufficiently varied to give any direct information on this point, but experiments show indirectly that the effect is much more complicated than might be supposed. The external loss of heat from the cylinder would be increased nearly in proportion to tho increase in the difference of temperature from the surroundings. The internal loss by re-evaporation and by the exhaust steam might be modified in a very different way. The experiments would appear to indicate, as has been suggested by Eirsch, that re-evaporation from the more highly-heated portions of the walls is of an explosive character; that is to say, that a portion of the water-film is blown off the walls without abstracting a full equivalent of its latent heat of vaporization. Condensing at atraospherio pressure with an initial steam temperature of 830° F. to 325° F., the temperature of the cover was over 800° F., and the platinum thermometer in the cover appeared to show that re- evaporation was complete almost as soon as the indicated tempera- tare fell below this point. This observation at once suggested the partially explosive character of the evaporation as a possible explanation of the high temperature attained by the cover. The same explanation probably applies to a less extent to the hotter parts of the barrel surface, which appeared to have been gaining much more heat by condensation than they were losing by re-evaporation. On the cooler parts the balance of heat supplied would be simply that due to the small difference in total heat between the steam condensed at a higher and evaporated at a [HinatM of I {rfaton. The Bd at an early ont-off. The rel portion of ksation on the ^han one-half. i at the orank )f the engine, iratore at that It is evident of the steam er review, the say, the total ed, would be mate. In the aeglig^ble, the Msnre was not Dn this point, s much more A loss of heat >portion to the surroundings. >xhaust steam ) experiments y Eirsoh, that B of the walls i a portion of jraoting a full Condensing at ure of 830° F. )0° F., and the show that re- sated tempera- nce suggested i as a possible le cover. The to the hotter been gaining are losing by heat supplied in total heat bporated at a ! Prooeedinga.] LAW OF OONDBKSATION OF BTIAM. 47 lower temperature. At ou<:)-fifth cut-off the evaporation apparently ceased to be explosive at a temperature between 270° F. and 260° F. The steepness of the temperature gpradient along the sides of the cylinder cannot otherwiae be satisfactorily accounted for. If the possibility of explosive evaporation at higher temperatures is admitted, depending partly on the diminished surface tension of the water and partly on the greater density of the steam, it is dear that the condensation may not necessarily increase con- tinuously with increase of initial pressure. This result was arrived at independently of Eirsoh, from the evidence of the observations. Kirsch makes the suggestion, not from direct experiment, but as a possible explanation of the smallness of the condensation observed in practice as compared with that which would theoretically be required, supposing that the surface of the walls were raised to the temperature of the steam, on the usual assumption that the rate of condensation of steam is practically infinite. Without further evidence, it would not be fair to con- clude that re-evaporation at higher temperatures is always of this character. The conditions of the hole in the cover are not quite the same as those of the plane surface; but the observations suggest a possibility which obviously requires consideration. On one occasion a curious eifeot was accidentally obtained as an illus- tration of the possible consequences of explosive re-evaporation. When the electric-lighting engine was unexpectedly shut off, the boiler-pressure rapidly rose nearly 10 lbs. above the usual limit. This produced a rise of temperature of 6° F. at the back end of the cylinder, but the temperature at the middle of the cylinder rose more than 16° F., from 264° F. to nearly 280° F. The other con- ditions of running were unohanged with the exception of a slight increase of speed. As a general rule the changes of temperature at this point of the side were less than those at the back end of the cylinder. It is difficult to account for this abnormal rise of tem- perature, except by supposing that the re-evaporation at this point ceased to be normal, and became partially explosive. It will be observed that the change of temperature took place at that part of the scale which for other reasons appears to be the critical point in a non-condensing engine. (d) Variation of Wetneu. — It has long been recognized that the presence of water in the cylinder or of priming in the steam, must have the effect of increasing condensation. In a small single- acting engine, at low speeds, and without special precautions as to lagging the cylinder or drying the steam, practically conclusive evidence was obtained that the action of the metal alone wiM -fl 48 OALLKKTVAR AND KIOOLSOIT 05 TBI [MinntM of ' competent to prodnoe the obserred effeota. At the mme time an illustration waa fnrnished of the abstraction of heat by the orm- densation and re-evaporation of wet ateam, which shows that initial wetness of the steam in probably one of the most power* fnl factors in increasing cylinder condensation. Proyided that the initial wetness is small, and that the lowering of the wall- temperatnre produced by it is not sufficient to greatly change the other conditions upon which the balance of heat depends, it ia possible to represent the effect in a simple manner by combining the formula already given with the law of condensa- tion. If t — f he the mean difference of temperature between the ateam and the walla dnring condensation when the steam is dry, the balance of heat supplied by condensation and evaporation ia approximately • 60 (<' — <") W per hour, where W is the weight condensed. For a amall lowering of wall-temperature, the balance of heat required would not be greatly reduced, and the condensa- tion would be increaaed nearly in the aame proportion aa the difference of temperature < — f. Upon these assumptions, it ia found that, for initial steam of a percentage dryness lOOx, the initial . mdenaation, as compared with that due to dry steam, is increased by a percentage given by the formula .); — ~. If the temperature difference for dry steam is 80 F°. and L = 900, the effect of 1 per cent, of wetness would be to increase the initial condensation by 25 per cent. In the majority of partiftl condensa- tion oydea, the errora involved in the above aaanmpi.Dna are of auch a nature aa to make thia formula hold through a aomewhat wider range than would otherwiae be the caae, but it cannot be trusted beyond 60 per cent, incieaae, and ahould be regarded, in any caae, aa showing rather the general nature of the effect than its absolute magnitude. (e) Variaiton of Back Preuure. — For a given initial preeaure, the wetnesa of the exhaust ateam, due to adiabatic expansion, will depend on the back preaanre. The cooling of the internal aurfaces during exhauat, apart from re-evaporation, will depend on the wetneea of the exhaust steam qnf te as much aa on its temperature. From the reporta of triala in which the jacketed aurfiMses of cylinders, valve-oheata, and receivers are given, it ia poaaible to eatimate that the rate of abstraction of heat by wet ateam in motion under such conditions does not probably ever exceed 1 T.U. per square foot per minute per 1° F. difference of temper- ature, and may be very considerably less. If the wetness per cubic foot, and not per pound, of the steam, ia considered, it would , » [MinntM of I Rame tima an it by the orm- ih shows that e most power- ProTided that ; of the wall- reatly change heat depends, manner by r of oondensa- e between the steam is dry, ivaporation is is the weight re, the balance the oondensa- ortion as the mptions, it is ess 100 x, the dry steam, is 00L(1 - a>) l-2(<'- f)- and L = 900, ase the initial ■ti»l oondensa- ipi.Dns are of 1 a somewhat it cannot be e regarded, in effect than its itial preranre, cpansion, will emal surfaces spend on the I temperatnre. I surfaces of is possible to ret steam in ▼er exceed 1 » of temper- wetness per ired, it wonld ProMedfagt.] law 0I> OOlTDIiraATIOir Or Rli.!!. 49 appear probable that the oooliug effect of the ezhaost steam in a condensing engine may often be actually less than in a non-con- densing, bnt that on the average there is no decided difference. (/) Effect of 0(mpr«$$ioH. — If there were no interchange of heat between the walls and the steam, compression to the initial pressure would restore the cushion steam to its initial state. The volume at compression in the present case was one-third of the volume of the cylinder and clearance, and probably in- cluded at least two-thirds of the heat abstracted by the exhaust steam. This would explain the very considerable superheating of the steam observed during compression olose to the walls. The effect of an early compression may thus be regarded aa equivalent to a considerable reduction of the loss of heat due to the exhaust steam, which is probably, next to initial, wetness, the most potent factor in abstracting heat from the walls. The effect of an early release is probably similar to that of an early compression in reducing condensation, though it acts in a different way. In the case of partial condensation an early release allows lees time for the condensation of wet steam on the colder parts of the walls towards the end of the stroke, and more time for the walls to dry before the return stroke of the piston, a condition probably unfavourable to piston leakage. It is possible that an early release may diminish the condensation and the exhaust waste sufficiently to more than compensate for the loss of area on the indicator diagram. In the cases of limiting con- densation, on the contrary, the effect of an early release may be to increase the condensation, and to lower the temperature of the walls by increasing the available evaporation and condensation areas. Effe The reduction of oondenaation in the caae in which » = H - t" amounts to about 33 per cent instead of 41 per cent, as giren by the above formula. The difference of tempera- ture between the steam and 'he clearance surface is diminished approximately in the same pr ^portion. Besides diminishing the initial oondonsation by raising the temperature of the wall-surface, the use of superheated steam diminishes the wetness during expansion, and therefore consider- ably reduces the exhaust waste and the abstraction of heat by the condensation of wet ateam towarda the end of the stroke. It ia also probable that it may tend to diminish leakage. Effect of Jacketing. — The effect of jacketing a C7Under with steam at boiler-pressure, is to raise tne temperature of the jacketed walls very nearly to that of the boiler if thb jackets are working properly. According to the law here proposed, the condensation on the jacketed surfaces would be practically negligible, and the clearance surfaces are by far the most important In large engines it would consequently be of little use to jacket the sides, but in small engines the clearance surfaces would also be heated by conduction so as to be practically jacketed. It is not improbable that a part of the economy due to jacketing, especially in small engines, is owing to the reduction of leakage. The valves and valve-seats become so heated by oonduotion that the possible water-leakage is minimised. From the same point of view, the drying of the steam in jacketed receivers must have a beneficial tendency, as thero is then less water avftilable to cool the valve- surfaces by re-evaporation in the exhaust. Variation of Size and Surfa^x. — The effect of variation of surface exposed, and particularly of the extent of the clearance surface, is probably different according as the condensation is of the partial or limiting type. If the main factor in the abstraction of heat from the clearance surface is the condensation of wet initial steam, as is probably the case when the condensation haa its limiting value, the amount of heat abstracted and the initial condensation will vary simply as the surface exposed. In thia case it is of [MinntM of diflTerenoe of aoe when the lyole, the per- eoted by the he ezpreosion ding approxi- ii not greater n the case in tead of 41 per le of tempera- i« diminished J railing the rheated steam tfore consider- of heat by the stroke. It is C7Hnder with >f the jacketed ts are working ) condensation legligible, and »nt. In large 3ket the sides, eiIbo be heated act improbable jially in small 'he valves and t the possible t of view, the ive a beneficial cool the valve- tion of surface irance surface, B of the partial raotion of heat t initial steam, la its limiting il condensation B case it is of ■2i fjjf K %■ . PraoMdlmts.] LAW OF OOHSIMIATION OF BTKAM. H primary importance to know the foil extent of the clearance surface on which the greater part of the condensation takeu place. In order to rednoe the Icid as far as possible, the extent of the clearance surface should be reduced to a minimum. If, on the other hand, the initial iteam is dry, and the condensation is uf the partial type, the extent of the clearance surface is a matter of much leofl consequence, because increase of surface has the effect of raising the temperature. It ib probably best, in the case of partial condensation, to neglect differences of clearance surface in comparing different engines, and to take the clearance surface at each end of the cylinder as being ir d\ where d is the diameter. The actual clearance surface cannot be more than 60 per cent, less than this, and is saldom more than 5 J per cent, greater. For all practical purposes the equivalent clearance surface forms a suffi- cient basis of comparison. But the barrel surface exposed up to cut-off, allowing for the time of exposure, may be very simply represented, if desired, by the addition of the term Ide, where o is the cut-off fraction, and / the stroke. The state of the surfaces is not important if the rate of condensation of steam is regarded as the main factor in limiting the amount of heat absorbed. The presence of a thin film f* grease or rust may make the cycles observed at a given doj ih in the metal of smaller range, but will not really make much difference in the amount of con- densation, unless the film is so thick and o jstractive as to greatly increase the surface range of temperature, which is probably seldom the case. Effect of Chnduction. — In two similar cylinders of different linear dimensions, but with the same distribution of temperature, the loss of heat from the admission surface by conduction, will be pro- portional to the thickness of the metal. The initial loss due to conduction, reckoned .as a percentage of the steam, for cylinders of the same thickness, will vary inversely as the cube of the linear dimensions. It is necessary to exercise caution in applying the results deduced from small engines to large. In order to estimate the probable effect of conduction, the temperatures have been compared with those obtained by Donkin with a cylinder 6 inches in diameter, and 8 inches stroke. The Authors conclude that the effects of conduction in their engine are not to be neglected as compared with larger machines, but probably do not amount to more than five or ten per cent, and are not such as to seriously vitiate the general nature of their conclusions. In cylinders of different shapes, under similar conditions of running, the loss due to oonduotiou at cut-off reckoned as a percentage of the in- u 52 OALLBNDAB AND NIOOLBOM ON TBM [tfisntoiof dioated steam, would vary as j^, where I is the thiokness of the metal, d the diameter, and I the length of the cylinder. For instance, the effect of oondnotion in the cylinder above mentioned, wonld be nearly five times as grea^ as in that of the Bobb engine. Formuhe of Condentation and Leakcqe. — From the foregoing oon- siderations it will be evident that, vhen the required data are available, or when it is possible to observe the temperature distri- bution, no formula can be regarded as being at all satisfaotory. It is, neve'.'theless, convenient to have a simple approximate formula for the purpose of making rough estimates and comparisons, and also as exhibiting the results of the investigation in a brief and compact form. In the case of limiting condensation, when neither the oards nor the extent of the clearance surface are given, the limiting value of thr condensation W at cut-off, expressed in pounds per hour, ma^ be estimated from the formula, W = nd? {t - f), - where d is measured in feet and ( in degrees Fahr. This for- mula assumes that the clearance sur&oe 'j v d^ but makes an allowance of 20 per cent, for the barrel surfoce. It also assumes that the cut-off is at or near mid-stroke, and that the drop of pressure during admission is small. The factor {f -' f) is supposed to represent the total range of the steam. The latent heat of the steam is taken as 900. The result, thus estimated, .•pay be corrected by eu^istituting the proper value of the latent heat in each case, and may then be reduced in the proportion }. , "^ ; , (.8 + V«) to allow for the effect of the probable range of the metal cycle. The case of partial condensation may be roughly represented by the formula, W = C X S = S («' - i°). where is the condensation in pounds per hour at out-off per square foot of total equivalent clearance sur&oe S of the cylinder, supposed unjaoketed. The condensation factor is a function of the initial and exhaust temperatures and of the external con- ditions, but may be taken as being approximately independent of the speed and the ratio of expansion. Apart from superheating or jacketing, the value of is probably most affected by ths degree of compression. The condensation factor may ako be inteipretedas the mean difference of temperature {( — <°), betwoMi the walls and the admission steam, rtw^ticri to one-half cut-off. Ai ¥ ' iLi'Wii: CMisntMcf liokness of the oylinder. For ove mentioned, e Bobb engine. foregoing oon- [nired data are peratare distri- atisfaotory. It dmate fomrala •mparisons, and in a brief and 1, when neither are given, the r, expressed in a, ihr. This for- but makes an !t also assmues at the drop of •r ((C - f ) is n. 'Che latent thns estimated, ' the latent heat .. (1 + » metal cycle, represented by ■ at ont-off per if the cylinder, is a function of I external oon- indcpendent of 01 superheating effected by ths may also be -• <°), betwcMi ne-half cut-off. :'^§. Prooeedingi,] LAW 07 OONDBNBATIOM Or 8TBAX. H The effect of sice and surfiuse, and of double or single action, may be supposed included in the expression for the surface 8. This factor should be taken as 2 vcP double-acting, and as ird' single-acting. The effect of jacketing in large engines may be represented by simply omitting the jacketed area from the factor S ; but the effect of conduction in small engines cannot be satisfactorily included in the formula. If W is the weight of feed in pounds per hour, aooounted for by the indicator at cut-off, and W* the total missing quantity per hour, W -I- W° represents the total cylinder fee^. If the oondensation W can be estimated, either by a formula or by observing the temperaiuio ulitribution, the remainder W° — W, which may conveniently be reprewutAd by the symbol W, may be most probably attributed to leakage. According to the Authors' experiments, the main part of the leakage at any point of the stroke may be represented by the formula, when (p' — p") represents the difference of pressure between the valve-ohest and the exhaust. If the factor L cannot be measured, it may be estimated by the method of p. 33, from the mean over- lap and the perimeter of the ports. Since the area of the steam- ports is generally designed to vary as d' N, where N is the piston speed, L may be taken as being proportional to d ^N in simiL\r engines. Nluttraiiont o/ Partiai Condensation. — As a test of the validity of the method of reducing to the equivalent clearance surface in the case of paitial condensation, and as showing the relative un- importance of the barrel surface, the following cafles are cited, having been selected by Ootterill (loc. cU., p. 334) for a similar Tabli Vin.— CoHDKroNO Tbiau. stroke, Dia- meter, Revo- i,»'jni H^te. CntHiit, e. AtMO- lute Preemire at Cut- off. Mlnbig Qiuntttjr. Equiva- teut Cleer- ince Surface, 8.i Oonden- AnthoritiM mi . TiUllUrk. J-ta-per Hour, W". Per Oeot.of Feed. aation Faouir, C. Mftir . .(L) .. . .(M) Feet. 6-6 5-6 Feet. 2-67 2-67 20-8 20-8 0-26 0-26 46 46 770 1,200 29 87 62 o 28 23 Dallas . (D) 2-6 1 8-0 56-9 0-20 47 1,460 29 60 24 Him and (2) Hallauer (S) (HH.). . (6) 6-5 6-5 6-5 20 20 20 80-6 800 80-4 0-26 016 0-16 54 55 55 850 460 760 80 26 86 82 / Supe \ 8V 29 S7 rheat, 'F. 26 Calculated by formula, 8 = 2(ncP + lde). if "s4 ■m ^t^ :§ ^0m 64 OALLENDAB AKB NIOOLSOM 6N THB [MtautMof purpose. There is considerable range of speed and of relation of stroke to diameter, but the oonditions of pressnre are fairly comparable, and the leakage similar, and probably small as com- pared with the condensatioiL. The engines are also sufficiently large to make the effect of condnotion practically negligible. In the trial Mair (L),» the sides ar J base of the cylinder -were jacketed, which would have the effect of reducing the unjaoketed olsarance surface by about one-third. The condensation constants given by Cotteriirfor thnse three oases are, (M) 5-3, (D) 7*0, (HH) 3*4 This wide range of values is probably to be explained by his taking the condenaation as proportional to the barrel sur- face. The agreement of the values of C in Table IX is as close as can be expected, and may be taken ap showing that, in the case of partial condensation, the equivalent clearance surface ia the better basis of comparison. Tabu IX.— XoR-Ck)MDiNBnra Tbials. AnlhorlUea and TrUlUark. C ard N (4-6) Willans . . Simple 2^2 . O.E. 1898, p. 174 Willana . . 90 Compound ^~^, Loe. eit. . . i 3 a I Feet. Feet. 1-0 0-88 74 0'20 0-6 117| 110 0-44 O-b' 1-17| 20i' 0-44 08 lir 408 0-44 Mining Quan* Uty. Hour W. 96 Gkktely and . Kletzsoh (Thurston) . (16) (17) (18) Col. English . Mech. Eng., 1887 P1.90, Fig. 5 . J. W. Hill, I BO. (Peabody, } H 0. p. 265). I Wbk. 0-5 0*83 ?22 0-60 0*5 0-88 21l! 0-60 0-5 0-88 40l! 0-60 35 1-5 I 68 0-41 8-5 1-8 69 0-42 8-5 1-8 69 0-40 1-5, 1*83 40, 0-80 40 4 1-5 18 40, 1-3 I 78 16 76 14 76 0-17 0-25 86° 88' 28° 83 82 80 IT 13° -T-8' 65 SO 40 48 87 187 88 64 88 24 Bq.Ft. 2-6 20 18 6 I 84 i 000 890 800 372 CotidenMtion Facton. Tt 4-6 4-6 4-6 2-4 2-4 2-4 470 99 : 0- 08 1,040 100 : 012:1, 100 91 I 0-OS,l,124 48 18-8 18-6 18-6 12*2 5i 170 2-8|19'' 2 •8,20° 2-8 80° 1-8 28= 1-828° 1-8 16- 1 160 16-2 16° 21° 48" 20° 80° 66° 69° 70° MinutoB of Proceedings Inst. C.E., 1884, vol. Uxlx. p. 840. [MinntMof of relation of re are fairly imsdl as com- 10 snffioiently eligible, jylinder ^ere he UDJacketed don oonstanta 6-3, (D) 7-0, ) be explained he barrel ror- X is as close ; that, in the Lce surface is CondenMtioo Facton. Si Bq.Ft. 2-6 Tt 4-6 4-6 4-6 2-4 2-4 2-4 16 18-6 18-5 18-5 12-2 H ■3 -/J 170 2-819"' 2-8J20'' 2-8J80° 1-8 28= 1-828° 1-8 21° 48" 20° 16- 1 16-0 16-2 16° 89° 66° 69° 70° p. 340. Pnweediogi.] LAW OV OONDKNBATION Or 8IKAM. 55 Table IX illustrates, the fact that the resnlta obtained by the thermoelectric method, even when reduced to the equivalent clearance surface, are smaller than any similar results, except perhaps a few of Willans*. The effect of an early compression is also wtiU marked. In the last three oases the steam is recorded as having been occasionally 5 per cent, wet, and the condensation was probably limiting. Iktimation of Leakage. — In the case of partial condensation, it is evident that the result may be so profoundly affected by differenoect of wetness or of compression, that it would not be justifiable to take the small result found in the case of the Authors' engine, and to assume that the larger results shown by other engines were to be entirely attributed to leakage. With the Bobb engine, running double-acting under the same conditions, making allow- ance for piston oonveotion, etc., tho condensation factor ought not to exceed 13 lbs. or 14 lbs. per square foot of equivalent clearance surface pe; hour. This result is much smaller than that shown by any of the double-acting cases cited. Although this value of the condensation factor cannot be directly applied to the estimation of leakage in other engines, the formula may be of use in 00m- paring different trials of the same engine, or similar trials of different engines, under strictly comparable conditions. In the case of limiting condensation, if the clearance surface and the steam cycle are known, a definite limit to the condensation is at once afforded. Any excess may be reasonably attributed to leakage. Unless the steam is known to be wet, the temperature of the walls should be observed, because it cannot otherwise be certain that the condensation is really limiting. If it happens to be partial the leakage will be under-estimated. If = 20 is taken as being a probable average value of the condensation factor in the case of simple engines, the total loss would be given by a formula of the type io • W°=40ird»-J. which would make the leakage relatively more important in small engines and at high piston speeds and pressures. Prof. 0. A. Smith, using a formula of the type W = Q'd {( - f) to represent the total losses in simple unjaoketed engines, measured in lbs. per hour per degree of steam range and per foot of piston diameter, finds values of C which vary between 1-85 and 4*72. The same series of trials may be represented on the Authors* formula by values of from 15 to 25, and valnek of the leakage oonstant from 5 to 10. This' does not appear, to be an unlikely 66 OALLBHDAB AMD NIOOLBON ON THB [Minnies of I- range of variation, but unless the condensation is often limiting in simple engines, the leakage must frequently be the greater. The data required for the exact application of the propo.Md method of estimating leakage are not generally available in any extant trials ; but to exemplify the general conclusions to be derived from the analysis, the triple-expansion trials of the experimental engines at Owens College, described by Osborne Beynolds,* may be selected, not only on aooount of the unimpeachable accuracy of the observations, but also because these engines would appear to have achieved some of the best performances on record for engines of so small a size. If it can be shown that, even in these engines, in spite of their record performance, a large part of the missing quantity is probably to be credited to leakage, it follows, a fortiori, that a similar cause may be suspected in more ordinary oases. These trials possess the farther advantage of an unusually complete and iinreserved record of all the leakage tests which were applied, and of the various minor leakages which were discovered and rectified from time to time. The trials 41, 35, and 40, in which the receivers were jacketed but not the cylinders, will be taken as an illustration. The effect was probably to dry the steam completely at the low speed for the low-pressure cylinders, but less completely at the higher speeds. The following are the data for feed and receiver-jacket condensation in the three trials : — Gtuhskb Jackets Bhptt. Bichtib Jaokbts at Boilib PBxssuBn. Triftl nnmber Boiler pressure, absolute . . . Feed per hoar (bot well) . . . Jacket condensation, lbs. per hour . The following are i he data for the three separate engines :- - Xngioe Norabwr I. U. m. Trial nnraber. . . . Expansions, r . . . Bevs. per minnte, n . . Pressure range, p' — p" . Temperature range, t'-t"\ degrees/ 41 2-7 146 187 83 8S 2-3 229 182 78 40 20 822 128 72 41 2-4 127 4S 68 85 2-4 216 50 70 40 22 820 52 69 41 2-7 109, 20 128 35 30 184 21 116 40 2-6 276 28 107 Hissing quantity. t° \ per cent./ 40 29 22 41 88 80 51 48 82 Missing quantity, wP \ lbs. per hour/ 185 200 282 100 262 817 287 882 887 Condensation limit . . 43 41 89 88 88 88 810 800 285 ' MJcntes of PfooMdings Inst. C.E., to], loix. p. 152. [Minatesof >ften limiting e greater, the proposed liable in any i to be derived experimental lolds,* may be xjnraoy of the ppear to have engines of so §^ne8, in spite ling quantity 'ortiori, that a oases. These complete and ) applied, and [ and rectified nrere jacketed >. The effect [ow speed for ,t the higher eoeiver-jaoket B Pbxsbubb. A )5 40 )6 201 38 1,055 17 158 ingines : in. 41 2-7 109, 20 85 80 184 21 40 2-6 276 28 12S 115 107 81 48 82 287 882 887 810 800 285 Pnweedinga.] LAW Of OONDKNBAIIOH Or BHAM. 67 52. The condensation limit given in the last line, in default of the cards, and of the actual area of the clearance surface, is estimated by the method explained. No allowance is made for the variation in the ratio of expansion, but since the three engines wore similar, the vrlues estimated may probably be taken as showing the relative order of magnitude of the condensation-limit to be expected in the three cylinders. Taking engine Na III, it would appear probable that at the lowest speed the steam vras practically dry, and the condensation consequently partial. At the higher speeds, the drying in the receivers was probably insuflSoient, and the con- densation had nearly, if not quite, reached its limit. In the case of engine Ko. I, the condensation-limit is obviously insufficient to explain the missing quantity. It is likely that both the condensa- tion and the leakage increased slightly with increase of speed, but the leak appears to have been much more important than the condensation. The rate of leak required would have been between 1 ' lb. and 1 * 6 lb. per pound pressure per hour. A leak, roughly estimated at 30 lbs. per hour, was observed on one occasion with this valve standing. The rate of leak required for the valve when running would appear not unlikely, consideri^ig the type of the valve, and that it was specially designed for a high speed. Part of the leak may have been due to the piston and to the cut-off valve, but the pistons appear from other evidence to have been fairly tight. The valve was probably mainly responsible. In the case of engine No. IT, the same observations are applicable. The rate of leak required would have been between 3 lbs. and 4 lbs. This would appear somewhat excessive, but it is recorded that leakage of the cut-off valve was inferred from the cards in this engine, and was rectified at a subsequent date. It is evident that a leak of the kind described as having been discovered in this valve might be expected to- increase considerably with the speed, as the amount of leak would depend on the inertia. The rate of leak in engine No. Ill may have been upwards of 2 lbs., but is evidently much less important than the oonaensation, owing to the smaller pressure difference on the valve. If the leakage therefore is rejected, the rate of condensation of steam must increase rapidly with the density. Setting aside the experiments described in the present Paper, this would appear a perfectly tenable hypothesis. Unless, however, the surface of the cylinder has been under- estimated, the metal in engine No. I seems hardly capable of aocountio'; for the whole missing quantity at the lower speed, even if Ine temperature range of the surface of the metal were the same as that of the steam; at least, on any reasonable theory ^?J *%a%| is" 68 OALLBNDAB AND mOOLSOM ON THB [MinntMof of oondensation. The inferences drawn from the above case may be taken as typical of a great number of other oases which might be g^von. Direct experiments by Colonel English* on initial oondensation with a portable engine appeared to show a rate of condensation varying roughly as the density of the steam. The method employed was similar to that nsed by the Authors for meaanring the exhaust leak, except that steam was admitted at each revdln- tion to the clearance space at the back end of the cylinder, and that the sides of the cylinder were jacketed. The speed and pressure were considerably varied, but it was assumed in reducing the observations that the condensation per hour varied as the square root of the speed. The effect observed when measured per hour appears to be much more nearly independent of the speed. It also appears to be nearly a linear function of the pressure difference. At a mean density of 0*15 the results observed by Colonel English would correspond to the condensation of about 40 lbs. per square foot per hour, which is nearly four tiraf« the value found by the Authors at a higher density in an unjacketed cylinder. These results may be more naturally explained by supposing a comparatively moderate condensation varying as the temperature difference, combined with a muoh larger leakage varying as the pressure difference. The rate of leak thus required would be about 1 'C lb. per pound pressure per hour in the non-condensing experi- ments, and a slightly smaller rate in the condensing experiments. It may be observed that this rate of leak is the same as that found in one of the Authors' valves, which was proved to be absolutely steam-tight when stationary. It does not imply that the valve was in bad condition, or that the engine was at fault in any way. Expression in Terms of the Steam and Feed. — It is frequeuJy convenient to express the condensation and leakage in terms of the indicated feed and also of the total feed. The percentages of indicated feed being d-signated by y, and the percentages of total feed by z, the following notation is afforded : — 100 W «° = a° = W 100 w° , y= 100 w "o™:.,"^^^,.. w «'= 100 w ,. «"= 100 w (w°-f w)' " {w+wy ~ (w°-i-W)' ;. »°=«'-f-«". ^. »' = N. »" = (100 -I- y°y (100 -I- yy - (100 -|- y")' » (100 ^°)" > Prooeedings of (he Inititutian of Ueohanioal Engineen, 1887, p. 506. [MinatMof wve oase may I which might 1 oondenaation ' condensation The method for meaanring t each revdlii- cylinder, and le speed and d in reducing raried as the len measured t of the speed. ' the pressure I observed by tion of about im<« the value :eted cylinder, r supposing a e temperature rarying as the rould be about insing experi- l experiments, as that found be absolutely bat the valve ; in any way. is frequeuJy ;e in terms of percentages of itages of total 1887, p. 506. Trooeedlaga.] LAW OF OONDBNSATION OV 8TBAV. 69 If y is the piston displacement per revolution, n the levolntions per minute, w the density of the steam at ont-off ia lbs. per cubio foot, and r the ratio of the volume of the feed steam at cut-off to the piston displacement, the formula for the indicated feed in Ibn. *^^L!i?. Ah»W'=OS = ?^,W''=LO/-/). r 100 per hour is W ■■ So that. ,^. i^?^. ,-= ^Z'!;K'^"^' The pi.ton displace- ment Y is to be taken as —7— single-aotinj-, and as -^ double- 8 4 acting. The value of the ratio ^^ is therefore j either single- acting or donble-ftoting. If the rate of leakage L is taken as proportional to the product of the diameter of the cylinder and the square root of the normal piston speed N, L = L' d V N, where L' may be called the leakage constant, and has values which probably vary between 0*20 and 0*05 in engines of different types. The formula for the leak percentage y" may then be written in the simpler form, y". T II _ ____ , where L" is a new constant, which is avN practically proportional to L' in all oases which occur in practice of engines running at their normal piston speed. This shows that the effect of leakage on the performance of different engines may be expected to diminish in proportion as the diameter and the square root of the piston speed are increased. If, however, an engine running at its normal speed is being compared with the Ii" r a/N same engine running below its normal speed, y" = - ^ , , which shov/s that the effect of leakage on the performance will increase in direct proportion ais the speed is diminished. The formula for the condensation percentage y' may similarly be written, V = ^— , where C = —5-, and N is the piston speed, normal or otherwise. Comparing this formula with the leakage formula, the effect of condensation on the performance is not diminished by increase of diameter, but is much more affected by increase of piston speed; it is also considerably diminished by increase of initial pressure, which is not the case with leakage. The formula for the total percentage loss due to both causes is — y''=y' + y"= 2!^ 4. Kl. ^^■i 60 OAUBNDAB AMD mOOLSOM ON THB [MlnntM of ■ K: It 18 interesting to oompare this formula with that of OotteriJl. 0" loir r y° = " djH ~' ^^^*^ " intended to represent the resnlta of experiment on the an'ramption of negligible leakage. As an illustratioa of the nature of the oonsequenoes involved in the assumption thit the rate of oondensation of steam varies as the density, the triple expansion trials already cited may be taken. Prom Cotterill's formula the values of the oondensation constant C" for the three engines at the lowest speeds, are:— No. I, 3*4; No. n, 6-0; No. Ill, 11-0. As the s^oam was probably fairly dried by the receivers .-f tbs speed, it is difficult to ee why the value of th? condensation ^nstanv should be so very different for thiM engines of a similar type and speed, unless '.he rate of ocndensa-A^jn of steam does not vary is the initial density. S'jmviary of Conclun(m*.~The observed wall-temperature cycles bhow that the range of surfa/je temperature, and the interohange of heat between the walls a7jd the steam, is determined chiefly by the temperature of the walls in each case, and by the finite rate of condensation of uteaui, 7^r this finite rate of condensa- tion, the value 0'V4 B,T,U. per square foot per second per 1° P. difference of temperature, at eOO° P., is obtained, a result approxi- mately equival'int to 2 ■ 7 lbs. of steam condensed per square foot per degree per hoar. The form of the wall-cycles, and other evidence show that the law of re-evaporation is ttw same as that of condensation, and that both are probably independent of the pressure. The amount of oondensation in any cylinder can, therefore, be deduced by the observation of the distribution of wall-temperature while the engino is running. Prom the form of the law of condensation there appears to be for any cycle a limit of condensation when the temperature of the walls is the time average of the steam temperature-cycle. Under this condition of "limiting" condensation, the re-evaporation is incomplete, and the tempera'uure of the cylind*" is maintained by the mechanical rejection of condensed water, so that it cannot fall much below this point. The oondensation observe-^, and the initial steam wa»^ piolubly uniformly dry. But a marked effect due to the co idensation of wet stean was observed, which bads to the inference that, owing to the abstraction of heat by the subsequent ro-evapr ration of the dej^osited wetness, the condensa- tion must always be limiting in caHbJ where the initial wf inetu of the steam is considetableT flffiPWI" [HiontMof t of Ootterin, lie TMulta of 98 involved in I varies as the tay be taken, ttion constant —No. I, 3-4; ■obably fairly ' ee why the ^ery different M» '.he rale of isity. irature cycles 3 interohangn mined chiefly b^ the finite of condensa- ond per 1° P. Miult approxi- juare foot per how that the tion, and that le amount of iuoed by the e while the appears to be )ratnre of the yole. Under B-evaporation s maintained lat it cannot nitic)^ va'.**f, 3nt a mbb9 condensa- al wf inetiij of I'raoMdiaga.] LAW OV 00NOIM8ATIOH OV STliK. 61 The case of " partial," w opposed to limiting condensation, is probably more common in ^mple engines. The amount of initial condensation per hour in this case appears to be nearly independent of the speed and of the ratio of expansion, and to vary little with the initial and exhaust teraperatuias. Simple approximate ex- pressions are given, deduced Arom the law of condensation, for the effect of initial wetness or of superheating of the steam, whi^h are probably, together with the degree of oompression, the most important factors in determining the result. Illustrations have been given of the method by which, if the requisite data are available, the condensation a,t any point of the cycle may be correctly computed by means of the condensation areas on the temperature-cycle diagram. This method requires a knowledge of the extent of the clearance onrfaoe, and also, in the case of partial condensation, of the mean tempeiuture of the surface, in addition to the steam cycle. The application of this method may be expected to throw light on other causes of loss, and particularly on the amount of leakage under the actual conditions of running, which, from these experiments, appear to be a much more ioiportant source of loss than is generally admitted. The thanks of the Authors are due to Mr. J. J. Guest, Assistant Professor of Mechanical Engineering, and to Mr. A. W. Duff, Demonstrator of Mechanical Engineering, for assisuince in pre- paring the figures for this paper, and in taking the observations for the measurement of the conductivity of cast iron. Valuable assistance has also been rendered by Mr. H. M. Tory, M.A., and Mr. H. T. Barnes, M.A. So., Demonstrators of Physios, and by Messrs. MaoDougall, Butherford, and Laurie, students of Applied Science. The Paper is accompanied by twenty-four drawings, from which Plate 6 and the Figi. in the text have been prepared. mc [Appendix. rf,t ml 62 OAXXmrDAB AHD XIOOiaON OK man OOHDmriATIOX. [lUantMoC APPENDIX. VmnoATiON or thi TnmnuTCu Oarmioir fob Lmrniro OommATioil. Ai K further oonfirmftUon of the Uw of oondeniktion and rB-«T»por»tion pro- poaed in the preeent P»per. it wm OTidently deairable to obeerTe the Um- pentnre of the oylioder-walla of "^a engine, in lome oaae in which water wm andoubtedly preeent in the oylinder. The temperature of the walla in thie oaae ahould be the lame aa the time-mean of the ateain temperature. A oaae of thia kind preaented itaelf recently in the MoGill workahop engine, whioh had to be run throttled, when the out-oif Talve had been remoTed for repaira. The whole of the indioator-diagrama, on reduction, ahowed the temperature of the mean of the ateam cycle to h«Te been the aame aa that of the clearance aurfaoe, within the limiia of error of the obaerrationa, although on aome occaaiona dia- grama were purpoaely taken when the load waa auddenly changed. On NoTember ISth the cut-oif TaWe waa replaced, and the oondenaation after thia date appa- rently ceaaed to be limiting. The following Table giTca a number of the reaulta obaerred. It will be aeen that iae rotation in queation appeara to hold OTer * conaiderable range of temporatun ao long aa the condenaation La limiting. Wall TupiBATDBii nr LmTixo CoromaATioir. Dkia, 18M. Ont- off. Btmarkt. Ttmpentnre* of Steun CyoJ* 1 t Max. Min. Bang*. HMa. Oct. 17 ,. 17 Mot. 24 » 6 ., 5 . Throttled . Half open . Throttled . Biaing . ! "F. 272 825 2o3 292 277 "F. 228 268 241 248 282 •F. 44 67 42 49 45 "F. 258 298 260 266 258 Mon-oondenaing » » >• "F. 254 299 260 267 250 Not. 6 » » „ 10 Falling . . Throttled . RiaiAg . ! 258 236 25fi 277 204 200 212 200 49 86 48 47 280 221 285 251 Oondenaing . » II • 2JW 221 286 248 Mot. 13 .> 18 t Half open . Quarter open 826 812 258 282 78 80 284 265 Oondenaing n • 297 288 [DIBC088IOM. ^r^smmmmmmm TDK. [MiantMoC OomiiinATioir. «T»pontion pro- ibaerre the Um- irhloh water WM lie walls In thie Uure. A oaae of ne, whloh had to >r repain. The npetatura of the learanoe stufaoe, le oooaaiona dia- 1. On November thia date appa- twr of the renilti ■ to hold over a I Um'.dng. ' iir. — ii mdenaing II II II II "F, 2M 299 260 267 2fi0 using -. . . • • 2JW 221 286 248 uing . • 297 288 [DuC0fi8IOM. PnNM«Un«S-] DnomMIOM OH C01IPINI4n01l or BTKAIC. DisoTurioii. Mr. J. Clarki Hawkbraw, OhairmaD, in proposing a vote of Mr. Hawk- thanks to the Authors, said the subject dealt with in the Paper **"*' was a difficult one, but that there were members present who would be able to discuss it and give the Institution farther valuable information on the points referred to. Mr. MiCHAXL LoHOBiDOK regretted he had been unable, in the Mr. Longridg*. time which had elapsed since he had received the Paper, to obtain a dear idea of the Authors' arguments or to form any decided opinion upon their conclusions. The method of making temperature measurements within the metal by thermo-electric couples appeared to have been very ingeniously applied, and he could understand the Authors being so much in love with it as to be led to build, upon the results obtained by it, a superstructure of conclusions perhaps somewhat larger and more elaborate than the solidity of the foundation warranted. Oertain discrepancies in these results were, for instance, attributed to leakage past valves and pistons, and this source of error was stated to be much more serious and important than it had hitherto been considered by experimenters. He thought it required stronger evidence than the Authors had adduced to establish their conclusion. On the other hand, they had satisfactorily proved, what had, doubtless, been long suspected, that the temperature of the inner surface of the cylinder did not follow that of the steam, but moved through a smaller range. Also, the temperature curves obtained drring compression seemed to settle the controversy between Him and Zeuner regarding the presence of water in the clearance spaces of a steam-engine. The point on which information was at present specially required was the relation between speed of revolution on periodic turn and initial condensation, and on this point he had failed so far to extract any information from the Paper. The Authors appeared to hold, p. 4, that cydioal con- densation — which meant, he supposed, condensation per revolution — was independent of speed, or, in other words, that the amount of condensation per hour varied as the number of revolutions, while, on p. 52, W, the condensation in lbs. per hour was said to be approximately independent of the speed, which was equiva- lent to saying that the condensation per revolution, or cyclical %i M er with great mation on the tta, Mr. Bryan naturally only irmometera the had thna been t of the rapid [e Bhonld, how- )liahility of the cl againat leak- lently been im- le with modem 1 the meana of hangea of oon- ed to be moat from them. Id be aeen, the ''t^. 6 differed biat ahown in waa obtained e waa preaaed ron, the other the iron wire romotive force >re wonld be a none. Other vea, and had, t waa new, and notive to have menta on the the orthodox he care whioh B experimenta. aed by others, londnotiyity of 9 experimented ban the other. need for the irmal jnnotions ig experiment. I'rocectllDgi.] DUOUBUON OX OONDmiATION OF BTIAM. 18 The rim of the ooTer wonld have to be ont off and the remaining Ur. Stromeyar. flat plate C, Fig. '2.3, surrounded by a thin aheet of roetal as shown. The dish thna formed wonld be filled with warm water and placed over another dish, I), containing cold water. The difference of temperature mig..t be maintained constant for a long time, and the transmission of heat could easily be estimated. Aa the one side of the plate would be warmer than the other, it wonld curve as shown, and this curvature, whioh oould be moaanred by the movement of the rods BR, or by angular diaplaoement of two mirrora attached thereto wonld give the temperature gradient. Thia experiment, applied to a different purpose, had been made by Mr. A. F. Yarrow,' and Mr. Stromeyer had ahown ' that, with the obaerved movements of the roda RR, whioh amounted to j inch, the amount of heat tranamitted muat have been 142 evaporative nnita per aquare foot per hour. On the anbjeot of the lagging of the in- dicatora, to which the Anthora had referred, an exhanative Paper > had been read at the Institution some years ago, and he had made experiments on it in marine-engine trials,* and had found that, with a 6- foot length of oord, the indi- cator always lagged | inch. In other words the upper or steam line on an indicator-oard taken under these conditions ought to be moved \ inch from the end of the stroke. Hei thought if that were ta-ken into account the amount of oondenaation mentioned would be materially reduced. The leakage of ateam'paat the alide- valves waa perhaps one of the most important anbjecta mentioned by the Authora, particularly aa it was now becoming a general practice to use no lubricants in marine-engine oylindera. He under- stood that many ateamers creased the Atlantic, or went to other parta of the world, naing no oil whatever in the oylinders. A few drops might be applied on starting, and a few drops before reach- ing port, to prevent the valve aticking fast; and he believed the engines worked perfectly. Whether the leakage under such conditions wonld be greater or less he did not know ; but the practice of doing without oil appeared to work very satisfactorily, '■J ..vl' I .^1 ' Transaotions of the Inatitntion of Naval ATohiteoto, 1891, vol. xxxii. p. 108. * "Marine Boiler Management and Construction," p. 94. ' Minutes of Prooeedin;^ Inst. C.E., vol. bcxxii!. p. 1. * Transactions of the Institution of Naval Architects, vol. xxxi. p. 145, and vol. XXXV. p. 407. [tHI INST. ai. VOL. CZXXI.] * ? 66 DIBOtrSSION ON OOKDBMBATIOM OF 8TEAH. [Minntea of Donkin. Mr. Sttomeyer at least at sea. In oonneotion with the leakage of steam, he had frequently noticed in very ?arge slide-valves, particularly single- port ones, that their faces did not wear fiat nor even round, hut were worn in facets. He had noticed thit) particularly in the case of a single-ported engine, with a very small stroke and con- sequently very wide ports. In that case the face was out of line fully jV inch, Fig. 24. Mr. Bryan Mr. Brtan Donkin had had the advantage of seeing at Montreal the steam-engine described in the Paper, and some of the instru- ments the Authors had used in their experiments. The Paper appeared to be a masterly treatise on the condensation of steajn ; and the experiments had evidently been made with some of the best electrical thermometers known. The Authors had described a mass of very delicate experiments lasting one year, and the account now given of them in the Paper appeared the best and most comprehensive yet published on the important question of the temperature of cylinder walls and on the conden- sation of steam. Great care appeared to have been taken on all points touched upon by the Authors. The electrical method of measurintr >vad determining the temperature of the cylinder-walls and covers as described in the Pa3:er peemed to be much more suitabh than the mercurial method, which he had adopted in his exporiir.^ntG. But it wcs a much more delicate method, and tho instruments used — the galvanometers and others — were difficult to observe in a boiler- or steam-engine house. The experinenis proved the enormous importance of cylinder-wall temperature compared with that of the working steam — or the difference in temperature between the walls and the steam — in studying condensation and steam-engine economy. They brought out clearly the fact that the clearance surfaces were much more important than the barrel surfaces. It was stated that 90 per cent, of tho condeusii.tion took place on the clear- arce su7! the slide-valve jd, it was very je inferred that valve used was leyer had men- t of the Sulzer M t.y;^.? ' . \; ^ i- 'JT'. » L i;i l> 1 ProoeodingB.] DISOUSSXON ON OONOBMBATION OF BTBAH. 67 type, which probably would not leak. Apart from the question Mr. Bryan of condensation, the Authors had made an interesting set of""" '"* experiments on the conductivity of cast iron, which appeared to show the value to be about one-third less than had been hitherto taken— 6*4 instead of 7*5. That was a very important result, and he hoped it would be conhrmed by other experiments. With regard to the depth to which the heat penetrated in oast-iron walls, it appeared, at 42 revolutions per minute, to be i^ inch. He might mention that, with a slower speed engine, he had found a still greater depth, | inch to about 1 inch. The question of time was vital in a matter of that kind. He agreed with the remark of the Authors, that the testing of the leakage of a piston and a valve with the engine blocked and stationary was of little use. It was much piraferab.1oth, tended to reduce condensation. He hoped that the classical experiments i! escribed might be continued, not only with regard to the tem- perature of oast-iron walls and steam, but with gas- and oil- engines. He had placed upon the table a very simple instrument, called by Him the revealer. It was screwecJ or? to the indicator- cook of any engine, and (after it was once heated, which shoald be very carefully performed to avoid breakage of the glass), the steam passed in and out at every stroke of the engine, and, aeoording as there were one or two glasses, there were different effects of condensation. It was extremely interesting to watch, through the glass, the rapid condensation and the equally rapid evaporation. When the walls were heated only, an extremely fine mist was formed. With different temperatures of the walls the size of the globules increased from fine mist with hot walls to -^ inch or ^ inch with cold walls, which then ran down the surfaces. Mr. E. B. Dolby noticed, in regard to the thermal conductivity Mr. Dolby, experiments, that the value the Authors stated as their result, 5 '65 thermal units, was for cast iron, whereas "the generally assumed value, 7' 5 thermal units, was for wrought iron." As far as he understood, the Authors did not assert that there was any difference between the value they had found and the value previously assumed. He had been much interested in the tran»>- mission between the surface of the cast iron and the steam, and F 2 .:«! "J ■"i v:-i ■§i & DIBOUBSION ON OOKDBKSATION Of 8TBAM. [Hinatoi of Mr. Draitt Halpin. espeoially bo in comparing it with the heat given off from ob»*-- iron hot-water pipes nsed for heating air. A short time ago I.) had had oooasion to examine the experimental data obtained by Sir William Anderson * npon the amount of heat given off from oast-iron surfaces to air. The experiments were made with a difference of temperature ranging from 0° F. to 200° F. It appeared from Pig. 9, Plate 8, in Sir William Anderson's Paper, that, for a difference of temperature of about 200° F., 2 thermal units were given off per degree of difference between air and cast iron per hour, or about 0'033 thermal unit per degree of difference per minute. Comparing that with the statement on p. 20 of the Paper that 6 thermal units of absorption were noted, it showed what a great difference there was between the two. He could not understand why there should be such an immense difference between the amount of heat given off from the surface of oast iron to a gas like air, and the amount absorbed from perfectly dry steam when there was about the same difference in temperature, and he should be glad to hear an explanation. Mr. DRurrr Halpin had had to deal, 26 years or 27 years ago, with a similar case to that illustrated by Mr. Stromeyer when experimenting with a balanced slide-valve — a valve that was very ingenious, because it oould be seen at work. He had taken the valve-chest cover off and had seen the valve working at a pressure of 150 lbs. per square inch. Trouble, however, was caused by the leakage. After successive trials he had at last found a simple method of demonstrating that the valve was out of truth. Tlie valve was soraped dead true to a standard surface- plate, and was afterwards placed with tongs in a bucket of water at 212° P\ On placing it back on the plate three or four layers of tracing-paper oould be inserted at different places, the valve had so buckled. That was a terrible lesson to show what leak- age might amount to. He thought proper caution should be exercised in using the Authors' statement as to conductivity. Fundamental data were given referring to a cast-iron plate 1 inch thick and 1 foot square. The number of thermal units passing through had been derived from experiments made with the appa- ratus shown in Fig. 12, ia which he thought the conditions were totally different. If a hole of 1 inch diameter were drilled into a 1-inoh iron bottom of a reservoir of water of any depth, there would be a certain efflux of water ; and the hole 1 inch in diameter and 1 inch long would be pipe, and there would be friction in ■ Minutes of Prooeedings Inst. C.E., vol. xlviii. p. 257. if! *?1 -;«;3E:t t. [Minntei of off from oa«*- \, time ago Lj i obtained by ^ven off from made with a 200° F. It erson's Paper, F., 2 thermal )n air and oast te of difference n p. 20 of the iei, it showed ro. He oonld inse difference oe of oast iron perfectly dry I temperature, 27 years ago, romeyer when &lye that was >rk. He had [ye working at however, was had at last valve was out .ndard snrfaoe- aoket of water or four layers oes, the valve ow what leak- ion should be conductivity, n plate 1 inch units passing rith the appa- mditions were re drilled into y depth, there oh in diameter be friction in 257. Proceedings.] DlBOnSBION ON OONDBNBATION 07 STHAM. e» 1 c - the pipe which had to be overcome by the water. But that firiotion M;. Dmitt in the pipe was only an exceedingly small part of the total resist- Halpin. anoe. There was tiiie resistance of entrance and the resistance of exit, which were exceedingly large in a pipe of those proportions. In the same way the amount of heat that could be reoeivjd by a plate, transmittdu through the thickness of the plate, tai then emitted were totally different quantities from the amounts trans- mitted, through the several sections, 1 inch thick, of the experi- mental bar shown in Fig. 12. Oaptain H. Biall Saskbt wished to add his praise to that Capt. Suktj. already bestowed upon the Paper, especially in regard to the method employed for thermo-electric measurement. He hoped to be able to use it in experiments he was undertaking on the specific heat of superheated steam. The Authors' oonolnsion that the missing quantity, instead of being largely due to initial oondensa- tiun, was principally due to valve-leakage, was uo doubt true in the engine tested by the Authors. The missing quantity was enormous, and the valves were such that they might be expected to leak very badly; but that all valves would leak to such an extent was a proj^^jsition that could not be maintained. About 9 months ago he had commenced experiments on the leakage of piston-valves. They were made with a small Willans engine driven by a motor, so that the valve was in motion as in an engine. The leakage found with piston-valves fitted with the ordinary rings and springs was considerably less than that forud by the Authors. On p. 54 the second example was "Willans simple ^r-^" and the missing quantity was 93 lbs. per hour. He had calculated the leakage in that particular trial from the experiments he had just made, and found that it was between 6 lbs. and 6 lbs. — a very, small percentage on the 93 lbs. He also gathered ttom. the Paper that the condensation due to the metal surfaces was comparatively^ small ; and, as he had shown, at any rate in the particular example quoted, that the leakage was very small, it followed that the remainder of the missing quantity was in all probability due to condensation produced by water in the cylinder. He might mention that that was precisely the view Mr. Willans maintained when he read his first Paper— that the condensation in the cylinder was due principally to water and not to the walls themselves. Betarning to the valve-leakage experiments, he found that the provisional law mentioned by the /J „ Authors (p. 32) k = ~ fairly represented his own experiments. ■ft-f * '":!'' y m - m:.:./-'-t ^;'«i*) ■'.fr-5j!-i ■i •JSiWrt""' sv^ to DI801TSSI0K OH CONDENSATION OV STBAH. [Minnteaof Capt. Snktj. This law asserted that the leakage varied with the pressnre and with the length of the valve — the lap, but he found that the coefficient depended entirely on the design of the valve, and was not a constant, as might be thought from the Paper. It not only depended on the design of the valve, bat to a con- siderable extent on the condition of the partionlar valve itself, whether the snrfaoes were properly polished or not. For example, in piston-valves fitted with rings and springs, he found that C was eqnal to 0*004, or about one-fifth only of the valae given in Table Y. But with a plug-piston, ground to fit the trunk of the experimental Willans engine so tight that it was on the point of seizing, the valii« of was 0*09, which was four times as great as the leakage determined by the Authors, and twenty times as great as with rings and springs ; but if the ' clearance was increased to j-^jV? inch, the value of C increased to 0'6, or about 150 times greater than with rings and springs. He had found that the speed of revolution had very little effect, and that the results obtained with the engine standing were also nearly the same as when running, a result he had not expected. He could confirm the surmise of the Authors that the weight of water that could leak through a given crack under a given difference of pressure might be between ten and fifty times greater than the quantity of steam. In one instance, that of a plug-valve with TT^nr i°c^ clearance, no less thsu 1 ton of water leaked through in an hour. He was making experiments with regard to the leakage of piston-rings, and he hoped on a future occasion to give the results. Mr. Walker. Mr. P. W. W ALKER thought that when a subject of such importance as that treated in the Paper was dealt with from the scientific stand- point the Authors had taken up, the result was of great value, but I the value would be immensely increased by having at the same time due regard to practical experience. In the opening of the Paper he noticed that the engine was arranged so that it could have the tumbler moved with a view of obtaining expansion, but when the experiments were being carried out the tumbler was set in a certain position. He should be interested to know whether that position was the best, and whether the lap en the valve had been made to suit many positions, or whether it vas such as to suit the position in which the engine was at work when the experiments were made. He thought nothing caused so much condeasation in the cylinder as having the valve over- lapped. He should have preferred the experiment to be made with an engine with a cut-off, that was, an expansion valve, the ' '-i!iJil,->,» . 'WMt AH. [Mlnntea of le pressure and found that the the valve, and the Paper. It , but to a oon- lar valve itself, or not. For id springs, he Fth only of the 1, ground to fit 10 tight that it '•09, which was ly the Authors, igs ; but if the C increased to id springs. He little effect, and vere also nearly loted. He could eight of water ^ven difference es greater than alug-valvo with .ked through in I to the leakage on to give the moh importance scientific staud- ^reat value, but ag at the same opening of the that it could expansion, but le tumbler was ested to know the lap en the whether it yraa e was at work thing caused so he valve over- int to be made osion valve, the Pnoeodinga.] SnOUSSION ON OONDINflATION OW STEAM. 71 '^main valve only being used for the inlet and exhaust, and not from Mr. Walker, the point of view of regulating the cut-off. lie could not help thinking that some of the condensation arrived at was due to the valve not being what would be considered the best valve. It was difficult in ordinary practice to build much on great refinements of measurement. He did not think too much should be founded upon the mea8ure.-('• W«lk« , area of 8,000 square yards. The engine and shafting were running light, driving all the countershafts, and the engine indi(»ted 80 HP. ; but 6 minutes afterwards work was started, the engine indicated 120 HP. The loss by the ahafong was, therefore, under the most unfavourable oiroumstances, only 25 per cent. Professor W. 0. Unwin said that one lived and learned, and he Pnf. Unwin. now heard that one way to improve steam-engines was to enlarge the steam-pipes. He would point out that steam-pipes could not be enlarged without producing two results. The radiating surfince was increased and the steam was made to stay much longer in the ■team-pipes. He did not think that either of those actions was con- ducive to conveying dry steam to the engine ; at any rate he could quote instances where quite as marked an economy as that which had been mentioned had been achieved not by increasing the size of the steam-pipe but by reducing it. He referred to that matter, however, for a more important purpose. In studying carefully the action of steam in an engine about the middle of the century, Bankiue and others came upon this problem ; ordinary engines used £0 per cent, to 60 per cent, more steam than there was any trace of on the indicator diagram. The progress of the study of the steam-engine during the last 60 years had been chiefly the study of where that steam disappeared to and why it disappeared. The Paper dealt with that problem — the 50 per cent, or 60 per cent, which went in the working of an ordinary engine — not a bad engine but a good one; and he ventured to say that i^e problem was a more important one than the problem of obtaining local economies by altering the steam-pipes. He considered the Paper was the most important Paper on the steam-engine which had come before any English Society since the first Paper of the late Mr. Willans ; and that from the point of view of getting at the bottom of the action iiiside the steam cylinder, it was the most important Paper ever presented to the Institution. He could quite understand the difficulty which the Paper would present to most readers. One speaker had said that he had not been able to master it in a short time. He was afraid that, as fiu as he was personally concerned, in 3 months or 4 months he should have to say the same thing, because the Paper bristled so with new suggestions and observations that he was quite sure that 3 months' or 4 months' study would not be too much to enable the whole of the valuable matter contained in it to be understood. It depended entirely for its results on the applioa- ^- H DISOUBSION OR OOinDENtATIOM OF STSAV. [Minutes of *i H Prof. UuviId. tion of oomparatirely new methocts of thermometrj. There again ha noticed thut a little donbt had been expressed as to whether those methods of thermometry oonld be depended upon. The Taper was really a gift to engineers fron the laboratory. The determinations of temp^^rature were snoh as coald hare beei: obtained nowhrrc else than in the laboratory, with an engine set aside entirely for the purpose, and in charge of obserTers with infinite patieiioe and abla to devote a very large amo ant of time to the Twork. WiV » gurd to the thermometry, "H inat i^td Le said 'V ih,: ofk Jr Callenaar had ^o rival in the world in rtoon, : ' If lAtJijint ?j. Fortunately he happened at Montreal to be ai'si.iifttetfl -mi,- Professor Nioolson, who looked at the question froii. tht- i : '(i^i.v d ,"'>int of view, so that they had a combination of Profossux Nicol! •:' knowledge of what problems there were to solve, and Professor ( allendar's skill and ability, which had led to the Paper being so valuable. The first real discovery of the action of cylinder walls in causing condensation was due to Him about 1857. It had been surmised before, but had never been really measured. So important, so large, and so complex was that action that Him announced there was no possible rational theory of the steam-engine. He was afraid that was still tme, but undoubtedly the Paper had taken a considerable step farther, towards a rational theory — a theory in which the amount of steam that any given engine would use could be predicted from the conditions in which it worked. The Authors had very modestly stated on the first page of the Paper the two most important results at which they had arrived. First they had thrown absolutely new light on the missing quantity — the quantity of steum used by engines and not shown on the indicator diagram. They hf.(' shown that undoubtedly, in certain cases, a very large part of that so-called missing quantity was steam which never reached tke cylinder at all, but escaped another way. How far that oonoiiision could be applied to all engines remained to be detenained; but that they had shown that the valve leakage was of much greater importance than "had been supposed, he thought there could be no doubt. It con- curred with the fact that the best Continental makers had been gradually driven back from all forms of slide-valves to valves which fitted on conical seaticgs. He would only point out how very thorough those experiments haJ been. Engineers had been sometimes content to say that a valve was tight when the steam was let on to it while standing, and it showed no leakage. But ,/. lAV. [Minutes of y. There again 1 aa tc> whether led upon. The iborotory. The mid hare beeu h an engine set ohflorTers with aoant of time to " mat i^cd Ut in the world in at Montreal to at the question d a combination sms there were \ which had led lisoovery of the •>n was due to but had never »nd BO complex ras no possible id that was still )nsiderable step iiioh the amonnt [d be predicted e Authors had Paper the two ed. First they 5 quantity — the shown on the tedly, in certain 5 quantity was U, but escaped i applied to all hey had shown mportance than doubt. It oon- lakers had been 'alves to valves point out how ;ineers had been when the steam ) leakage. But 4 Prooeediag*.] DI80TJ88I0K ON (XmStB^ftGS Of §TKAK. te it was felt that that w"? a verj imperfect experiment — so imper- Prof. Unwin. fee* as to be pron more Mkely to miwlrad than to be helpful. The Authors aad gone (. .ffeiantly to wor* ; thoy had worked out the i^aestion sepan.^eiy, and had determiD^d the leakage of the valve, in cond'ions such as those in wljii a valve was ordinarily vxl. They closed the stoam-porta with leud and drove the vaive by a separate motor a- d ihen measured the leakage. Their actual results f^'d not go beyond the two engines which they ^•l er-^4xmeu(«d with, which were vary diverse, a small high- pressure engi'-j, anu a triple-expansion engine. It was also to be note^. that in the calculation of the missing quantity of different engines, light was thrown by the results upon anomalies which could not before be explained. The second point of the , Authors was their detorminaticn of the fact that the cycle of ' temperature in a cylinder wall was different from th9 cycle of temperature in steam. That, also, was of very great importance It was not quite so new, because the best writers on the Theor of Engines, and especially Professor Cotterill, had insisted Vfe, j ■ strongly, on grounds not so directly experimental, but on r ;^ elusive grounds, that the metal cycle must be different fto^ ii 9 steam cycle. A further result not mentioned on the first ' ^ but a result likely to be of quite frst-rato importance, w^ ! f* discovery of the fact that the rate of condensation by siea^ri on inetal surfaces was limited. That would affect the whole theory of condensation in the most fundamental way. The Paper was far too difficult to discuss very much in detail at an ordinary meeting, but he should like to make a remark on one or two points. He should be glad if the Authors would expand a little the Table which they had given at p. 56, stating more clearly the meaning of the quantities in the first column. Although he believed he understood it, it had taken a good deal of trouble to find it out. He also thought that the Authors might have mede \ little more dear on the surface of the Paper what doubtless c uld be discovered by the reader, that at present the '^.hole of th 'r results hftd been on non-condensing engines. He did not kn' V that they would be different in the case of condensing e '^nesi but it would have added to the clearness of the Paper, if it had been more prominently stated that all the results had been obtained with non-condensing engines. He believed that even the empirical and approximate formula at which the Authors had arrived, and which T\.ried considerably from that at which Professor Cotterill had arrived, would be of very practical use to engineers. i ■1 T» DISOVBfllOR ON OONDBNSATION OF 8TXAM. [MlnntMof h:/' :■: ii'; EH < Prof. Unwli. An expreoaion of the probable amoxint of oondensation, in » ■team-oylinder, working nnder any oonditions, had been funnd by Profesflor Cotterill, which fitted a wide range of experiuiflnte, with a certain amount of ab-mraoy. He made the oondencation depend npon the barrel surface of the cylinder. It was found by the Authors to depend very much more on the clearance surface. That war a question which had already, more or less, arisen, and on ^'>ord had been said. Professor Cotterill was so careful in his ^ .K than hn nonid not help thinking that he had based his oon- jluaion that the barrel surface was important on a tolerably wide induction. On the other hand, the experiments in the Paper seemed rather oonclusive in showing that clearance surface was of more importance. The Authors had certainly given some ▼ery valuable results in helping to determine the action of the deaianoe surface. The Paper bristled with new sugges- tions. It was, for instance, absolutely new to him to hear of explosive boiling of steam in the cylinder-walls such that water was carried oC moohanioally without carrying away its proper proportion of heat. Another fact might be of practical importance, namely, that there was evidence of vortical motion of the steam in the cylinder itself, vrhich made the pressure wid<)ly different at different parts. There were different pressures in the cylinder at the same moment, so that it was possible that indicator diagrams taken from tlie side where the pressure was greater than near the centre, might be considerably erroneous. All those 8..ggefltions were of remarkable interest, and he thought the Insti- tution was greatly indebted to the Authors for their researches, which must have involved an extraordinary amount of labour and l-'itieaoe. Prof. C«pp«r. Professor ''>. S. Capper thoujjht that more information as to the manner in which the practical measurement of the temperatures had been carried out by the Authors would be of interest, particu- larly as to the calibration of the thermometers. The thermopile appeared to have been connected in the differential experiments with two different points at a small distance apart in a hole sunk in the cylinder; consequently it was the electromotive force produced by the contact between a wrought-iron wire and the cast iiron of the cylinder, which had been a measure of the tempm the observed results at different depths until it cut the vertical ordinate at the surface, a very wide margin of error might be introduced, owing to the increasing curvature as the curve approached the inside surface. The nearest point to the inside surface that could be reached was about j^ inch from the actual contact of the surface with the oteam. It had previously been renerally assumed that the rapid alternations of steam tem- perature in the cylinder were not closely followed by the metal to a greater depth than about y^ inch. He did not, however, mean to suggest that in these experiments a smaller depth ought to have been attempted. Extraordinary difficulties had been over- come with remarkable ingenuity and success, and very great credit was due to the Authors for their truly wonderful results, even up to that point; but he would like to know how that 1-^,^ inch had been bridged over in arriving at an estimate of the temperature at the surface. It was highly probable that in read- ing over the Paper again he would find that the Authors had really satisfactorily explained this point, and that he had mis- understood their meaning, but more detailed information as to oalibratiou of thermopiles and the bridging of th^s gap would be an immense help. Ho had always been under the impression that polishing the outside of pipes would prevent radiation, and not conduce to it, as inferred by Mr. Walker. It had also been stated * by Mr. Druitt Halpin that conduction in cast iron was totally dif- ferent when taken along a bar from what it was when taken through a plate ; nnd he arrived at his conclusion by comparing these results with the penetration of water through orifices of different proportions. He failed to realize the distinction, and if he had not known that Mr. Druitt Halpin was sound upon this point he should have imagined that his remarks were based upon an old-time belief in caloric as an absolute substance penetrating through the surface of the metal. ■WtT r^^i^fimfmntm Mr. W, Wftlktr. JV^*. «0. I DiaouauoN oh oondinbation 01 vmu. [HioatMof 0. Mr. W. O. Walkkr had carried out exp«rimenta to compare the effect on the transfer of heat of variously-arranged aurfaoea pro- jecting from the pri juiry unrfaoe of a plate in contact with steam, air and water. Two cylindrical smooth veaaeb were conatmoted of exactly similar dimensions, cut from the same brass tube, each 6| inches long, 2,>b inch in external diameter, and ^ inch thick, tho} were fitted with movable O water-tight lids, through which ^"-^ thermometers wore inserted. When ""y \^^ filled with water, no difference was ( " J found to exist between their re- ""Vy^^^,/"*^ speotive powers to absorb or dis^ charge heat ; this was ascertained bj' heating them together in steam to 212° F., and allowing them to cool either in air or water. The thermometers registered alike throughout the scale. One of the cylindrical vessels was then tried with copper ribs, 6i^ inches long, J inch wide, and 0*012 inch thick, soldered on longitudinally, the ribs were spaced equally and radially round the cylinder, and were tried internally and externally, as shown in Fig$. 26 (I to IV). The external and internal ribs, when tried together, were in the same plane. The two vessels were heated by suspending them in a tin vessel. Fig. 26, 12 inches high, and 8 inches in diameter, having 2 inches of water boiling by a powerful burner ; the wooden lid of the tin vessel was movable, and the thermometers passed through it. The two cylinder* to be compared were filled with water, and when their temperatures were equal, generally about 66° F., tho wooden lid to which they were attached was placed on the vessel containing boiling water; the lid formed a good fit, and the metallic surfaces at once commeaced to abf rb heat. The reading of the thermometers, together with the time, was noted simultaneously at every 10° of the plain cylinder; the time wan taken in seconds by a chronometer, and the temperature to |°; the thermometers became stationary at 210° F. The lid, together with the cylinders, was then lifted off and suspended either in air or water, and allowed to cool, the respective temperatures and the time being noted simultaneously at every 10°. Contact with steam, air or water always referred to the extenua surface of the Fig. 26. ikM. [MlnntM of I to compare the id Burfaoes pro- wst with steam, ero oonHtrnoteii ^ram tube, eaoh i ^ inch thick, with movable through which Qserted. When I difference was ireen their re- absorb or dis- ras ascertained ^ther in steam >wing them to •r water. The istered alike e. One of the wtm then tried od 0-012 inch spaced equally internally and I external and le plane. The vessel, Fig. 26, imetor, having il burner; the irable, and the I two cylinders ter, and when y about 66° F., attached was ig water; the lie surfaces at reading of the le, was noted the time wax erature to ^°; i lid, together d either in air ttures and the Contact with surface of the PiooMdingn.] DuomnioN ON oosttwankTiov or tnuii. ft pylinders, the internal surface being, of course, exposed to water only. The difference which arose in the conditions between some of the experiments due to the variation in the supply of heat from the burner, change in the atmosphere, temperature and cooling we ter, did not affect the value of tue comparative nature of the experiments, as any change had the same effect on eaoh cylinder. The results of the experiments were given in Tables I-III, p. 80. The difference between the temperature of the corresponding plain and ribbed cylinders increased from sere and reached a maximum after a certain time, after which they »gain dosed to equal degrees of temperature. The presence of the ribbed surfaces increased to a considerable extent the rapidity of transfer of heat either when absorbing heat from steam or discharging it into air. With the ribbed surface in contact with water the speed of transfer was also increased, but not nearly to so great an extent. In the case of the externally-ribbed cylinder (II), Figt. 26, the greatest difference in temperature between the corresponding plain one was 18° P., 38° F., and 8° F. reupeotively when in contact with steam, air, and water. When exposed for 63 minutes in the atmosphere the temperature of the externally ribbed cylinder had fallen 102° F., whilst the temperature of the plain one had only fallen 69° F. The addition of the internal to the external ribs (III), appeared to produce little or no effect on the result previously obtained. The temperature of the ex- ternally and internally-ribbed cylinders showed a maximum difference of 18° F. and 33° F. compared with the plain cylinder when absorbing and discharging heat in steam and air respectively. In the case of the internally-ribbed cylinder (IV), the external ribs having been taken off, the maximum difference in temperature over the plain cylinder was 1° F., 1° F., and 6° F. in steam, air, and water respectively ; their outer or external surfaces were, of course, the same in this case. It was interesting to notice the greater effect of internal ribs in the case when the cylinder was cooled in water. The comparatively slow transfer of heat when in contact with steam or air would allow the temperature of the water in the cylinder to adjust itself; when, however, the transfer became rapid, the ribs became more effectual. The external ribs were more effectual in discharging heat to the atmosphere than in absorbing it from steam. This difference might be due to the condensed layer of steam which was deposited on the surface, and which retard*) to some extent the transfer of heat. The external surface was aUc increased by coils of copper wire -^ inch and 33 iuoli in diameter. With a space between the turns from a Mr. W. a. Wtlkar. / ■'M 4 ■'■:<. 80 CISOUSBION ON OONSBNSATION 07 STEAM. [Minates of Mr. W. 0. WalJrtr. Tabu I.— Plain ScBFAcn ahd Subfaob vitb Extiknal Bib*. Stetm. Air, Water. Plain Exteraal Time. PUin External Time. PUIn External Time. Sarface. Bibs. Baibae. BIbe. ''.nrface. Biba. Temp.°P. Temp.oF. S«cond«. Temp.oF. 209 Temp.°F. 209 Hlnntea. Temp.oF. 209 Temp.''F. 209 Seconds. 80 92 85 200 193 8 200 198 10 . 100 118 167 180 160 16 180 174 80 120 188 295 160 132 80 160 154 65 140 154 480 140 107 58 140 182 110 160 168 590 120 89 84 120 112 170 180 182 757 100 76 182 100 93 254 200 201 950 80 68 265 80 77 480 209 209 Tabli n. — Plain Subtaob and Scbtaoi with Extirnal and Intibnal Bibs. steam. Air. Plain Snrface. External and Internal Blba. Time. Plain Snrfiioe. External and Internal Ribs. Time. Temp.^F. 80 100 120 140 160 180 200 209 Temp. » F. 90 116 187 155 168 184 201 Seconda, 85 115 220 825 480 560 690 Temp. °F. 209 200 180 160 140 120 100 80 Temp. °F. 209 191 160 180 107 90 78 10 Minutes. 5 17 84 67 88 138 261 Tablb ni. — Plain Subfaob and Subtaob with Intbbnal Bibs. .'■• Steam. Air. Water. Plain Internal Plain Internal Plain Internal Time. SLrface. Ribs. Surface. Bibs. BurfiuM. Bibs. TenM).°F. Tenm.oF. Temp. °F. 209 Temp. °F. 209 Temp. "F. 209 Temp. °F. 209 Seconds. 80 80 200 199-5 200 196 IS 100 100 180 179 180 174 48 120 120-7 160 169 160 155 72 140 141 140 189 140 134 118 160 161 120 119 120 115 185 180 ICl 100 101 100 96 295 200 199 80 79 80 7» 640 209 209 -i. _ 'KAM. [Minntei of rmiNAi. Bibs. Water. In loe. ExtanuJ Biba. Time. ."F, Temp.°F. SecoDds. i 209 ) 198 10 ) 174 80 > 134 65 9 182 110 9 112 170 } 93 254 ) 77 480 tTD IMTISNAL BiBS. Jr. nal und Time. ?.°V. UiDQtei. )9 »1 5 SO 17 JO 84 )7 67 m 88 78 188 10 261 iBNAL Bras. IVater. [ntemal Ribs. Time. emp. " F. Seconds. 209 19? 13 174 48 165 72 184 118 115 185 96 295 79 640 Prooeedinga.] DISOUBSION ON OONDSmilTIOX OF STBAH. 81 mean of nine experiments, the temperature of the coiled cylinder Mr, W. O. fell faster than the plain one, reaching a maximum difference of Walker. 6*7° F. When, however, absorbing heat from steam the effect was reversed, the temperature of the plain cylinder rising the fastest, and reaching a maximum difference of 4*7° P. This apparent irregularity could be accounted for by the thick layer of steam which was depo8iteMD1N8ATI0N OF STEAM. [HinntM of Mr. Davey. details of the engine, he ooald not attempt to oritioize the eleotrioal inatruments and the thennometiy ; bnt, with regard to the engine, he noticed that the valve was not an ordinary slide-valve, nor was it a piston-valve. It waa a plate of metal pierced with ca'^ties, assumed to be steam-tight between, two snrfeMWS. I'hat was a very old type of valve, and was called an equilibrium slide-valve. It had been used in England 30 years ago, and he believed it w^ used very largely in America at the present time. In England the experience with valves of that type was the great difficulty of making them steam-tight and keeping them so. Again the results would be vitiated by a difference in the condition of the surfaces t-' r t INITI AL P mas UKC Fig. 87. ■ WS'T LBa. P«N. S9UAIia INCH \ [ } ■ \ N ^ acsa IBUM jm^ ^*ll I IHCI — ".'i^Z^ 1 , , , ^ ?*s ».. TOTAIi; STCAU PBII. I.H.K PMK- HOUR ■ ICUl LBS. exposed to the steam. Thero were largo valve-ports in that engine, and he assumed that the valve-ports were in the condition in which they usually were in such engines, that was, just as they came from the foundry, the sand having been washed off either before the engine was put to work, or after, by the rush of steam, so that it was with what the founders call the skin on the surface. The surface of the piston probably was a machined surface, and pro- bably that of the cover also. He thought definite information was necessary upon those subjects, because the condition of the surfaces affected the rate of uondensatioTi. very materially. Mr. Liversidge. Mr. J. G. LiVEBSiiKjE, R.N., wished to point out one use to which Mr. Donkiu's revealor could be applied. There were still many U XAH. pSinntes of oize the eleotrioal ird to the engine, ie-valve, nor was led with catties. That was a very I slide-valve. Tt a believed it ■wm In England the Teat diffionlty of Again the results n of the sarfaoes Praoeodings.] DIBOXTBSION OH OONDEiraATION OT STBAM. 83 I INCN Las. J in that engine, idition in which it as they came oS either before tf Rieani, so that 9 fiurfaoe. The iirfaoe, and pro- ite information ondition of the jrially. ue use to which vere still tuauy L. engineers who had a prejudice against the use of steam in the Mr. Livewldg*. jacket on account of lubrication ; they were afraid of having the steam too dry in the cylinders, and consequently scoring them. He had himself made one of the revealers by utilizing a sight- feed lubricator glass, and had fitted it up in connection with the pipes for the indicator, and the action of the jacket was most usefully shown in the revealor. Without steam in the jackets there was far mqire deposit on the sides of the glass revealer than was the case when steam was 'n the jacket, but there was still water to be seen on the sides of the revealer even when jackets were in use. He should like to assure Mr. Walker that the Admiralty steam-pipes were very well lagged. If they were not, the temperature in the engine-rooms would be unbearable. He might also mention tliat steel waa displacing the use of copper for large steam-pipes. Mr. G. M. Olabk bad had the privilege, some years ago, of Mr. Clark, assisting Professor Oallendar in some thermometrio researches which he had carried out in Cambridge. He had, therefore, an intimate acquaintance with the dexterity and skill with which Professor Oallendar had carried out tliat class of research. He thought that some difiSoulty in grasping the Authors' meaning was due to the compression of their statement, rather than to omisidons in the experiments. If the Authors had put their sum- mary of conclusions on tlie first page instead of on the last, it would ha^e been much easier to follow the general drift of the experi- ments. An attempt had been made in the Paper to measure directly, by means of thermo-couples, quantities which had been deduced formerly by more or less indirect methods; and, as the value of the Paper depended on how that either carried out or did not carry out previous results, he thought it was of great importance that throughout the Paper great stress shorJd: be laid on which were direct measurements and on which were more indirectly calculated quantities. That waa specially the case with regard to Table VI, p. 34, where the Authors brought to bear all their xmnlU of leakage, and proved that the leakage oftloulation or letikage oorreotion was a quantity which could be accepted. It was somewhat difficult to see which of those quantities were measured by thoir new method, and which were old quantities that could be calculated iadependontly. The pith of the whole thing was oontisuaed in line 6, but there was nothing to show how that calculation was arrived at. A simple uumerloal example would have cleared up a grent deal of the diSioulty. The leakage correotion might have been calculated from K = 3, which a 2 •»1'i?tfl^?*t^''" ■•M M DISCUSSION ON CONDENSATION OF STEAM. [Minutes of Ur. Clark, had been g^ven a page or two earlier. Attention had been directed in line 16 to the order of consistenoy under the different trials, bnt it would be seen from line 16 that the quantities varied between 0*0074 and 0*0159, but these were not the quantities which caught the eye as being exactly consistent. The greatest Vf&s twice the least. He did not know whether that meant that the first three columns were to be compared; if so, he did not think the comparison was of much value, because the experiments were conducted under very similar conditions. The trials were run over a longer time, and were only repetitions of experiments carried out under similar conditions. If another line were inserted treating the means in such a way as to bring the numbers together, it might be much easier to compare them immediately. Any assumption the Authors had made with regard to the leakage coefficient, if they only wished to compare these . three columns, was not of very muof value. Another example of the want of distinguishing between directly and indirectly measured quantities, though cf a rather different nature (which p,^■«*»»"'■ iAM. [MinuteB of Sid been directed different trials, lantities varied the qnantitiea The greatest hat meant that so, he did not the experiments rhe trials were of experiments ther line were i to bring the compare them side with regard compare these Dother example and indirectly nature (which on p. 14, in the Y of cast iroE. ked out by the ever, used that had eventually by multiplying ilated from the >int, and some n consequence. ratio ^. That A. nt at different it be some error that Professor 1 into such an i of the thermo- mrt of it must > the outer air ; ud consequent leads, and the that way be without more I given in the B cylinder was rticukx section ProoeedingB.] DIB0US8ION ON 0ONDBN8ATION OF 8TEAM. 85 depended not only on the difference in temperature between Mr. Clark, the outside and inside, but was also iniBuenced by the gradient across every previous section through conduction along the walls. If the piston was moving from hot to cold, the previous gradients would be all steeper than the particular one chosen. There would, therefore, lx% so to speak, a gradient pe :allel to, as well as perpen- dicular to, thy wallfi Iz extrapolating these from observations taken near the oziiiuoe to tb*) temperature at the actual sirface, perhaps not the real thickness, but some virtual thickness should be taken depending on this con\pound gradient. That might, to some extent, modify the Authors' oonolnsions as to the temperature of the surface of the walls, and hence also of the quantity ^i the rate of condensation. The criticism might be superficial and perhaps not very sourd, but it was a significint fact that throughout the Paper the results obtained from observations taken from the cover were much more concordant than from those taken in the side walls. If the Authors' Table III, on page 19, were plotted for j- in terms of speed, it wojld be found that the obser- vations did not lie along with those taken from the cover, and that would be expected from the arrangement of the thermometers. All the cover thermometers were arranged in a circle of IJ inch radius round the axis of the cylinder, so that they were in such a position that a correction for a virtual thickness would not have to be applied. The central portion of the cylinder cover would scarcely be influenced by the changes in tlio walls or the changes of temperature in the flanges, whereas those in the side would be more immediately influenced. It was pointed out by the Authors that it was essential that no foreign material should be introduced into the thermo-couple circuit, but on pp. 6 and 9 two places were referred to at which such foreign matter whs introduced, first, in the mercury-cup contacts, and secondly, by the tin used in soldering the couples in place. It would b interesting to know whether the formulas given (p. 6) he3< for both the original and subsequent calibrations of the couples, and to what extent these angles were permanent. The formula^ too, should, for the sake of uniformity, be expressed in tl t Fahrenheit scale. Again, as V the differential metho(i : working, changes which affected joth junctions were not fa- served, the Authors might more definitely have explained the " small portion of the oast iron of the cylinder," which was used as part of the circuit, especially with reference to freedom of one ^mw-' ^ DI80U88ION OH OOHSEmATION OV 8IBAM. [MinutM of Mr. Clark, junotion from oyolioal ohangefl. In the desioription of die c' tnation of the thermo-oonples, it was not evident where the jtinotioD - S, inches, subsequently referred to in the Table on p. 19, was placed. Also, it was not oleasr what the " four similar holes " were "abcve and below" (p. 10). Were they neatly on a vertical diameter of the cylinder ? Again, junction No. 1 side woald appear to refer to a couple placed opponite to the clearance space (p. 10) and not to the one previously specified as being situated 2 inches along the cylinder. If, however, "at this point" referred to time and not to place, the meaning of the /uthors would be quite different. The word "depth" had been us 1 by the Authors to . indicate thickness of metal measured from the inner surface of the cylinder. The mind-picture was that the tuermo-couples were in8er\'ed in holes drilled in the metal. The word " depth," there- fore, b'^ore more strongly on the depth of the hole than on the thirkmss of metal left undrilled. The Authors had explained their ute of the won. on p. 6, line 6, which, if understood, might bi sanctioned for its brevity. In Fig. 10 it was shown that the emperature began to rise at each point before steam reached it, wliy Hhould not horizontal conduction play an im- portant par\? On p. 13 it was stated that condensation on the admission surface was due to the lowering of temperature by coii 1km. It would be of interest to know under what con- ditions the engine was running whilst the measurements of the outward temperature gradients were taken. If it was at one- fifth cut-off, it would be more natural to put tho statement in that paragraph rather than in the next. Again, in the same paragraph, were both the temperature of the "inner surface" and also of the outer surface extrapolated from observations at different depths, or was the outer directly me the difference iu electrical conductivity. The alteration of the brackets from one sentence (line 8) to the next would render matters clearer. In the paragraph headed " Density," it would have . been more logical, and ut the same time drawn tho attentioi^ from 1 C, - e d f i a I 8 a 1 ( ^•i«jy>^:ai!~Miiifft\A.-N. 1^^ i^iimatsm. JNeateJjgi LM. [MinntM of of die c'tnation a the jtinotioD I on p. 19, was lar holes " were ' on a vertioal e woald appear le space (p. 10) tuated 2 inches "eferred to time TOuld be quite )he Authors to r surface of the o-oouples were • depth," there- le than on the had explained if understood, it was shown t before steam n play an im- nsatiou on the emperature by der what con- rements of the it was at oue- statement in 1, in the same nner surface" }bsexvationB at ired? In con- it should be hble amplitude, sphere and the perature range. ; wati doubtful ning that the intrustworthy, for oast iron expect od from ^eratiou of the would render " it would have atteutiob fzoiu •^^i^S^^IM Pioeeedings.] DIlCmSSION ON OOVDBMSATION OF 8TBAM. Fig. Q e 87 4'6Hr.aark. ^ = 1' 12, to have said that the viilnes assumed were e 20, and hence k— 5'4. An amplification of the portion dealing with cyclical heat absorption, showing how to advance from actual observations at definite depths to a fully drawn-out temperature-depth curve, to surface temperatures and to heat absorbed, would be of real utility. There were three dotted curves shown on Fig. IS; the boundary curves were explained, but the intermediate one was not referred to. The wave-length and heat absorbed might also be expressed algebraically as well as numerically. There was no explanation given as to the sepa- ration of Table III into two parts. An example of the method of calculating column 5 should be given, and it might be ex- plained that column 6 was obtained from Table II, column 5, by multiplying the values here given by the ratio column 5, Table III, to 10. In Table YI, line 21 gave act the usual lbs. gross feed per LHP. hour* but lbs. corrected feed (line 7). This should be noted particularly, as it was an unusual method of stating their con- sumption in the engine. The usual method would be to use line 6. Column 7, line 21, would then be 66 ■ 5 lbs. against 23 * 8 lbs. It was to be remembered that the engine was desigied to run at 260 revo- lutions per minute, double-acting, whereas in these tria! ' was run simple-acting and far below the designed speed. Dr. A. B. W. Kennedy wished to carry Mr. Davey's dia|^;«.^ a Dr. Kennedy, little further. He would take an extreme case, and suppose that there was no expansion at all, but a perfectly rectangular diagram. Putting 10 per cent, or 20 per cent, or even 100 per cent, more water into the cylinders certainly neither would nor could make the least possible difference in the power, but to have to pay for the additional water (in cost of coal) would, nevertheless, be a very serious matter. He used this extreme case to show that it was not the proportion of the power, but the proportion of the water that was of importance to practical engine users, for here not even the small gain in power mentioned by Mr. Davey would follow from tho saving of the whole additional water, but the actual practical gain in cost would follow all the iiame. The important matter in the present case was the matter of condensation, and it must be measured in condensation, os it had to be paid for, and not in power. He was glad to hear Prof. Unwin's references to Prof. Cotterill. Probably it had not come into the Authors' pro- vince to mention him; but his work in connection with the theory of steam-engines in regard to condensation in the cylinder and to the sfiiion of the jacket was most notable, as having IIUMWIII ■PMippMllii ir IS DIB0XT88I0K OK OONDSKSATION OF STEAM. [MinaU« of Dr. K«nu6dy. been thought out by a man who, at the time he wiote, had little experimental work to go upon. It was one of the most splendid pieces of work in the theory of the steam-engine, and it was not reduced in importance in any way by the more ooirtipleto experimental work which the methods of I'rof. Oallendar and others had rendered possible. He thought it very unwise for gentlemen who had not used electrical methods to throw doubts upon them. Platinum thermometers and electrical resistance in- Btrumeuts generally were at least as easily verified as mercury thermometers. There was some difficulty in verifying either, but the electrical tharmomotora were correct within a degr-" of accu- racy unattainable by nierourial thermometers. If a doubt was felt as to Prof. Oalleudar's result, by anyone who was prepared to swear by an indicator card (which the speakei was not), ho should refer to Fig. 18, showing a curve diawn representing the tem- perature calculated from tlin pressures on an indicator diagram throughout the stroke. On ♦ha^. continuous ourve were placed a number of crosses, which indicated the temperature measured by a platinum thermometer e aowu in another Fiy. It was evident that the readings of the platinum thermometer and the readings of tlie indicator springs had been practically identical. He him- self considered this to be a check upon the indicator springrt, not upon the platinum thermometer ; but engineers who had a kindly belief in indicator spiings, as friends of their youth, might, if they thought proper, regard it as a reasonable corroboration of the platinum-wire thermometer. Mr. Davejr. Mr. H. Davey said he ought to have added " missing power as the result of cylinder oondensaticn." He wished to point out that the missing power, in the best modem englnos at least, was a smaller quantity than was asually associated with the large amount of cylinder initial oondensaticn which took place in some engines. In the example he had given he saw no advantage in extending the diagram for the actual eagine, as the terminal pressure of the adiabatio would be below the pressure necessary to overcome the friotional resistance of the engine. Dr. Hopkinaon. Dr. JoHN HoPKiNsoN desired to refer to one point connected with the particular methods of thermometry used. On care- fully reading the Paper, it appeared that two electrical methods of thermometry had been employed : the thermo-electro method, which had been extensively used by Le Ghatelier in France, and the method by resistance, which Professor Oallendar had brought to great perfection in England. Lat on a superficial reading, the second might escape notice. He thought it would have t _^- :-^- iM. [Minattia of he wrote, had ne of the most engine, and it more oon^piete Oallendar and ery unwise for o throw doubts resistance in- ied as mercury ■ing either, but Jegr-" of acou- f a doubt was eas prepared to not), ho should iting the tem- icator diagram B were placed a e measured by It was evident d the readings ioal. He him- it springn, not .0 had a kindly outh, might, if orroboration of missing power d to point out IB at least, was vfth the large c place in some o advantage in the terminal sure necessary }int connected ed. On oare- trical methods leotro method, in France, and r had brought ioial reading, > would have Prooeedingik] DI80UB8I0M ON OONDENSITIOM Or 8TSAU. W been of great value to angineevs if those methods had been Or tiopkinton, described in detail, because he believed that was the first time that either method had been employed for measuring the instan- taneous value of rapidly varying temperatures. He believed that these methods of theimometry would bo largely employed . in future by engineers, because, compared with the mercury thermometer, they were more accurate, inasmuch as they were very much more rapid in their action. In the Authors' experi- ments it would be impossible to use the mercury thermometer for measuring the temperature at the various [K>ints in a com- plete cycle. That difficulty had been completely overcome by the two eleotriiml methods, and he ventured to think that the value of the Paper would have been considerably increased if a description had been given of methods which, though familiar to many physicists, were unfamiliar to most engineers. Mr. W. H. Patc'hkll had erooantered the difficulty re rred to, Mr. Patckell. but with the assistaQoe of Mr. ^. W. Burstall, had been able to simply accomplish with the platinum thermometer experiments which would have been im}M)S8ible with a mercury thermometer,^ He thought Mr. Walker would not find the electrical thermomrjter harder to use, or more brittle than the ordinary glass thermometer. There was now only one missing link. The apparatun had to be used by an experimenter accustomed to electrical instruments, but now, happily, even that difficulty was being overcome. Two indicators had been placed upon the table, and they appeared very much like the steam-gauge, with which Mr. Walker was no doubt familiar. If he would only get a thermometer of the kind described and oouple one of the indicators on to it, he would never want another mercury thermometer. Capt. Sanket thought it might interest members to know that Ctpt. Sanke;^. early in 1892 Prof. Callendar and Mr. Willans had undertaken steam-engine experiments with platinum thermometers, but the lamented death of th<9 latter had prevented the trials from being carried out. Sir John Wolfe Babbt, President, -wid the institution was greatly Sir J. Wolfe indebted to the Authors for their Paper, which from all points of ^"'7* view was of great value. It was a matter of regret that they were not present to hear the discussion and to verb&Uy supple- ment some of their descriptions with the explanations desired by some of the members. Another matter of re;,Tet was that neither of the Authors was able to write " Associate," " Associate Mem- .1 Prooeed'iga of the Institution of Mechanical Engineers, 1896, p. 134. m niSOTTBSIOir on OOVDBmiATION or 8TBAK. [Mlnntoii of i sir J. Wolfe ber," or " Member " after their names. He hoped at some fottire *•"^ time their names would be neen with some of those titles added. The Authori. I'be Adthors, i, -r ly, regretted their inability to be present during the disonsaion, as it was in many cases difficult to under- stand the exact meaning of questions in the report. They had been taken to task by several of the speakers for the obscurity > and difficulty of the Paper, and for the omission of important details and explanations, for which they were not responsible The Paper, as originally written, contained many details of physioal apparatus and methods, and much explanatory and illustrative matter, which were omitted in publication, owing to the necessity of bringing the communication within the limits of space at disposal. This compression of the Paper had rendered it more difficult to understand, and from a physical point of view, less interesting and oonvinoing. The Authors, however, hoped that the general results would prove to be suffioiontly intelligible on caretui study, and that they might have the benefit of the doubt in case any obvious and important test appeared at first sight to have been omitted. The ingenious method of measuring the conductivity of a plate by the curvature, as described by Mr. Stromeyer, could not have been applied to the cylinder cover used by the Authors without destroying it. The two methods adopted in the ease of the 4-inch bar were better known, and appeared to be capable of greater aoouruoy. They had not, however, been content with assuming that the metal of the cover and the 4-inoh bar, though oast from the same ladle, possessed identical properties. The temperature cycles had been observed at different depths in the metal between •j^ inch and | inch. The ajjniement of these cycles on reduction fr- .ned a satisfactory test, not merely of the quality of the metal, but also of the accuracy of the observations and of the validity of the method of reduction. The indicators used throughout the trials were of the Crosby pattern. They were tested against a platinum thermometer, and in various other ways, with consistent results. The indicating gear was specially fitted and tested for high-speed work. At the V low speeds actually employed, it was not possible that the lag should have been of the order suggested by Mr. Stromeyer. ; I It might be observed on a close inspection of the section of the valve and relief-back shown in Figs. 1, Plate 6, that the probability . ®^ bending or buckling of the valve as an explanation of the leak- age, wua minimised by the symmetrical form of the valve and its f . [M iMtM or t aoine fatnre those titles to be prMent nilt to nnder- They had the obeourity of important responsibi. ly details of lanatory and ion, owing to t the limits of d rendered it t of view, less r, hoped that ntelligible on of the donbt first sight to ity of a plate mid not have ithors withoat ) of the 4-inch )le of greater ^ith assuming ugh cast from > temperature aetal between I on reduction of the metal, ' the validity f the Crosby mometer, and tie indicating rork. At the that the lag oeyer. leotion of the e probability I of the leak- t^alve and its ProoMdiBgB.] DnOVSnOM OH OOHDIRBATION Or nWAU. •1 freedom from strain. The surfaces were also tested at interraliTlMAathen on a true plane, and were not found to have undergone any marked deterioration as a oonsequenpe of wear or of unequal heating, as suggested by Mr. Druitt Halpin. The lubrication of the valves of marine engines without the qm of oil was readily explained on the Authors' hypothesis with regard to the nature of valve leakage. The Authors had recently been able to verify by direct ex periment the conclusion that a large part of the economy obtained by superheating or jacketing, especially in small engines, was due to the great reduction of the possibility of leakage in the form of water. They regarded these experiments as a strong confirmation of their views on the nature of valve leakage, and hoped shortly to be in a position to publish full details of the method employed. They had not intended to advance the method of testing the piston leakage by running the engine single-acting as original. Tlxey were unable, however, to quote any case in which the leakage had actually been measured in this manner, and they were not aware that any previous obsorvers had given a method of deducing the leakage between any two points of the stroke. The reason for making the engine single-acting was that the conditions at the two ends of the cylinder were very difinrent, and that it was impossible to obtain satisfactory observations of the metal cycles at the crank end. The reason for confining attention to non-condensing trials had simply been that the time at their disposal was limited, and they preferred to work out one case as completely as they could. The experiments of Captain Sankey on tho leakage of plug- and piston-valves were regarded by the Authors as an important verification of their formula. It was evident that the leakage of plugs must depend on the fit, and that of piston-rings on the strength, of the springs. The dependence of the value of the leakage coefficient C on the width of the fissure was quite clearly stated in the Paper, p. 82. The important points were that the leakage of any particular valve was approximately pro- portional to the pressure-difference, and independent of the speed, assuming that inertia effects were absent. The Authors had also made tests two years ago on the leakage of the Willans engines belonging to McGill College, which were regularly used for electric lighting. They had followed the method described by Willans of blocking the eng^e in various positions. This method was certainly unsatisfactory in the case of an unbalanced slide- valve, but might, as Captain Sankey had stated, be trusted to give a fair indication of the state of the piston-ring»in a Willans engine. ■mi ^^'^mmm 'mmsmmmm mimmmmmmm ^^^^^^^'^mM\mi^mfm^ ■ji^if^ee s'X t ss t. 'Tf -v'fi^'i'^eiti ^ IMAGE EVALUATION TEST TARGET (MT-3) '^O // '<^<^ .V^ 1.0 I.I 1.25 141 ^ m 2.2 Hi I 40 1.4 1.6 .>;feS ' ' i^^n\i.;/dl*»*3«*^^ '-'^^ ■ "-^*- ■r--'^-'-^^¥:iW¥^^'^^'^ n DIS0X7SSI0N ON OOKDENBATIOM OF 8TBAH. [Minutes of i V The Authori. They were grateful to Professor Unwin and Mr. Bryan Donkin both for the generous appreciation they had expressed at the meeting, and for the interest they had taken in the work during their visit to MoGill College. That the temperature oycle of the surface of the metal must be different from that of the steam, had been shown by Eirsoh and Cotterill on theoretical grounds, to which adequate reference had been made in the original paper. They had, however, assumed too large a value for the oonduotivity of iron, and they explained the difference in the cycles by supposing that the temperature of the metal surface ceased to fall as soon as re-evaporation was complete. They both assumed that the temperature of the metal surface during conden- sation practically followed that of the steam. The latter result appeared also to follow from Hall's experiments, and from others which might be quoted, but was in direct opposition to those of the Authors. The explanation given by Kirsoh would evidently fail in the case in which water was continually present in the cylinder. This case had been specially discussed in the Paper, p. 36, and afforded one of the strongest confirmations of the views there advanced. With regard to the desirability of extending these experiments to engines of a larger size, the Authors were glad to be able to state that, with the kind permission of Mr. Wanklyn, the general manager of the Montreal Street Railway, they were engaged in testing a large compound-engine of 3,500 HP., directly coupled to a single generator of 2,500 kilowatt capacity. The engine could be tested condensing or non-condensing, and with or without steam in the jackets on any part of the cylinders or receiver. All the quantities, with the exception of the condensing' water, could be directly measured, and arrangements had been made for auto- matically recording the temperatures of the metal and steam throiighout by means of electrical thermometers. The phrase " cyclical condensation " had been used by previous writers on the subject. It meant simply condensation ooourring cyclically, and not condensation per cycle. The temperature cycle of the surface of the metal, as given in Table III, had been de- duced by the aid of Table II, assuming simple harmonic formulas. As explained on p. 19, this method might lead to some error ii^ certain cases for the surface range, but the value of the quantity > of heat absorbed, upon which the results mainly depended, would be very nearly correct. Eirsch hud given several ingenious methods of deducing the heat absorption for a oycle of any form at the surface of th§ metal. Ue assumed the surface cycle to be »':,y WM ' ^-—wwwww^am EAU. [Minntes of r. Bryan Donkin ixpressed at the the work during nperature oycle rom that of the 1 on theoretical )n made in the •o large a value the di£Perence in he metal surface ete. They both 9 during oonden- 'he latter result and from others ition to those of would evidently present in the d in the Paper, Qns of the views ese experiments id to be able to ^lyn, the general vere engaged in eotly coupled to lie engine could ith or without »r receiver. All ins' water, could ; made for auto- etal and steam sed by previous Ation ooourring mperature oycle I, had been de- monic formulas. > some error in of the quantity epended, would ireral ingenious ole of any form face cycle to be Piooeedinga.] DISOUBSIOlfr ON OONDBNSATIOK OF STEAU. 98 the same as that of the steam. The Authors were unable to make Th« Anthon. use of his methods, as the form of the surface oycle could not be satisfactorily deduced from the observed oycle at a depth of 0*040 inch in the metal. Some not unreasonable doubts had been expressed with regard to the results obtained by means of the electrical thermometers. Dr. Hopkinson and others had asked for further details of the con- struction and testing. These details had been excised as being of too purely physical a character. The Authors might fairly claim considerable experience in the use of these instruments ; they had spared no pains in the testing, and they hoped that the evidence of Dr. Kennedy and others who had spoken in their defenoe would be taken as sufficient proof of the capabilities of the electrical methods. They hoped, with the kind permission of the Council of the Institution, shortly to publish some of the more important of these details. The account of the conductivity experiments had been com- pressed from ten pages into two. This would account for some obscurity in the statement of results. The same apparatus had been used for both methods. Of these, the second method was the most accurate and direct, because it most closely corre- sponded with the actual propagation of heat-waves taking place in the metal of the cylinder itself. The Authors were inclined to agree with Mr. Clark that the headings of the different Tables were not sufficiently clear. Nearly ten pages of space, however, had been saved by omitting the explanations given in the original. The agreements referred to in Table YI were between the observed and calculated values in lines 16 and 17 respectively. This agreement proved that if the calculated value of the condensation were correct, the leakage must be very nearly independent of the speed. The subsequent remarks of Mr. Clark referred chiefly to matters of taste and expression, which oonld not be profitably discussed at that stage. He had succeeded in finding the correct answers to most of his own questions, and the Authors hoped that time would resolve the remainder of his difficulties. With regard to the effect of the state of the surfaces discussed by Mr. Davey, the Authors had explained, on p. 61 of the Paper, that, according to their views, the amount of condensation was limited chiefly by the possible rate of condensation of the steam, and that the state of the surfaces was, therefore, of less con- sequence than was generally supposed. In confirmation of this view, they had made experiments with surfaces of various kinds 'M f-^ \. * (iSSItoliSiSs&.iVviS'til i ■ : 94 0OBBB8PONDBN0B ON CONDENSATION OF 8TBAX. [Hinntea of The Anthon. from oaat iron to polished platinnm, and with various amotints of adherent water. Some of these experiments had been recently oommnnioated to the British Aosociation, and had already been published in Engineering. f-^,. . .; ; Oorrespondenoe. Prof. Ewing. Professor J. A. EwiNQ congratulated the Authors on the immense amount of important and novel matter which the Paper contained. It was a long Paper, and, notwithstanding that, the amount of information it contained was so considerable that being paoked into a relatively small compass it became somewhat hard reading. Especially was this the oase towards the end, where the general conclusions were stated. It was diflGioult for the reader to realize the great quantity of experimental work which the Authors had succeeded in accomplishing. Much remained to be done, and iu several places tae Paper must be held to open questions rather than settle them ; but, as it stood, it was undoubtedly the most important contribution which had yet been made towards the elucidation of that difficult matter — the action of the cylinder walls on the steam. It was all the more interesting that in several respects it conflicted with existing notions on the subject. On one matter of primary importance, namely, the time rate of condensation of steam on a metallic surface, he understood that one of the Authors had recently been making experiments by a more direct method, and it would be an advantage to have the results of these added to the Paper. He would be glad of fuller information as to the data from which it was inferred that the rate of evaporation was conditioned in the same way as the rate of condensation. It would be interesting also to know whether the rate of condensation was affected, to any considerable extent, by the presence of air mixed with the steam. It was at first sight somewhat startling to find the steam in the cylinder superheated during the admission part of the stroke ; but that was no doubt compatible with the presence of a quantity of water on the exposed surfaces. The condition of the mixture was then far from one of equilibrium, t»ad it was conceivable that the main volume of the freshly-admitted steam should show superheating, while at the V same time there was % layer of water condensed on the walls, so that the subsianoe taken as a whole was wet. It would be very interesting to know whether the apparent fall of tempera- :. ture during expansion, to a value below that which corresponded to the indicated pressure, was due to a local reduction of pressure ■-iJ^ «ai 'f^si' "^Sm"^^ lAM. [Minnten of ions amounts of 1 been recently ftd already been nthors on the tter which the lotwithstanding i so considerable pass it became lase towards the It was difficult erimental work iishing. Much er mnst be held it stood, it was h had yet been ter — the action Lore interesting notions on the imely, the time , he understood ixperiments by ^ to have the I glad of fuller erred that the f as the rate of w whether the .ble extent, by I at first sight er superheated was no doubt on tlie exposed far from one of volume of the while at the on the walls, It would be II of tempera- L corresponded on of pressure "• \mm m m . • H>^Hr^*. ProceedingB.] OOBBKSPONDENOB ON OONDENaATION Or BTBAM. 95 through eddies, or to true super-saturation, such as the Authors Prof. Ewing. had conjectured. In oases where "the limiting condensation" was reached, namely, when the walls remained wet during exhaust and compression, it might be questioned whether an amount of condensation greater than the stated limit might not occur. The accumulation of water in the clearance-space might lead to an intimate mechanical mixture of the water with the incoming steam, resulting in an amount of condensation which would exceed the limit attainable if the steam were merely brought into contact with metallic surfaces. The evidence by which the Authors were able to supper'-- their conclusion as to valve-leakage was very strong — so far as the engine on which their experiments were made was oonoened. But it could scarcely be supposed that such a leakage was other than exceptional, for in many other engine trials the whole "missing quantity" was so moderate in amount that such a leakage would more than suffice to account for it all, and would leave nothing, or less than nothing, to be accounted for by condensation. The Authors did well, however, to insist on the importance of considering the leakage in all estimates of condensation based on the " missing quantity." In that, as well as in many other respects, their conclusions were of the highest importance. The Paper was an admirable example of scientific analysis, and it was to be hoped that other experi- mentalists would take up the work and extend it. Professor T. Olaxtos Fidleb remarked that the theory of the Prof. Fidler. steam-engine could never be made complete by any study of thermo- dynamics unless supplemented by research of a different kind con- ducted on the lines of Fourier's investigations ; and the Authors' experiments on these lines afforded a valuable contribution to the subject, even if it should be found that their results had been affected by an unusual leakage of steam at the valve. The Authors appear to regard the leakage as normal, but the measurements would hardly be accepted by engineers as applicable to steam- engines in general. In the compound engine at University Oollege, Dundee, the leakage of the valves had been repeatedly measured, and it was difficult to conceive that the observed leak- age, which was always a negligible quantity when the valve was stationary, would be increased a hundredfold or thereabout by setting the valve in motion. Moreover, it was obvious that if the Authors' correction for leakage were applied to the recorded resulUi of Mr. Willans and other observers, the " missing quantity" would wholly disappear from the account, and in many cases the steam admitted from the valve-chest would appear as a smaller quantity h^^?i*=S?iv'^- -'-!«m^'a*a^-" l^^^^^^^p^^pf^i^p^^^i 06 COBRXSPONDEMOB ON CONDENSATION OF STEAM. [Minntet of Prof. Fidler. than the steam present at ont-off, and the trials would show initial evaporation instead of initial condensation. But in spite of this accidental defect of the engine, the Authors had obtained resnlts which could not fail to be of interest to engineers who had attempted to devise a quantitative method for estimating condensation in the cylinder. The question was beset by innumerable difficulties, and while condensation was in progress at one part of the surface it was probable that evaporation had already btigun at another. Confining attention ;to an unjacketed clearance surface, and neglect- ing the barrel or replacing it by an equivalent augmentation of clearance surface, a fsw equations, which must be true in all oases, could at once be written. Thus, after due allowance was made for the loss of boat by radiation, the metal might be brought into a debtor and creditor account ; and if a denoted the area contained between the two extreme phases of the Fourier curve, it would appear that the heat received from the steam at one part of the cycle and afterwards returned, must be A^ = a e F, where F was the area of the surface. Whatever ambiguity might exist as to the exaov form of the curve, this consideration was yet sufficient to show that in most cases the extreme range of temperature in the surface-temperature cycle must be far less than in the steam- temperature cycle ; for if it were not so, the quantity h^ would suffice to effect a much greater condensation than was generally observed. If the mean indicator diagram for a prolonged steam trial was analysed at a succession of points along the curves of compression and expansion it would give further indications as to the temperature changes at the wall-surface. By this analysis it was possible to fix, within close limits of error, the beginning and the end of each period in the alternating interchanges of heat — points which must apparently coincide with the two inter- sections of the curves of surface-temperature and of steam-tem- perature. And, at several points, it was possible also to calculate the varying rate at which heat was being transferred between the steam and the metal— or at all events between the steam and some external body whether metal or fluid. Plotting as abscissas the successive angular values, d, of the orank motion, and as ordi- nates the calculated differentials j^, a pair of curves might be traced, or partly traced, whose areas must, em hypothen, represent the quantity of heat A, as found by the actual trial. This curve, so far as it could be traced, would represent observed facts which were not dependent upon debatable opinions ; and although the ourve could not be completed, at least its length, its average height, and a I 1 ■■■■■■ AU. [Minntef of iM show initial in spite of this >btained results } had attempted londensation in ible difiBoalties, of the surface an at another, oe, and negleot- iigmentation of ne in all oases, mce was made brought into a area contained nrve, it would le part of the \ where P was ^ht exist as to 'et sufficient to Brature in the in the steam- ntity Aj would was generally olonged steam ' the curves of iications as to ' this analysis the beginning banges of heat he two inter- of steam-tem- 80 to calculate jrred between the steam and ig as abscissas 1, and as ordi- -ves might be , represent the s curve, so far ts which were gh the curve height, and a Proceedings.] OOBBKSPOKBBNOB OM COXDBNSATION OF SHilC. 07 few of its actual ordinates, were known. Then if it was true that Prot Fidler. the heat-loss (shown by analysis) had really passed into the metal, the following relationships must be exhibited : —(1) The area of the curve above mentioned must represent the saitae quantity of heat as the area, a, of the Fourier curve. (2) The ordinatas of the first- named curve must everywhere be proportional to the inclination of a tangent drawn to the corresponding Fourier curve at the wall- surface. Up to this point the equations were independent of any opinion that might be formed as to the causes which govern the rate of transfer between steam and metal ; but again, if it was true that the rate of heat-transference was simply proportional to the instantaneous difference of temperature between the steam and the wall-surface, then the following relationships must also be exhibited: — (3) The observed ordinates r-r should be all pro- portional to the temperature-differences or to the height inter- cepted between the curves of surface-temperature and steam- temperature; BO that the surface-temperature curve could be (in part) constructed by simple addition or subtraction, or the curve -=-2, constructed from the observed cycle of surface temperatures. (4) Drawing at each successive phase of the cycle a tangent to the Fourier curve at the wall-surface, these tangents must all radiate from a single point so long as the pressure of the steam remained constant — as it did approximately during the period of admission, though not very closely. (6) If it were possible to assume, further, that the restoration of heat was governed by the same law, then the two loops of the curve must either have equal areas or must exhibit only such a difference of area as could be attributed to known external gains and losses. With all these simultaneous equations in view, it was possible, without insisting on their accuracy, to see that the fluctuations of surface-temperature must have certain phaseis and must be confined within certain limits of probable deviation. It might be interesting if some such comparisons couM be n^ade between the related quantities in the experiments recorded in the Fktper; but it would be hopeless to attempt the analysis of a mean diagram if it were found (as in this case) that nearly nine-tenths of the steam missing at cut-off bad to be written off as the probable leakage at the valve. From the imperfect, firagmeatary evidences that could be obtained from experiuvintal corves of -j^, it did not appear possible that the rate of transfer could depend wholly and simply on the tempera- [THI hist. C.I. VOL. OXXXI.] H 98 OORBBflFOMDENOB ON OOMBBNBATION OT BTSAM. [MioutM of another variant, dryness of the ^'^ Prof. Fldler. ture-differenoe. it seemed to depend also or> whioh might be either the density or the steam. Jfr.FitjiG«r»ld. Mr. MiURicB F. FitzGerald regarded the Paper as a yalnable contribution to experimental knowledge of the temperature cycles in steam-engine cylinders, containing, as it did, a large number of observations on points as to which, np to the present, limited experimental information was obtainable. The Authors' observa- tions might even be described as the first real attack, except that of T^fessor Hall, on the problem of directly measuring, with reasonable exactness, the cycle of temperature in a steam-engine, for either metal or steam ; mercurial thermometers being hopelessly sluggish for the purpotte. The first general f ondusion of the Authors, that the rate of condensation on the surface of the metal was simply proportional to the difference in temperature between the steam temoerature and the temperature of the wall, seemed to be in fair accordance with the results of Mr. Bryan Donkin's and Colonel English's experiments.^ The agreement was not dose for small differences of temperature, of 3° or 4° only, between metal and steam, but with differences of 10° to 20° it appeared that condensation took place at rates, in Mr. Bryan Donkin's and Colonel English's experiments, not differing notably from that inferred by the Authors. But, on .he other hand, it appeared anomalous that there should be this agreement when, as Fig. 17 would appear to show, the actual contact-differenoe of temperature between the steam and metal differed extremely from that between ihe general body of the steam and metal, and the " condensation area," measured in Fig. 17, from the dotted curve, would be about three times as large as that obtained from the full line. On comparing the period of heat absorption in Fig. 14, with the position occupied by the " condensation area " of the full line in Fig. 17, the time- temperature dotted ourve of whioh resembled that of Fig. 14, i( ' would appear that the " oondeujation area " actually oocnpied, as fairly as might be judged, the part of the revolution in which heat absorption took place ; and if the law of proportionality put forward were the true account of the matter, he would expect the results to agree with a condensation area estimated from the dottod curve of Fig. 17, rather than the other. There was, therefore, reason to suspect that the law put forward might be really the result of a compensation 3f errors, and would not hold so nearly true in other engines. But whatever might be the correct ex- > Proceedings of the InBtitntion of Meohanioal EngiBeers for November, 1896. i M ■■M <^P M. [Miautet of [Other variant, ryneas of the I a Talnahle >eratnre oyolea rge number of resent, limited thors' observa- ik, except that Mtsnring, with steam-engine, )ing hopelessly slasion of the le of the metal ratnre between rail, seemed to L Donkin's and 18 not dose for Mtween metal appeared that Donkin's and bly from that d, it appeared ben, as Fig. 17 of temperature u that between " oondensation Tonld be about On comparing sition ocoupied J. 17, the time- t of Fig. 14, a ly 'jconpied, as ition in which Drtionality put aid expect the rom the dotted nras, therefore, ) be really the hold so nearly he correct ex- Norember, 1888. Prooeedingi.] OOBBBBFOMDEMOS ON OOMDBHSATIOIf 09 SntM. 90 planation, he thought the presence of the layer of steam indi- ifr.FltiGtraldL cated by Fig. 17, must be regarded as a discovery of the Authors, and, unless it proved to have been a mere pocket of steam in the |-inoh hole referred to, a discovery of some importance. He did not think the presence of a layer of steam, with metal on one side aud steam on the other, both many degrees colder than itself, towards the end of the lead, could have been anticipated as probable, or, in the abeen'.^ of distinct evidence of its existence, even credible. He had examined a good many diagrams some years ago,^ and in all it appeared that, taking the steam on the whole as saturated up to cut-off, a large amount of oondensation must have occurred during admission, and no moderate amount of superheating would modify this result, while special experiments made on a large Corliss engine, in which the indicator showed only the part of the diagram belonging to compression on an extended horizontal scale, appeared to show that, reckoned as saturated steam, there was Uttlo or no change in the quantity of steam present during compression till the admission-valvis oponed ; the slight apparent variation of the weight of steam present being probalsly fully accounted for by superheating. There was thus evidence against the highly saperheated layer, fcund by the Authors in their engine, existing jaftsr admission began, as it was difficult to conceive how condensation could go on so long as it existed. Fig. 17 appeared to be a complete over- throw of all the usual expressions as to sudden oondensation on cold surfaces during lead and admission, and so forth, commonly regarded as almost axiomatic, and he would expect that it would with difficulty be accepted as a true record of facts. He suggested that oondensation might Cake place under and through a super- heated layer of the kind if the vapour near the ipetal surface consisted, like a gas, of molecules having velocities differing widely among themselves, through having a definite average energy, and that the surface of condensation selected the slow- ' moving molecules for absorption into itself in a marked manner. A room full of flies could be imagined, some active, others lazy, the walls papered with sticky fly-catcher, which captured the lazy flies near them so quickly that an observer near the wall would find that the flies in the air there were nearly all moving quickly. If the flies were considered as steam molecules, and the fly-catching pa^r ooli metal, the description applied ■ Beport of the British Aaaooiation for the Adyanoement of Science, 1888, p. 819. H 2 • vi 'I : -4 } 100 OOBBMPOKPWCTt OH OOHDKfSATlOT? OF BtlAM. XMlnnUi «fr-^-fe*i.-sj^^ STKAM. [Mioutea of aed to be that no uture layer next a oa wai in progress, )pear to Lave been on pp. 28-29, on the >anding steam, were le events whioh took things in whioh the perfectly, to one in -tension, but not so >ved from moleonlar and experimentally. y, that if drops were ize, somewhat larger ium, the work done r condensation might set free as to leave a Authors, unless the ad, with drops small might be sensibly ountered per second ire of the Authors, in ary for verifying the irface-range and heat i compared with the other points, like the It-iron used, rendered itigations as to valve- further pursuit, as T to tally with the rate, could be made oh heavier than any ecisive experience to ' metal could be made irfaces, and kinds of to draw in or squeeze I shown by the great bearings and end- or to dogmatize on the int, but the general i of lubrication, when consistent with the Proceedings.] 0OBRI8PONDBNOB OK OONDKNSATION OF BTBAM. 101 Prof. M. F. GuTERiiirni, of Darmstadt, thought the trials Prof Ontar- described in the Paper not only led to most important results, as °^^^^- to the interchange of heat between the cylinder-wall and «team in the experimental engine, but gave a practical basis for the judgment of the similar proceedings in other steam-engines under different circumstances. He regarded the experiments as a marked advance towards the explanation of the complicated process ^a the steam-cylinder when at work. Prof. 1). 8. Jacobus, of Hoboken, N.J., considered the approxi- Prof. Jwobu* mate formula, by whioh a condensation factor C had been derived, of special interest. A similar factor had been deduced for tests made by Prof. Denton and himself in 1888 on a simple engine of 17-inch bore and 80-inch stroke, running non-condensing. Any formula for the condensation must necessarily be approximate. In the Tables giving the results of their tests, a factor was ■ i u- deduced whioh represented the heat-equivalent of the conden- ■ation per stroke divided by the product of the time of admission the surface exposed, the point of cut-off, and the difference ot temperature between the steam admitted to the cylinder and thu steam at release.^ This factor had h^en suggested by Bankine, in discussing the famous experiments of Isherwood on the steamer "Michigan," which afforded the first systematic data for the appreciation of the important influence exerted by the phenomena of cylinder-condensation upon economy.^ The tests to whioh he referred were specially adapted for determining the constancy of such a ratio for a given engine, as the speed varied between 9 revolu- tions and 70 revolutions per minute, and the cut-off from 7 per cent, to 60 per cent, of the stroke ; the variation in HP. I aing from about 10 to 160. These factors were found to vary considerably and to be largest for a given cut-off when the speed was highest. In studying these figures, it seemed probable, that, if the conden- sation were assumed to be proportional to the time of admission raised to some power less than unity, the results wculd fall more in line with each other. This was equivalent to assuming that the rate of condensation at the instant that the steam struck the metal surface was greater than the rate of condensation at any Buooeeding interval and decreased as some power of the time. The results of all the experiments, except those in whioh the steam was throttled, and some special experiments in which there > Transaotioiu of the American Society of Mechanical Engineers, vol. x. p. 722. ' TransaoUooB of tho Institution of Engineen in Scotland, 1861-62. ■ " ' .^s-aatB^T'www- ^ wmm i,. 102 OOnlllSPOnDtirOB on OOUDBIBATIOlf or itlilf. [MtaatMof Pr«r. Jkeobti. wH a very low boiler-prawnre, w«r0 employed to determine the exponent of the time. Thia exponent was fonnd to be 0*68 in their testa, and the miaaing quantity in Britiah thermal nnita per atroke, divided by the prodaot of the temperature raised to this power, the aurfaoe in aquare feet to out-off and a range of temperature flrom admiaaion to the end of expanaion, waa found to be much more nearly oonatant than the aimple factor obtained by dividing by the time, aurfaoe and range. Designating the miaaing quantity in Britiah thermal units per stroke by Q, (.;e ' time of admiaaion by i, the exponent of 0-68 by y, the aurfaoe to the out-off by S, and the range of temperature from admission to release by R, the average results of Cie experimenta gave Q -f <* BR s 0'20. The higheat value of this ratio waa 0'41 and the lowest 0*22. A similar ratio wan then obtained for Willans' testa, where the low-presanre cylinder was run as a non-condensing simple engine, and the average of all the tests was found to be 0*81; the highest value being 0-4 and the lowest 0-18. This waa a remarkable agreement for two engines of such distinct types, the Willacs engine being of 14-iiioh bore and 6 inches stroke, running at 400 revolutions per minute : whereas the engine on which their tests T7ere made was of a much larger size and ran at a low speed. The factors for each of these tests were gi^en in col. 10, Table 1, and the factor 0, described in the Paper, in col. 12. An examination of Table I, p. 103. would show that the factor varied to a greater extent than the faokor given in ool. 10. The factors for Willans' steam-engine trials were given in Table II, p. 104, which showed that the same remarks applied to the constancy of the two factors, as in the case of their tests, the ratio of Q -j- C S R being more clearly constant than the factor C. If the pressure at release was greatly different from that of the atmosphere, the missing quantity divided by C S R might be considerably different from the value of 0*3 here deduced, which might be assumed to apply only to nnjacketed engines cutting off at less than 0*6 stroke and exhausting against atmospheric pressure. The time ,of admission was taken as the time required for the crank to move from 10°, before the dead- centre point to the point of out-off ; or, if a be the angle from the dead centre to the point of out-off and N the number of revolutions per Edinnte, the time of admission would be —^ — • The surface was taken as the olearanoe-surfaoe, and that portion of the surface of the cylinder and piston-rod exposed to the point of out-off. He noted that the leakage correction, given in Table YI, p. 34, was \ VriAM. [M iantM of to determine the tonnd to be 0-68 in ritiah thermal unite I temperature raiaed it-o£f and a range of [tnaion, waa fonnd to iple factor obtained Designating- the er btroke by Q, {.xe by y, the surface to tnre from admission e experiments gave 8 ratio was 0-41 and ibtained for Willans' as a non-condensing sts was fonnd to be lowest 0*18. This nes of snch distinct bore and 6 inches ainnte : whereas the M of a much larger >r each of these tests r 0, described in the ble I, p. 103. would itent than the faoior steam-engine trials that the same remarks 8 in the case of their riy constant than the reatly diflferent from ity divided by <* 8 B value of O'S here r only to nnjaoketed id exhausting against on was taken as the 10°, before the dead- be the angle from the lumber of revolutions -^ — . The surface portion of the surfaoe point of cut-off. He Table VI, p. 34, waa PkWXMdlaga.] OOBMBPOVDBNOIOM OONDntlTIOll 09 RTJUUI. 103 greater than the corrected feed-water used by the engine. Although Prof. JaoubM. the axperiments determining this leakage appeared to have been Tabli I. Simple •ngine— 17-iDoh bore ; >9|-inoh atroke ; loda 2 ^ inches and 2} inohna in diameter. Cle«nuioe inrfaoe b S'81 aq. feet. Average cleManoe rolnme s 6'8 per cent Hi MMoa — T Rang* i •s QntBtity. Bor* OfTMD- MMna Cnt- Revo* ImioM Ttm« ofAd- miMion fkM to Cni- uiriii 8. ^i^.-alnrt from Ad- OSB B.f.U. EqniTalenI ClMranm hctor 1 alt. Hlnutc. LIM. iCur. B.T.O. Stroks. t 11 mdi . mb"' t- nd miwlon K. U-M. SIrok* 08B. SorhM 81. 0, 1 t S 4 • 1 T 1 9 10 n M 1 0A99 70-88 600 72-8 0-264 18-48 40-4 220 0-83 .?-80 41 2 A99 61 001 070 680 80218-48 40 4 241 28 16 80 84 8 099 20-69 466 182-9 724 18-48 86 4 894 84 16 80 28 4 099 17-84 410 178-0 I 07818-48 34 6 489 80 16 80 24 099 12-68 408 284-1 1 47813-48 82 8 070 41 16 80 24 6 099 10-27 291 207-0 1 81118-48 82 7 660 81 16 80 17 7 099 8-9? 287 280-6 2 08818-48 88 8 789 82 16 80 17 8 818 87-CO 700 09-8 148 9-82 7P 7 214 28 14 88 48 8 ;o 818 61-80 672 80-4 210 9-82 n 207 81 14 88 40 10 818 08U2 606 82-9 224 9-82 70 6 269 81 14 88 44 11 818 28-00 479 128-8 400 9-82 72 6 417 30 14 83 82 12 818 1606 867 167 2 809 9-82 71 608 27 14 83 20 IS 818 18 07 829 184 8 998 9-82 66 7 653 28 14 83 22 14 818 8-68 270 229-6 1 006 9-82 61 8 795 29 14 83 18 10 182 8600 680 08-0 117 8-14 99 9 18i> 31 18 00 40 16 ie2 6410 644 74-8 108 814 99 1 230 32 13 90 46 17 182 60-88 480 08-8 166 8-14 98 9 237 25 18 90 80 18 « 182 09-94 078 70 7 168 814 98 8 239 30 13 90 41 19 182 09-76 648 79-6 169 8-14 97 9 288 83 13 90 47 80 C 182 20-62 800 101-2 894 814 92 6 400 20 18 90 26 21 182 28-66 870 114-0 427 8-14 91 9 420 27 13 90 27 88 182 17-80 804 128-0 084 8-14 89 9 508 25 13 90 22 23 182 18-80 270 148-8 709 8-14 87 090 20 13 90 19 84 182 1818 S»8 168-8 769 8-14 87 2 098 28 18 90 21 8fi 182 1800 299 169-0 777 8-14 88 2 600 28 18 90 22 26 182 1000 279 194-0 962 8-11 81 7 648 SO )S 90 20 27 182 9-98 249 188-1 1 012 814 90 7 744 20 13 90 18 88 182 8-64 266 220-9 1 169 8-14 84 760 30 18 90 19 29 •126 62-80 004 09-4 187 7-48109 210 28 13 50 37 80 |0 •126 60 07 486 09-2 148 7-43 108 213 28 13 50 36 87 099 62 01 482 07-6 29818-48 86 7 2!7 27 16 86 29 88 •818 6266 470 06-8 207 9-82 69 7 280 24 14 88 82 89 lO •182 0960 392 49 6 169 814 89 8 219 23 18 90 28 40 •126 60- 10 891 48-7 -143 748'102 1 202 24 13 00 29 41 0-069 69 00 298 88-0 0-110 670 110-7 1 171 0-22 13-10 23 I ^vc rage 0-29 Ave rage 30 \ b -J BMOHIi 104 OOBBEBPONPEMOE ON COMDENBATION OF STEAM. [MinntMof Prof. Jacobns. i Table II— VTiLLAiiaf BftMAM-Msanm Tbuul Simple engine — 14-inoh bore ; 6-inoh itroke. Gleamnoe BorfBoe = 2*77 iqnaro feet MlMlOg lUnge Qnantlty. Q (ifTem- Missing Cnt- Revo- luiion* Time ofAd- mlulon r«ce to Cnt- offln Sqnare Feet 8. peratora from Ad- fSR Qusn- tit; In B.f.0. EqnlTtlent Clearance Factor 1 off. Ili^at*. Lb«. ^nr. B.T.n. per Stroke. in Seconds mission to Knd ofEx- pension* R. y — 0-«8. Stroke + CSB. Snrftoe 8. a 1 S 8 4 6 6 7 8 9 10 11 18 1 0-804 898-a 82-0 8-270-0474'4-192 85 18-5 0-18 4-68 18 2 0-487 408-* 187-0 5-170-08794-106 52 28-1 0-22 4-58 80 8 0-88S 409-1220-6 8-170-088i;3-568 66 28-2 0-35 4-47 49 4 0-296 408-2il88-8 7-050-08148-466 77 25-4 0-28 4-45 42 fi 0-264 400-9,211-0 7 -86,0 -02988 -892 85 26-6 0-80 4-43 48 6 0-287 897-7278-9 10-26 0-0286 3-829 89 26-4 0-89 4-41 62 7 0-216 406-2258-4 9-44 3-0268 8-278 95 26-5 0-86 4-40 69 8 0-487 200-61 93-1 7-14 0-0771 4-106 48 84-5 0-21 4-53 21 9 ;0-8S9 205-2166-5 12-26 00660 8-568 62 35-0 0-85 4-47 87 10 0-264 223-0115-2 7-66 0-0537 3-392 83 38-6 0-20 4-43 26 11 0-216 223-7 261-8 14-34 0-0487 8-278 86 36-1 0-40 4-40 60 12 0-487 110-5 SS 8 11-97 0-1400 4106 46 49-7 0-24 4-5.S 19 18 0-839 112-7 14t> 19-86 0-1201 8-668 58 491 0-40 4-47 S» 14 0-264 122-8 144-0 17-62 0-0974 8-392 75 62-2 0-84 4-43 88 15 0-216 188-0 181-9 19-48 00790 8-278 8(3 50-2 Ayerage 0-80 4-40 41 0-81 Aventge 89 made with the greatest of skill, the results oonld not be as reliable as when the leakage was shown to amonnt to but a small frdotion of the total feed-water. The main difficulty in determining the leakage was that the water condensed on the surfaces formal a great portion of the escaping fluid, and it was hard to determine how this condensation eii'ect varied when the engine was run under a load. There was no necessity for having as large a leak in the valve as was found in the trials. In careful tests of the valve of a single non-condensing Ball-Wood engine of 10-inch bore and 80-inch stroke, the line used to determine the effect of compression on the water consumption, the leakage of the regular valve was found to be 10 lbs. per hour, and of a special valve, which was afterwards fitted to the engine, 2 lbs. per hour. This included all the leakage, both in the form of water and steam. ' Temperature at end of expansion taken aa the temperatnTO oonciponding to the pre88Ute at end of the alioku, as givou in Willaiui' Tublee. )r BTEAX. [Hinatetof fTbuia ih itroke. feet HMdk Qutn- SR titT In B.T.U. Equivalent Cletrince ractor •88. per Surfkoa a Struke B. + t>BB.. 9 10 11 18 8-fi 0-18 4-68 18 8-1 0-22 4-58 80 S-2 0-88 4-47 49 5-4 0-28 4-45 42 6-5 0-80 4-48 48 6-4 0-89 4-41 «2 16-5 0-86 4-40 A9 *-6 0-21 4-53 21 SO 0-85 4-47 «7 8-6 0-20 4-43 2ft «-l 040 4-40 60 9-7 0-24 4-58 19 9-1 0-40 4-47 88 52-2 0-84 4-43 88 W'2 0-80 4-40 41 age 0-31 Avenge 89 »nld not be as reliable ) but a small fraotion of y in determining the I the surfaces form>d was hard to determine 1 the engine was mn tiaving as large a leak [n careful tests of the od engine of 10-inoh etermine the effect of leakage of the regular id of a special valve, 2 lbs. per hour. This ' water and steam. empomtnre corresponding tab' TubloB. I>ruceeding&] OOBRIBBPONDBNOB ON 0OMOBM8ATION Or STBAH. 106 Prof. 0. H. Peabodt, of Boston, Mass., remarked that the ^raf. PeaWj. experiments detailed in the Paper gave the first real insight into those intricate interactions between the walls of the cylinder of an engine and the steam, which were first demonstrated quantitatively by Him. None except those who had made such tests could appreciate the labour, care, and skill shown by the Authors, who were to be congratulated on having shown by quantitative results that the phenomena of condensation and , re-evaporation were due to the action of the walls themselves and not to adhering water or any other cause. The fact that the steam in the cylinder during compression was highly superheated in this engine was not at all surprising, considering the large amount of compression and the high mean temperature of the walls of the cylinder. It was to be regretted that the tests were not accompanied by analyses by Hirn's method, showing the correspondence between the heat that had disappeared from the steam and the heat taken up by the walls of the cylinder. The ' same theory had been exhibited indirectly by the comparison of the actual and the calculated condensation by the Table on p. 34. One of the most startling and important results of the Paper was the direct leakage to the exhaust, which appeared to change the steam consumption in the diagram shown by Fig. 8 from 27 * 1 lbs. to 68 * 6 lbs. per HP. per hour. It could scarcely be conceded that snob a rate of leakage was unimportant even at the full speed of 250 revolutions per minute. If the power of the engine was ^ assumed to vary directly as the speed, then for the same cut-off the engine running at 260 revolutions donble-actinor would develop about 45 HP. and the leakage would still amount to more than 6 lbs. per HP. per hour, and the consumption would be increased to about 33 lbs. instead of 27 lbs. Comparison of piston and slide-valve engine with engine with separate steam- and exhaust-valves did not show such discrepancies, provided the ^ valves were in gdod condition. Thus Mr. Hoadley had found ^ 25*6 lbs. of steam oonsumption for an engine running at 125 revolutions under 120 lbs. per square inch steam-pressure and developing 80 HP., while Mr. Hill had found ' about 25 lbs. from three engines running at 75 revolutions under 96 lbs. steam- pressure. The first engine had a piston-valve controlled by a shaft governor, and the other engines had four vaivf>B arranged as on Corliss engines. On the other hand, it might be admitted ' " Thermodynamica of the Steam Engine," p. 266. » Ibid, p. 265. ^ ,* j 106 OOSBBSPONDBNOS ON 0ONDSN8A11OK or STKAK. [Hilinteiot Prof. Vtthoij. that high-speed engines with piston-valves or balanced valves often showed very poor economy in their use of steam. These tests showed very clearly that the condition of the surface exposed to the entering steam had maoh to do with the effect that surface had in producing condensation. On p. 10 it appeared that the fluctuation of the surface temperature of the head was 4° • 0, while that of the adjoining barrel surface was 13° '6. Again, on p. 22 it appeared that nearly equal surfaces at the side of the head and : the side of the oarrel were accredited with 79 thermal units and 123 thermal units per minute ; that was, the metal at the side was nearly one and a half times as active as that of the head. Now the surface of the barrel was polished and kept clean by the rubbing of the piston, wh:^8 that of the head was probably merely turned, and must have been more or less fouled with grease and dirt. Considering the surfaces of the ports, that were commonly cored out in the casting and were afterwards cleared from sand as well as might be by pickling or otherwise, was it not probable that they were less active than the surface of the head? It appeared probable that all investigators of condensation had over-estimated the influence of the clearance surface, unless these same surfaces vrere considered to be steam-jacketed; then they were assumed to be entirely innocuous. The conclusion that it was unnecessary to jacket the barrel of a cylinder provided the beads were jacketed did not follow from these tests. Such a conclusion required tests with and without steam in a jacketed engine, and such tests showed just the contrary. For example, three series of tests on the engine in the laboratory of the Massachusetts Institute of Technology,^ with steam in the jackets on both barrels and heads, with steam in the jackets on the heads only, and with steam in none of the jackets, showed that jacketing the heads only gave a gain of about 3 per cent., while jacketing both barrels and heads gave a gain of about 14 per cent. The Author reported that the indicators were tested as nearly as possible under the conditions of the trials, but did not give the probable errors of the indicators except as incidentally they concluded they were less than the difference between the temperatures of the steam as inferred from the indicator diagrams, and as measured by the platinum thermometer. The maximum error of the thermometer appeared to be not more than 0*6° and the discrepancies referred to appeared to be 2° or 3°, which at 300° F. corresponded to about as many lbs. per square inch— an error > TrsnBaotioua of tho American Society of Mechanical Engineers, Tol. xtL SIKAM. [MinntM of or balanced valvM of steam. These the surface exposed effect that surface appeared that the sad was 4° -3, while Again, on p. 22 de of the head and thermal units and etal at the side was the head. Now the ean by the rubbing My merely turned, th grease and dirt. ere commonly cored d from sand as well t probable that they [t appeared probable over-estimated the ) same surfaces were were assumed to be was unnecessary to heads were jacketed lusion required tests nd such tests showed ies of tests on the usetts Institute of n both barrels and ads only, and with jacketing the heads hile jacketing both r cent. The Author I nearly as possible ot give the probable illy they concluded n the temperatures >r diagrams, and as rhe maximum error 9 than 0-5° and the 3°, which at 300" F. aare inch — an error nl Eugineen, vol. xvL Prooeedinga] OOBBSSPONDKNOB ON OOMDmiBATIOH Of 8TBA1I. 107 which might fairly be attributed to the indicator unless that P»of. Paabody. instrument was tested by some method that was more reliable that those used by the makers. Mr. W. W. F. Pollen noticed that the Authors* piston platinum Mr. PolUn. thermometer was first used in a hole in the cylinder cover, and that the dotted curve. Fig. 17, represented the variation of temperature it indicated in that position. It would be interest- ing if the Authors could give data from which a drawing of the thermometer in the position stated, together with that portion of the cover in the immediate neighbourhood of the thermometer, could be constructed, as the indicated temperature of the steam varied so widely from that obtained when the thermometer was placed in the piston. It would appear that the steam in the tube surrounding the thermometer, when fixed in the cover, was highly superheated near the end of compression, while only a few degrees of superheat were noticed when the thenuomecer vraa fixed in the piston and occupied a similar position at the end of the stroke. It was not mentioned whether steam was admitted into the jacketed cylinder cover during any of the experiments, and what was the difference in the platinum thermometer curve when steam was and yas not admitted to the cover. He believed that a more extended description of the Authors' very beautiful platinum thermometer would be much appreciated. Professor Bobert H. SurrH desired to add his tribute of admira- Prof. Smith, tion of this Paper and the experimental work it described, which he regarded of the highest scientific character. He had long considered the method of thermo-electric measurement of the variation of cylinder temperature as the only method capable of giving accurate results, but it involved many instrumental obstacles. These difficulties were not even indicated in the Paper. Attempts to use the method had previously been made both at Harvard and at Sibley College, but without substantial success. The greatest merit of the present work lay in the fact that special experimental dexterity seemed to have overcome these obstacles. The electrical registration of a slowly changing temperature had for long been easy; the problem of accurately following and registering a very rapidly varying temperature was wholly different, and immensely more difficult. He hoped, therefore, the Authors would furnish a detailed account of their apparatus and mode of using it. The method of the "condensation area" appeared very successful. Its principle, that of finding the time- average of the difference of temperature between the steam and the metal, and multiplying this by the area of the metal surfiice ■"TsaNaWswiWifB^sss***' miiiiiii warn ]08 OORnsSPONDENOB ON CONDENSATION OF STBAM. [Minntet of 'L- Prof. Smith, exposed to this difference, was the same as that used in the final oaloulation he had given » in 1888; but at that time neither the average metal-surfaoe temperature nor its variations had been measured. The "limited rate of condensation " appeared to be due largely, if not wholly, to surface resistance to conduction of heat. In Peolet's and all other rules for the conduction of heat through boiler-plates, it was seen that the greater part of the resistance to conduction was credited to surface-resistance. In the transmission of every other kind of energy, viz., sound, light, Ac, the passage through the dividing surface between two portions of different kinds of material was obstructed so as to give rise to a difference of energy-potential on the two sides of the dividing surface; and it was only reasonable to suppose that a similar phenomenon occurred in the conduction of heat-energy, giving rise to a temperature-difference. He demurred to the Authors' statement that it had been hitherto universally supposed that the rate of condt sation of steam upon a metal surface of lower temperature was infinite. Engineers who, in considering the matter, had remembered Peclet's well-known results in heat- conduction from gases through metal plates to water, must have recognized that in steam-metal heat-conduction there was probably, »nd almost certainly, a surface-resistance which would make the condensation rate a strictly finite one, however great the steam- metal temperature-difference might be. In one set of experiments only did the Authors appear to have measured the temperature so close as ^hv ^^^ to the inside surface, and they did not use the results of these experiments in their subsequent work, relying solely on the readings at ,V inch and greater depths. This was disappointing, and it would be of great value if they would state what instrumental di£Boulties prevented the continuance of the measurements at ^^ inch. It would appear practicable to arrange the contact at this and even a less depth without a blow- through of steam, and he could not think of any other hindrance. . The result was that the Authors had to rely upoa a somewhat uncertain calculation of surface temperature from measurements taken so far away as ^ inch. On page 20, for Trial XIX, the Authors calculated a maximum surface temperature by adding half the range of 20° F. to the mean wall-temperature, 291° F., obtaining 801° F. But from the diagrams ^ven it would appear that the addition of seven-tenths to eight-tenths of the range to the mean would be a fairer calculation of the maximum. If three-fourths > Minutos of Proooddh ^i Inst. G.E., vol. zoiii. pp. 285-292. k- rEAH. [Hinntet of used in the final t time neither the iations had been ' appeared to be to conduction of Dndaction of heat -eater part of the oe-resistanoe. In ▼iz., sound, light, ice between two loted so as to give two sides of the to suppose that a >n of heat-energy, demurred to the iversally supposed metal surface of tio, in considering n results in heat- water, must have here was probably, 1 would make the r great the steam- set of experiments )he temperature so y did not use the ent work, relying lepths. This was ' they would state ontinuanoe of the ir practicable to h without a blow- y other hindrance, upoa a somewhat om measurements •r Trial XIX, the ire by adding half 291° F., obtaining I appear that the ange to the mean If three-fourths >p. 285-292. Prooeedingi.] OOBSBBPONVSNOE ON OOKDENSAHON OF 8TBAM. 109 instead of one-bilf of 20° F. were added to 291° F., since the result Prof. Smith, was to be subtracted from only 828° F., the resulting difference, namely, 27°F.,or only 22° F., would be largely affected. Begard- ing jFV^. t8, p. 25, the Authors staged that the small discrepancies between indicator and platinum-thermometer readings were so regular and consistent, Uiat they " cannot be attributed either to errors of the indicator or of the thermometer." This conclusion appeared based upon the vertical differences on the diagram. It would be noticed that horizontally the differences all corresponded closely with what might be due to the well-known horizontal displacement of the indicator-diagram due to the variation of the stretch of the indioator-oord. He had little doubt that the Authors were correct in attributing to valve-leakage a greater relative importance than has previously been given to it. If the valve-leakage depended on deformation of the faces and seats due more to unequal-temperature strains than to mechanical pressure, it would indicate great superiority in this respect in piston over fiat slide-valves, and the great importance would also be deduced of equal and opposite ports, so that the fiow of steam should be symmetrically divided, and equal temperatures, t;j far as may be, thus maintained. As the to-and-fro motion of the valve was so important in increasing leakage, it appeared strange that the speed of reciprocation should have no visible influence on it : this point seemed worthy of considerable further investigation. The conditions of heating of a single-acting cylinder were much less complex than those of a double-acting cylinder, and were, therefore, much more favourable to the reliable deduction of conclusions from experimental results ; but it might perhaps have been useful to emphasize more than the Authors had in the Paper, that the numerical results for double-acting cylinders must certainly differ very widely from those obtained by the Authors for single-acting cylinders. He thought, in the consideration of this subject, the Paper ^ contributed to the Institution by Messrs. Baraclough and Marks upon their extended experimental work under Professor Carpenter, at Sibley College, a highly important one. This work had been csjrried out most systematically, and the iPaper appeared not to have received the attention due to its scientific character and the amount of new information it afforded. Professor B. H. Thubston thought the Paper should be welco.^ed Prof.Tliaritoii. as the most complete statement as to the condensation of steam in the cylinder that had yet been presented to engineers. The > Hinntea of FNoeedinga Inai C.E., vol. oxx. p. 828. 4- ..ii ii .u w wiii,Wj"i iii M.'. HP liMiliiilillllillM^ Prof. p:,;':'. ^')-^ *,t.^ 110 OOBBBSPONDSNOI OH OONDKNgATIOH OV RnAM. [Minntetof ThvntoB. method had been admittedly important, bnt hitherto nnfmitfal in most instanooB. The apparatob used by the Anthon would seem to have proved itself exceptionally well fitted to the work; it certainly had been used by skilled and painstaking hands. All information relative to this form of engine waste had been derived hitherto through the study of the steam cycle, the pressare- volnme diagram, and the heat-supply. It long ago became obvious to the observer in this field that the temperature and heat-cycle of the metal of the engine cylinder must be observed accurately, as well as that of the steam, to give the needed data for a complete understanding and computation of the heat« wastes of the machine. Some 20 years ago Mr. A. A. Wilson, testing pumping-engines designed by him for the City of Ohioago, attempted to secure measurements of fluctuation of temperature of cylinder-wall by the use of thermometers, and, naturally, failed to obtain anything but a record of the mean for the engine-cycle.' Before the work of Professor Hall was begun, Mr. Churchill, in the Sibley College steam-engine laboratory, experimented for a long time upon the same subjerii, using the electric current as his measure of temperature of the metal, and gauging the resistance by a Wheatstone Bridg^. He found the temperature constant, apparently, at a depth beneath the interior surface of the cylinder- wall of 0* 238 inch, measured a fluctuation of 17° F. at 0* 115 inch, and of 25° F. at 0*0426 inch. The speed of the engine was 308 revolutions per minute. The load was light and the steam temperature at entrance was 300° F.' A number of other more or less similar investigations had been made; but none had given satisfactory determinations. Those reported in the Paper seemed far more perfect in their exhibition of the internal action of the engine than any hitherto made. It migLt prove interesting to compare with the diagrams here given, by these later investi- gators, those obtained in Sibley College laboratories some years ago; Mr. E. T. Adams, the investigator, using a curiously ingenious and strikingly effective system of photographically recording the temperature-oyde.' The method consisted, in brief, in the insertion of a thermo-couple at a measured depth into the cylinder-wall, thus following the fluctuations of tempera- ture of the metal. The current of the pile operated a galvano- meter of enormously high rate of vibration, and a mirror, 5 > "Hannal of the Bteam-Engine," put i. p. 493. i uCawie' s MaKazino," 1895, and Tranaaotioiu of the Aiaerioaa Society of Mechnuical £cgineer8, 1895, pp. 445-7. * Ibid. ■m'm^^i^s^i < IKAM . [Minntn of lerto nnAmitfal in ithon would Beem to the work; it ^ing hands. All had been derived ole, the preflsnre- long ago became temperature and mnst be observed give the needed ation of the heat- [r. A. A. Wilson, e City of Chicago, of temperature of laturally, failed to the engine-cycle.' . Churchill, in the aented for a long io current as his ing the resistance >erature constant, \oe of the qylinder- P P. at 0-116 inch, r the engine was bt and the steam ir of other more or t none had given the Paper seemed rnal action of the >ve interesting to lese later investi- tories some years ing a curiously photographically lod consisted, in i measured depth itions of tempera- erated a galvano- t, and a mirror, Ainerioan Society of ProoeediDKB.] OOBBBSPOMDSXOK OM 00VDKN8ATION Or flTKAM. Ill fixed upon the needle, reflected a ray of light npon a sensHive Prof.Thanton. plate compelled to move synchronously with the engine-piston. The diagram thus produced was the pressure-volume diagram, as taken with the "indicator," transformed into a temperature- volume diagram ; the 7 ^ir exhibiting the simultaneous variations of temperature and pressure, Fig$. Q8. Many months had been spent by Mr. Adams upon his experiments, and a large proportion of the time was given to the attempt to produce thermo-piles, at once delicate, accurate and durable. These investigations had now been in progress for several years, and their success had encouraged continuance of the work. The diagrams reproduced the changes due to heat-flow -j^ inch below the interior surface of the cylinder-wall. The particularly interesting point to be here noted was the singular and unexpected rise in temperature of the metal of the cylinder-head the instant after exhaust had occurred. Whether the diagram thus automatically produced F»g«. 38, was in any way misleading, or |^ whether it actually represented correctly a curious and unantici- pated phenomenon, it was evident that something occurred at that point the revelation of which was likely to throw light upon the-working of the engine in this respect. His own explanation was,' that the outrush of steam at the opening of the exhaust- valve cooled the face of the metal almost instantly to a minimum temperature; while, immediately after, the current having ceased to flow past with such exaggerated effect in refrigeration, the store of heat just under the surface, in Lhe cylinder-wall, at once raised the temperature again to a new maximum, the cooling taking place again more gradually. Corresponding portions of the steam and metal diagrams were similarly lettered. The dotted line from D" to C, Fig. 28, showed what, presumably, would have happened had the outflow of steam at exhaust been as slow and smooth as during the return stroke. The rapid cooling period was that evidently during which the water which might have condensed I •< Catriei'a Maf^zine," 1895, and Traniaotions of the Amerioan Society <^ Mechanioal Engineen, 1895, pp. 445-7. MSTON VNAVak riarON TMAVBk m I 112 COBnEgPOMDKNOB OM r. Eirach as his rkable disonasion od of variation ts waves as was ame form of tem- Fig$. 5, 6, et •eg.' oth by theoretical i have been anti- ith, however, the emperatare-range e depth of metal of research, with sating, and must eedingly obscnre. )»' investigations ■esults. The one tion between the r-wall in oontaot lent now largely 13d years should nal wastes of the Glasgow,' and that tly describe and aperature in the T him attributed. k,« M. Inst. O.E., rer 40 years since ies on stationary longen der Dampf- l889;aIaoThanton's Sou; and London, :, D. Appleton ft Oo^ forking of Steam in oal Bngineen, 1852, ' Proocodingt,] OOSBBBPONDSNOK OH OOmUKmhTlCIS OF 8TBAH. 113 engines,' and over 80 years since Isherwood,' repeated that work Prof.ThnntoB. on marine enginos. So di£9oult of observation and measurement had been these fluctuating temperatures of the cylinder-wall that all the resources of modem physical science had failed to reveal their precise method and the extent of their variation, until now, except through such academic studies as were made, first, by Professor Ootterill,' M. Inst. O.E., and, later, by Kirsoh, adopting what might, perhaps, be called the Fourierian system of investi- gation.3 Writing of Mr. Adams' remarkable research. Professor Dwelshauvers-Dery concluded*: "Ces diagrams oonfirment Ies id^es que Him a ^mises et que ^ n'ai ceswS de dSfendre. lis montrent I'importance qu'il faut attaoher & bien prot6ger le oylindre contre lea refroi- dissements; & en entretenir la haute temperature par le moyen d'enveloppes com- pletes & vapeur ; a diminuer autant que pop'-' jle Ies con- duits ; enfin & enduire celles de ces surfaces qui ne sont pas ezpos6es an frottement du piston d'une substance, q jeloonque, huile on vemis, qui arr^te la transinission de la chalenr et augmente la resistance de la couohe superfioielle du mStal, me- thode precouisee par le Professeur Thurston." As illustrating the latter action, thus advocated, the succeeding figures {Figs. 29) showed the results of similar experiments made by Professor A. Barnes, in the course of his work in. Sibley College, with the heat-meaKureinents taken at a greater depth within the cylinder-wall, and with a non-conducting coating employed to check the heat-waste. The diagram (a) was the steam-eugine indicator or pressure-volume diagram; (&) was, a time-temperature diagram with the cylinder- wall clean ; (e) showed the effect of simply oiling it ; while (<2) gave the results of coating it with a mixture of oil and (Of (d/) ■ " Thforie M^caniqne de Ohalenr," Paris, 1876. ' " Engineering Beaearohes," 2 vols., Philadelphia, 1863, 1865. » "The Steam-Engine," 1878. * Revue G^n^rale des Sciences, 1895, p. 773. [THK IM8T. C.K. VOL. CXXXI.] I ■i 114 OOBBUPOKDIKOa OM 00MDXM81TI0X OF BTIAX. [MinutM of Prof. Thunton. graphite. The moet effeotive ■ystem yot employed in (mrrying this proposal into effect wm that of leotiring first a poroiu raper- ficial coating and then filling this with a drjring oiL This had been found to reduce these heat-wastes very greatly, and to increase the economy of the machine some 10 per cent, or more.* This fact might, perhaps, throw some light upon the phenomenon referred to by the Authors, when induced to speculate upon the probable effect of a surface coating of grease. lie had found that, by simply wiping such a surface with a greasy cloth, and rubbing as clean as practicable, the surface-conductivity, or resistance, was modified 10 per cent.' The oils and greases were almost perfect non-conductors with respect to electrical currents; they were also effective interceptors of heai-onrrents. The Authors' conclusions and deductions from this admirable investigation seemed to him generally well sustained, both by his own and by other researches, of which tLey were in torn confirmatory. They computed a heat-transfer of 0*74 thermal units per second (44 thermal units per minute) per square foot per 1° F. tempera- ture range, measured between the steam and the cylinder-wall. His own experiments, of many years ago, g^ve one-third this quantity per degree of difference between steam and exhaust temperatures, which was not very far different, and Professor Ootterill's results, reduced to the same system of scale, would accord fairly well. The effect of wetness of steam was probably still an unsolved problem. Experiments at Sibley College, and those of others upon engines with low expansion-ratios, proved it to be un- important; but it remained to be determined what was the effect of the presence of water in the working steam, with large expansion-ratios and in large engines. It seemed to be the con- viction of engineers of experience that its presence was decidedly objectionable in such oases, on the score of deleterious effect upon economy, as well as for other and more obvious reasons. It was certain that the reverse, the superheating of the steam, even but a few deg^rees, had a decidedly good influence, and effects marked economy. The conclusions of the Authors relative to the effect of varying cut-off (page 45), must be read, having in view the fact that they resulted from experiments upon a small, simple, nn- economicoJ engine, and one in which a three-ported valve was ■ "The Finatl ImproTement of the Steam-Engine," TnuuMtione of the United Statei Naval IiuUtate, 1892. * Ibid. II mmnmm T lAM. [MlnutM of yed in oarrying t a porotui miper- ; oil. Thi« had greatly, and to ir cent, or more.* the phenomenon Boalate npon the a had found that, loth, and rnbbing by, or reeiBtanoe, ues were almoet il onrrents; they a. The AuthoTs' ble investigation by hia own and am oonflrmatory. units per second ^t 1° P. tempera- khe oylinder-wall. ,Te one-third this earn and exhaust nt, and Professor 1 of scale, would ' still an unsolved d those of others ved it to be un- ed what was the ; steam, with large led to be the oon- ence was deoidedly iteriouB effect upon as reasons. It was B steam, even but a and effects marked ttive to the effect of ig in view the fact small, simple, un- e-ported valve was " TnnMMstioiu of the Proeoodlntri.] OOBBXSPOVDINOI OH OOWDimATIOir OF BTIAM. 115 employed which complicated every case by its necessary variation Prof.Thutr»t<«. of the ratio of compression with that of the ratio of expansion. His own formula, there referred to, related to the more usual case of a fairly eoonor "al type of engine, with separated steam- and exhaust-valves and no compression, or none variable by the regu- lating mechanism. He should expect their proposed formula, how- ever, to prove a better expression for use in connection with the now common "automatic " type of small engine. Mr. A. L. Rice, of the Pratt Institute, Brooklyn, had, while at Sibley College, made a very complete examination of the various expressions proposed by investigators to date, and constructed the modified formula — „^ 0(r-I). in which o was usually about 0"6, but varied somewhat with pressures, and b was about 0-4, varying somewhat similarly, and was about 100, and of the values of which constants he gave the figures for a wide range of working conditions, and thus secured practical coincidence with all accepted experimental determinations. The average varied but 0*2 per cent., where Professor Thurston's formula, as employed for many years past, had an average variation of 2 per cent, or 3 per cent' He thought that engineers compelled to deal with such computations would oome to the adoption of the simplest form of his own measure, a = a i/r, finding values of a for the type and class of engine with which they had to deal, and avoiding all other complica- tions. He did not see how the action of leakage could be taken .' into account. It was purely accidental, without law, varying with different engines enormously, and with the same engine constantly. It might ' :i a serious tax on the engine to-day, and, the engine-driver discovering it, it might be checked and become nil to-morrow. So far as it occurred its effect was practically the same as that of equal initial condensation. The Authors were entitled to thanks for having thrown some light upon its method of ooourrence on the process of flow. The reference of the magni- tude of the heat-waste to the fluctuation of the now-ascertained temperatures of the cylinder-wall gave important and previously- lacking knowledge of the case. The next essential for the purposes of the designer and of the practitioner was a reliable system of reducing his computations to expressions of relation > TnmBSotioiu of the Amerioao Society of Mechanical Engineen, vol. xviii. p. 990. \ . I 2 i -5 116 OOBBEaPOKDXMOB OM OOKDnWATION Of ITIAll. [MiantM of Prof.T1ianton. between thRi T»rt*tioo »nd the Tari»tion of iteam temperatarea and preaaurea, in anoh manner that he might compute probable loaaea from the indicator diagram, or ita ideal oa laid down upon hia drawing-board. Lacking thii connecting link, he had been compelled to adhere to the original form of expreuion for thia waate, aa deduced many yean ago— a 'Jri Th« Anthon, in which the percentage — the fraction, rather — of oondenaation waa measured by a function of r the ratio of expansion, ( the time, in seconds, of 1 revolution, and i the diameter of the cylinder in inches; the values of the constant a, for such case, Iwing approximately 4 for the simple engine, unjacketed. External wastes were taken as 0-5 B.T.U. per square foot per hour per degree difference of temperature lietween ateam and surrounding air.' The whole subject was one of intense interest, as it was of serious practical importance, and the Authors' investigation must rank among the most famous, aa it unquestionably ranked among the most extensive, elaborate, ingenious and fruitful of ita class. Its prosecution and its outcome entitled the Authors to the thanks of the profession. They had entered an important and promising field of experimental work. The Authors, in reply to the Correspondence, regretted that, owing tc the lateness of the mail, they had had only 2 days to consider the points urged by the various writers. They hoped that they had not in consequence overlooked any important ques- tions. The performance of their engine had been severely criticised in respect of leakage, and it had been suggested that tlie correctness of their conclusions had thereby been seriously affected. They oould not insist too strongly that the fundamental advantage of their method of investigation, as compared with Hirn'e analysis, was that the results were entirely independent of leakage. It was also immaterial in their method, so far as the main conclusions were concerned, whether the engines were large or small, single or double-acting, condensing or non-condensing. The leakage of the Robb engine was not abnormal, but compared favourably with that of other engines. They had proved this by experiments under the conditions of running, and not merely by the usual stationary tests, which they had shown to be inadequate. It was ' "Promiie and Potency of High-rreseuro Steam," Trsnaootionk of the AuioricKU Hoc'icly uf Muclmnical Enginuvrs, vul. xviii., No. dccxviii. H«.^MiibikiiVkiyjUH««j)i«nflM««H«MMI««lW »U<7D [MtaatM of n t«inp«ntiirM mput« probable aid down npon k, he had bMO roaiioD for thii of condensation ezpanaion, t the r of the cylinder iioh oaae, being eted. External ot per hour per and rarronnding treat, as it was of vestigation must ly ranked among itful of ita class, kors to the thanks nt and promising e, regretted that, tid only 2 days to era. They hoped Y important ques- severely criticised lat tlie correotnesa icted. They could dvantage of their rn'a analysis, was leakage. It was main conclusions or small, single or 'he leakage of the favourably with I by experiments ely by the nsual adequate. It waa Tranaootions of the I. docxTiii. ProoeedinKi.J OOBaKSPOMDBNOH ON OOMOaHIATION Or iTIAM. 117 not fair to take a high-apeod engine, with a relatively large Th« Auth«ta. bulauoed valve, und work out its steam-oonsuiuptiou lingle-aotiug at one-quarter of ita normal apeed. Taking the Itobb engine, working at oiio-half out-ufT and 250 revolutioua per minnto, with Hteaip ut lOU lbs. (Msr sciuare inch presaure, as fairly reprosonting tho avorago praotioe in mills where auoh engines were uaed ; it would develop 00 I.IIR, double-acting-, non- condensing, with a ounsuiaptioa of alwut 85 lbs. per I.IIP. Only 3 lbs. per I.III'. would then be debited to leakage, which, aa they had atated in the I'liper, waa not an unreasonable proportion for su small and aimple a machine. It was incorrect for Prof. Fidler to anppoae that, if the Anthora' correction fur leakage were applied to the leaulta of Willana and other observers, the missing quantity would disappear wholly from the account. This atatement reated on a complete mia- apprehension, which it waa important to remove. The leakage was very different for different valves, but could be approximately estimated by the Authors* formula, p. 32, piovided that the fit were the same. For instance, the leak calculated for the valve of the Willans engine in trial 77-77, assuming the aame cloeeneaa of fit as in the Robb engine, would amount to between 20 lbs. and 30 lbs. out of a total missing quantity of 90 lbs. By actual experiment, following the Willana method, the Anthora had found much larger values of the leakage. They attributed thia to wear of the piaton-ringa, and to leakage of the cut-off ring. Much amaller valuea might no doubt be obtained by the uae of new and tight piston-rings under the conditions obtaining in a factory. That the leakage of a valve in motion was not always a purely accidental circumstance, without law, and constantly varying for the same valve, aa atated by Professor Thurston, might be inferred from the Authors' tests, which they believed to be the only measurements hitherto published under conditions which approxi- mated to thobe of actual running. They had recently re-determined the leakage of the Robb engine-valve in the same manner, after an interval of more than 2 years of constant use, daring which the valve had not been scraped or re-fitted, or the engine repaired in any manner. The results were in precise agreement with those given in the Paper. Him's analysis rested on the assumption that steam during expansion, and in rapid vertical motion, followed the kwa of statical equilibrium, whioh was known from other experiments not to be the case. It ^Iso depended on difficult and delicate -mm 118 OOBBBSPONUKNOB OM 0ONDEM8ATION 07 STEAM. [Mlnutoa of The Authois. ilieasarements of indicator diagrams, and assumed that the leakage at different points of the stroke was negligible. The Authors, therefore, considered that this method ooold not be applied with any reason ezcep^' to specially designed, large, low speed engines, like Hirn's, where the expansion was slow, and the leakage relatively small. Even in this case they considered that the leakage should be measured under the conditions of running, and that corrections should be applied for it as far as possible at different points of the stroke by some such method as that which they had indicated. Hirn's method of deducing the heat exchange from card measurements had been described by Professor Fidler, who proposed to go further and to differentiate these curves of heat flow. Considering the imperfect and fragmentary chturacter of the original curves, and the variety of possible sources of con- stant error, there was grave danger that the differentiated curves, if they did not depend upon debatable opinions, would afford wide scope for the exercise of taste in the matter of interpretation. '' It would be interesting to know whether such curves of -j- had been published, and what was the nature of the evidence to be derived from them that the rate of heat-transference did not depend on the temperature-difference. The Authors had themselves obtained decided evidence that the condensation and subsequent re-evaporation of wet steam produced a marked effect in the abstraction of heat from the walls. With regard to the experiments of Mr. Bryan Donkin and G«jlonel English on the condensation of steam by the surface- condenser method, it would have been a source of great satis- i'uotion to the Authors if they had been able to deduce from tbem such a confirmation of their law of condensation as that given by Mr. Maurice Fitzgerald. The conclusions of these able and careful experimentalists had, however, been of a different kind, and their method of calculation did not permit any conclusion to be fairly drawn in favour of the Authors' theory. In the opinion of Mr. Bryan Donkin and Colonel English, *' the results of their experiments showed that the film of water deposited by condensation and adherent to a metallic surface resists the transmission of heat in exactly the same way as an equivalent greater thickness of metal would do; so that the difl'erenoe of temperature between the steam and the actual outer surface of the metal would depend upon the mean thickness of the water-film and upon the rate of flow of heat through it " They had calculated the difference of temperature between tho steam and the surface rEAlC. [Mlutttos of that the leakage le. The Authors, not be applied large, low speed ', and the leakage isidered that the is of running, and it as possible at lod as that which he heat exchange Professor Fidler, « these curves of nentary character >le sources of con- erentiated ounres, would afford wide nturpretation. '' It . - dh , , jurves of -j- had a t he evidence to be Dsferenoe did not ors had themselves n and subsequent ked effect in the ryan Donkin and n by the surface- roe of great satis- e to deduce from densation as that elusions of these rever, been of a lid not permit any > Authors' theory, nel English, " the of water deposited urface resists the as an equivalent the difi'eronce of liter surface of the of the water-film bey had calculated a. and the surface ProoeediLgs.] OOBRBSPONDENOB ON CONDENSATION Or BTBIM. 119 of the metal by assuming a convenient ratio of the relative The Anthon. oondnotivities of water and metal. There was no evidence to justify such a procedure in the oase of a layer of drops of water in rapid motion, which could not be treated as if it were a uniform quiescent film of known oouduotivity. The differences of tempera- ture thus deduced could not, therefore, be used in support of a theory which was directly founded on the issumption of a dean metal sur&ce directly exposed to the steam, but differing from it in temperature, on account of the essentially finite rate of conden- sation of the steam itself. The Authors, recognising this difficulty, had taken pains to make a special serieo of experiments by the surface-condenser method,^ and had endeavoured to prove, by severe.) different lines of reasoning, that such a Ir.yer of waters drops ouuld not be treated as a quiescent film, that it did not oppose any material resistance to the passage of heat, and that it might even increase the amount of condensation in virtue of the greater surface exposed to the action of the steam. The remarkable superheating of the steam close to the walls during compression, as shown in Fig. 17, had been verified with five different thermometers, each of which had been independently calibrated. The piston thermometer did not show this effect to the same extent for several reasons. It was in motion, in a wide thin tulte, exposed to the air-jacket, and at a lower temperature than the rest of the cover. The cover-thermometer was confined in a narrow hole in the thickness of the cover itself, and the wire was at an average distance of less than ^V ^^°^ ^^m the walls. It would appear probable, considering the low conductivity of gases, and comparing the indications of the two thermometers, that this highly superheated surface-layer was very thin, especially on the more exposed surfaces. The Authors could not agree with Mr. Fitz- gerald that " it was difficult to conceive how condensation could go on so long as it existed." The superheat was a very small propor- tion of the total heat,' and it could easily be shown by experiment that superheated steam would condense on a metal surface very ,,''>"; nearly as fast as saturated steam, provided that the surface were below the condensing temperature. The other suggestion, made by Mr. Fitzgerald, that the dotted area in Fig. 17, showing the super- heat during compression, ought to be taken for the " condensation area," instead of the saturation curve, which the Authors had invariably made use of, appeared to be somewhat inconsistent '.ith his belief that superheated steam could not condense. The ' Report of the British Associatioa for the Advancemeat of Science, 1897. 'i 120 0OBBK8PONDEM0B ON OONDENBAIION OV STEAM. [Minutes of Tha Authors, adoption of this snggestion would also involve the assumption that steam, if sufficiently superheated, would condense on a surface above the saturation temperature, and that the rate of condensation of steam was, in any case, increased in direct proportion to tne degree of superheating. The Authors had not experimented on the leakage of stuffing- hoxes, but failed to see that the tightness of hydraulic apparatus at low temperatures and for excessively slow motion oould be urged as an objection to their views on steam-engine leakage. The attempt to make steam glands similarly tight, would either stop the engine, or make the piston rod red hot in a few minutes. The empirical formula so fully worked out and illustrated by Professor Jacobus, appeared to be capable of representing his own observations better than the Authors'. It would be more interesting, however, if it had a clearer theoretical foundation. It was not easy, for instance, to see why the condensation should be proportional to the range of temperature during expansion rather than to the whole range from admission to exhaust. The former supposition involved the aj arent anomaly that, if the steam were carried through the whole stroke, there would be no oondensatioa. A more serious point of difference lay in the effect of variation of speed. According to the results of Professor Jacobus, the missing quantity increased nearly in pro- portion to the square root of the speed. This might be explained on the Authors' hypothesis by supposing that the variation of speed involved a simultaneous variation in the quality of the ■ ' '' steam, which they believed to be a factor of great importance, as explained on p. 41. The diminution of the missing quantity with early cut-off, though somewhat greater than that given by the Authors' formula for the equivalent clearance surface, was in the same direction and of similar magnitude. The formula of Professor Jacobus appeared, however, to represent equally well (i.«., with a range of about 30 per cent, either way) the results of "Willans with a totally different engine, in whidi the missing quantity was nearly independent of the speed, and increased enormously, instead of diminishing, with an early cut-off. No dcubt, a part of this increase might reasonably be attributed to increase of leakage with higher pressure, but the conviction of Willans himself to the effect that water was always present in the cylinder was practically conclusive evidence, in the Authors' view, that the condensation in his trials was " limiting " and not " partial." In this case, as explained on p. 62, the condensation should be proportional to the whole range of the steam temperature, an assumption which [Miantea of he assumption ise on a surface )f condensation portion to the ge of stnffing- Aulio apparatus >tion oould be ngine t, would either a few minutes. illustrated by tpresenting his (Tonld be more sal foundation, e condensation erature during I admission to arent anomaly le stroke, there )f difference lay > the results of I nearly in pro- ht be explained ;he variation of quality of the ; importance, as g quantity with t. given by the 'face, was in the lula of Professor xrell (t.e., with a ults of T^illans issing quantity sed enormously, dcubt, a part of rease of leakage LS himself to the 'was practically ' be condensation In this case, as portional to the umption which Ptooeedings.] 00BBB8POKOSN0B W 0ONDKH8ATION OF STEAH. 121 evidently agreed more satisfactorily with the observations. This The Aathon. point had been stated with reference to the results given in the original Paper, but had been omitted in its abridgment. In the formula quoted by Professor Jacobus no allowance had apparently been made for the difference in duration of exposure of different parts of the surface. The portions of the barrel surface exposed for a very short time were reckoned as of equal importance with the clearance surface. This could hardly be justified theoretically on any view as to the nature of condensa- tion. In the Anthers' formula for the equivalent clearance surface allowance bad been very accurately made for the varying time of exposure of different part? of the barrel surface, without adding to the complexity of the formula, by the simple addition of the term /do (I = stroke, d = diameter, e = cut-off fraction), instead of writing vide, the whole area of the surface, as in the formula of Professor Jacobus. A corresponding formula had been given by Professor B. H. Smith in the "final calculation" to which he had referred,^ in which the portions of the barrel surface exposed up to the time of cut-off, allowing tor the time of exposure, had been represented by the expression — ( v/here A was the angle turned through by the crank. A Table was added giving the values of this function for different points of cut-off. On working out this formula it would be found that the results were in practical agreement with the simple expression proposed by the Authors up to one-half cut-off. Beyond that point the values of Ide were smaller than the true equivalent of the surfiace exposure. This was intended by the Authors to make approximate allowanae for the higher temperature of the parts of the barrel exposed during the latter half of the stroke in a double-acting engine. ' The formula proposed by Professor Smith did not, as he seemed to imply, make any reference whatever to temperature difference between the st^ >im and the metal, and could not, therefore, be considered as equivalent to the Authors' method of " condensation areas." No authorities were quoted by Professor Smith for the assumed "surface resistance" due to "difference of energy potential," which had led certain engineers to suppose that the rate of condensation was probably finite. On p. 290 of the article to I Minutu of Ftooeedingi Iiut O.B., toL xoiiL p. 291. Th« 122 OOBIUSFOMDIHOB OH OONBKNSATION OT 8T1AM. [Minntea of Authon. whioh he referred he had himself oalonlated the thickness of metal whioh mast be raised firom exhaust to admission temperature to aooount for the missing quantity. This appeared to imply that the surface temperature followed the steam cyde, whioh would require the infinite rate generally assumed ; and there was certainly no hint expressed of any limit, at least, in the pages referred to, which chiefly related to quite different matters. In the Authors' view it was not necessary to assume any discontinuity in the temperature at the dividing surface. The layer of steam in actual contact with the metal would be at the same temperature as the surface of the metal below its saturation point, and would be condensing on the surface at a rate proportional to its defect of temperature. It was well known that steam could exist at temperatures below that of saturation without condensing, unless there were some surface to condense on. There would be a continuous tempth as giving iter readings. I diminish the lie introduction )int8 had been and the depth I, owing to the pened that the from obeerva- . 20, for trial range to the wrreot, but it lent. In the ) draughtsman to the right, he propagation ) cover surface Pioouedlnga.] OOBBESPOMOKMCni ON OOBTDBnUTIOll OV BTBAM. 123 was responsible for the great difference in the ranges of the cover Th* Anthon. and side cycles as suggested by Professor Peabody. Junction No. 1 side was opposite the oonnterbore, and was less dean than the cover. The difference of cycle range appeared to be simply due to the much lower temperature of the side, caused by longitudinal conduction, as explained on p. 13, and was exactly accounted for by the increase of condensation area on the Authors' theory. Side jackets were doubtless an advantage, especially in long narrow cylinders, but the effect might be partly attributed to diminution of leakage. The Authors had found that a rise of temperature of only 20° F. in the metal of the valve^eat diminished the leakage of the Bobb engine by more than 80 per cent. The Authors wished to take this opportunity of directing attention to the case of "limiting" condensation explained on p. 36. They considered this to be a most important deduction from their theory, but it had so far remained entirely unnoticed, both in the discussion, and also by their numerous correspondents. A great deal of space bad been devoted to the discussion of purely empirical formulas, but the Authors had not wished to attach any importance to their approximate expressions, as compared with the method of the temperatnre^de diagram, which was founded on a definite and rational theory. The existence of this " condensation limit" would be found, when proper data were available, to explain many facts with regard to cylinder condensation which bad hitherto been regarded as anomalous. Up to this limit, for instance, the effect of wetness of the steam might be very great ; beyond this point there was practically no increase of condensation. This would explain the conflict of opinion referred to by Professor Thurston. Some consideration of this kind might also be found to give the key to the very puzzling observations of Messrs. Barradough and Marks, mentioned by Professor Smith. These observers had apparently found different laws of variation of condensation as dependent on speed at different initial steam- pressures. The Authors regretted that they were unable to give a more extended description of their platinum thermometers, as requested by Mr. Pullon, without unduly increasing the length of their reply. The instrument involved two or three novel points of construction, which had been briefly referred to in the Paper as originally communicated, but had been exdsed as bdng of too technical a character. It might, perhaps, be of interest to state that they had recently succeeded in making this thermometer draw the curve of the steam oyole automatically with pen and ink 124 COBBBSPONDEMOB ON OOMDSNBATION OF STEAM. [Minutes of The AutLon. on a revolving drnm. The arrangement was snoh that the effects of inertia and friction were absolutely eliminated. It took abont a quarter of an hour to describe a complete curve, during which time the engine had to be kept running steadily. The apparatus could be looked up and left by itself to write oyole-curves all day if desired, and did not require any interference or any special skill on the part of the observer. It had the advantage of producing a perfectly oontinuons curve on a larger scale than an ordinary indicator. A complete description of the apparatus, which was much simpler than might be supposed, would be published as soon as the Authors could find leisure for the work. They intended to make a comparison of the curves so obtained under a variety of different conditions with those deduced from the indicator diagram. They might hope in this manner to obtain more definite explanation of the discrepancies they had already observed between the two instrumenta. The errors of the two indicators chiefly employed in the trials were found to be 3 per cent, too high and 3 per cent, too low respectively. If they had act been corrected thoir errors would have been of the order suggested by Professor Peabody; but as they were tested hot simultaneously on the same steam against the identical platinum thermometers with which their readings were compared, it is not likely that their residual errors can have been of such a magni- tude. The variation of stretch of the indicator oord, as ordinarily observed, would have the opposite effect to that suggested by Professor Smith. In any oase, such variation, if it existed, was far too small in the present instance to account for the observed effects. It was, perhaps, also necessary to remark that accurate account had been taken of the obliquity of the connecting-rod in reducing the indicator diagrams to temperature cycle-curves. It appeared probable that there were real differences of temperature existing simultaneously in different parts of the cylinder, and that the temperature of the steam might fall considerably below the saturation point during rapid expansioi ADDENDUM. Pbess Abstbaot. At the Ordinary Meeting on Tuesday, the 80th NoTembe.', Mr. J. Olarke Hawkshaw, Member of Gonncil, in the Chair, the Paper read wag on "The Law of Oondensation of Steam," by Messrs. Hugh L. Gallendar, MA., and John T. Nioolson, B.So. In the disouBsion of steam-engine trials it had generally been assumed that smm [MinuloM of t the effects took about ring which e apparatus rves all day special skill producing a an ordinary , which was published as rork. They tained under ted from the ler to obtain had already 8 of the two to be 3 per If they had of the order :e tested hut ioal platinum ired, it is not uoh a magni- I as ordinarily suggested by b existed, was r the obeerred that aoourate teoting-rod in )le-ourves. It f temperature ader, and that bly below the ., Mr. J. Clarke ul was on "The indar, M.A., and len aaeumed that Proceedings.] THE LAW OF CONDBNBVnOH Ot STEAM. 125 the rate of condeniation of iteam on a nurfaoe was praotioallj infinite, so that any inrfaoe in direct contact with the iteam waa immediately heated to the ■atnration tempemtare correiponding with the preMuie of the steam. It liad aim been snppoeed that the amount of condensation nnder any giren oonditians was limited, either by the resistance of the fllm of condensed water to thvi passage of heat, or by the capacity of the metal or of the oironlating water to carry off the heat. Is many oases condensation was diminished by films of oil or grease, or by accnmnlstions of air, or by other inomstations or depodta, but these were not considered in the Paper. The Authors found, on the contrary, as the result of their experimeuts on a steam-engine running nnder normal conditions, that a praotioally clean and dry metal surface was not immediately heated to ttie temperature of the saturated steam in contact with it, that the rate of condensation of steam was not infinite, but finite and measurable, and that the smount of condensation in any given case was limited chiefly by this finite rate of oondeosation, and ooold be calcu- lated in terms of it The oynlioal Tuiations of temperature in the metallic walls of the cylinder, with each stroke of the engine, were measured by means of thermo-couples inserted at various distances from the inner surface. It was possible thus to deduce the amount of heat absorbed and given out by the metal, and to infer the quantity of sterm condensed and re-evaporated at different points of the stroke. The temperature-oyoles of the steam were simultaneously measured by a very sen:itive platinum thermometer. The observationB showed that the temperature of the steam in different parts of the cylinder differed in a systematic way from the saturation temperature as deduced flrom indicator diagrama In order to deduce the condensation firom the observed tempeiatnre-cyolea, it waa necessary to determine the conductivity and specific heat of cast iron. A series of experiments were made upon a 4-inoh bar of cast iron, and the result found for the conductivity was nearly 80 per cent, smaller than that generally aasur od. At the lowest speed of the experiments, namely, 45 revolutions per minute, the temperature of the surface of the metal at the end of the admission period was found to be never raised higher than within 20° F. of the temperature of the steam, and the rate of condeniiation at any moment was simply proportional to the difference between the temperature of the steam and tlie surface. The numerical value foond for the rate of condensation was 0*74 B.T.U. per second per square foot of surface per degree F. of difference between the temperature of the steam and the surface. This waa equivalent to the condenBation of 27 lbs. of steam per square foot per hour at 800° F., for a difference of temperature of 1U° F. Assuming this law, the total amount of condensation at any point of the stroke conld be inferred by measuring the "condensation areas" on the temperature-cycle diagram, t,e., the areas included between the curves repre- senting the temperature of the steaia and of the metal aurfaoe. To compare the results thus found with the missing steam deduced from the indicator diatj-rams and the feed measurements, the leakage of the valve and piston was determined as nearly as possible under the conditions of running. It was found to be proportional to the difference of pressure and nearly independent of the speed through a considerable range. The usual test for leakage with the valve stationary was found to be of little or no value. From a comparison of leakage tests, it was inferred that a valve in motion, however well fitted, was subject to leakage of a definite type. The leAkage took place chiefly in the form of water, by condensation and re-evaporation on the moving snrincos, and TBI LAW or OOKDlNBAnOK OV 8TB41C. [MinntM of waa directly proportionftl to the perimeter of the ports and liiTei«eIy to the width of the bearing rarfMee. The unonnt of oondeiiMtion obeenred during the admiuion period in a ainglenwtiog noD-oondenaiag cylinder 10 5 inohea in diaaaeter with a atroke of 12 inohea, waa only %0 per oeni of the feed at a apeed of 100 revolutinna per minnte. The amallnMa of thin leanlt waa probably dor to the early compreaaion and the dryneaa of the ateam anpply. It was ftnind that ro'^raporation waa completed rery qnlokly, and that the walla were dry for the g^reater part of the cycle. It waa inferred from the form of the tempeiatare onrvea and from other evidence that the rate of re-erapontian waa the aame aa that of condenaation. From the form of the law of condenaation it waa poaalble to make an Important theoretical deduction with regard to oaaea in which re-eTaporation waa incom- plete, and the walla remained wet throughout the whole cycle. Urder tbf ae oonditicna the mean temperature of tho walla ahonld be the aame aa the time average of the temperature of tho ateam to whioh they were ezpoaed, and the cyclical condenaation waa the maximum poaaihle for the given ateam cycln. If the extent of the clearance aurfaoea wm known, thia limiting value of tho condenaation in any oaae might be easily deduced flrom the indicator diagram. If the aurfaoea were dry during port of the atroke, the condenaation waa leA than the limit, and it waa neceaaary to know the mean temperature of the dearanoe aurfaoea in addition. Upon theae viewa of the nature of ocndenaation and leakage, the miaaing quantity of ateam W in pounda per hour might be expreaaed by an equation of the general type, Wa B (f- 1°) + L (j>'— p"), — where the firat term repreaented condenaaUon and the aenond term leakage, 8 being the equivalent clearance surface in aqnare feet, and f — (P the mean difference of temperature, in degreea Fahrenheit, between the walla and the steam during admisidon reduced to one-half cui^iff. L, the rate of leakage per lb. difference of preasure p'— p", might be taken to vary approximately as the product of the diameter and the square root of the normal piston-speed, for engines of differeit sixes. It wild appear ^m thia formula that the effect of leakage on tho f arformance was relatively more important in amall engines and at high preaaui"*, and that the leas due to condensation waa moat effectively reduced by increase of piston-speed. As an indiieot verification of thia law of condensation, the temperature of the clearance snifaoe in oaaea in which water waa present in the cylinder waa measured, and waa found to agree with that of Uie mean of the atecm cycle. The amount of oondensatlon waa alao correctly calculated in aeveral cases of published tests in whieh sulBoient data were available. The rate of oendensa- tion deduced was alao direetly verified by an mtirely different method. The experiments gave approximately the same rate of condensation, and appeared to show that the water-dropa condensed en the metallic surface, owing probably to their rapid action, did not appreoiably dimiuiah the rate. Assuming it possible to estimate the condensation ococrring in any given case by the method indi- cated, from a knowledge of the. indicator diagram aiid of the temperature and area of the clearance aurfaoes, it then became possible to determine the amount of leakage under the actual conditions of running. I '^'^•^tfgggif*^ " • ' '""7" [MinntM of ranely to the •enred during 10-5 inohM in !iBed ftt a tpeed I probably don It was ftnind Is were dry for to tempetatora aa the Mine m B an important on was inoom- Urder these le aa the time poaed, and the earn oyolfi. If ; value of the oator diagram, ation waa le* erature of the >f ocndenwtion lonr might be L(|.'-j."),- term leakage^ — tP the mean walla and the of leakage per imately as the •ton-apeed, for at the effect of dl engines and lost effectively peiatora of the » cylinder waa le steum oyole. everal oases of e of oendensa- method. The nd appeared to [ng probably to dng it possible a method indi- imperatnre and iue the amonnt Praoeediagi.] BLSOnom, KTO. m 7 December, 1897. Sir JOHN WOLFE BARBY, K.O.B., LL.D., F.R.S., President, in the Ohair. It waa announced that the Aasooiate Members hereunder men- tioned had been transferred to the dase of ASMOSft* WAi/rm DraoAH Babbow. Bdwabd Skkltoh Boulaus. JOHH ALBXAHDm BRODIB. OSBBBT CRADWIOK, O.M.O. WlUJAM GBATHAM. Eownr Krwos Oiabb, M.A. (Oairfab.) Hbmbt Tjffiho Obooe. William Dtaok. GBABLBS WttUAU VsuM Fabbwbix. Thomas Obiititbb. Thomas Habbt Houobtob. WiLUAM HVBCT. Isaac Mattbbws Johbb. Datid Hiohabl Litstbb. Abobibalo Thomas Haokbhiib. Jambs Mbldbcm. WlLUAM OTMJOH MHWTOM. Oau>bb HPKon Olivbb. PHOir Hbhbt Falmbb. WnXIAM AliBXABOBB FaTBBSOB. Johh Pbiob. Lbwis Hbhbt Babsomb. Albbbt Bdwabo SnuE. Datid Bimsoh. ' JOHABB PHIUF EdMOBD ChABLHI Stbombtbb. JOBBPH TTSOB. Thomas Dvboab Wbib, B.Sc. (0Im.) And that the foUowing Candidates had been admitted as BiudmUt. OBoa Babbbb. HABDIBaTOH Abtbub Babturt. JO0BCA BAOABDOn BBBJAMIB. Habold Mbbbuztb Kna Bbbbdob. Habbt Fowlbb Bioen, BJl. (floMb.) WnUAM HmBBBT Blakbb, Bdwabo Pai* Botbt, B.Bo. (MeOiO). Hbbbt Samubi. Boobbs Botajiab. Bobebt Jambs Botd. Fbbdbbiok Bbiohtob^ HooB Bbodhcbst, BJl. (Gontob.) Obobob Fbbdbbio Booki^bd. William Hbbbt Ohambbbs Bvllbh. Pbibb Maitbiob Stbwabt Gabmiohabl. Oboil Bbbtbam Gabb. Kbnhbth Mdbbat Ghadwiok. Obobob Plubbvit GbaHiIB. Biohabd Eluoit Glabxb. WiLUAM OOLLIB& William Oborob Golqctoiib. ABannm BAOinoBD Obaddook, B.Sc Gbbistofhbb Dabbt. JOHB NaBSOAWBB DAWB. GaoBOB Wblbbt Dbakib. Oboab Hbbbt Dbsbbbb. Edwabd Aluob Dovolas. Abthiib HoDocoall Duokham. Abthcb Btdbbt WBLrai Eldbb. GLAin» Vttiab Abmit Bspkut. Hbbbbbt Waltbb Fm Simohb. TcDOB Datid Folebs. 128 ■LwrnoKB, mo. £KiMlmlt— oontinned. [If inntM of Rnmu. Oeoroi Bla«i Fobdham. Hkuirt WaLUM Frank. Tromai Edwabd AKDfBfcni Olia DOWB. Omab THtoporai GHOMFiairi. Hbrbkbt Wood Hahbobt. Edwabd Staotjit HAnniAii. Bbxbr Yocmo Habbimr. Edwabd Cicil QuiXAHO Hbwkioiibb. UOOBB OABKBLI. HBTHBBHtOTOB, B.A. (Cantab.) HniBT Thomab Hildaob. UaKOU) COMLim BlLTOX. Jambb Uabbt Holudat. Adriam Jambs Bobbvt Hon. ebhbvr hodoboh howat««. Fbederiok Nobl Hvdww. Malcolm SuTMBBiiAin) Jaoomb-Hood. William Jamimom. Etbltm LurwBXTii HotrLaB Jona, B.A. (Oxon.) JOBM Abthub KmoHT. Wallaob SnoKLAiiD Lakb. JOHH MOBTAOCB ElXIB liABOTOK, B.A. (OohUA.) Albbvt ABOiwroni Lba», B.A. (Cantab.) Harold Brook Lbabotd. Jobb Biobardsoii Lino. Athol Lookbt. OwBM Lbohabd MaoDbbmott. WnxuM Matthbwb Uao1!'adtba>. Bahald Maobab. Btuabt William Buchabah Mao- Obboob. Oabdibbb Hbbdbbkw Maokolop, B8o.(Oto.) John Maooatib Maolsab. Joseph Mallaob, B.B. (Hoyol). Abthob Wallace MtMnx. Charles de Fatb Mbsbbbot. HbRRV WiLUAM MlBHlTT. Francis Hembt Niobolbob. Btavlbt Pabkibsob. Th« Hon. OtoFPRBT Lawbbnob Tab- sorb, B.A. (OgpoN.) Jambs Patebsob. Nbville Leckobbt Phipps. Hebbebt Jambs BiaoBAM Powell. MOBMAB REXD. Hbbbebt Libdslbt Riohabdsob. Fbeoebiok 8tbbll Bobbbtbob. William Hcbebt Bobihbon. OsoAB Fbidolp Alkxabder BANDBSna. BOBBET ABTHUB BABOE. Basil James Bbable. LOVn HiLOBOTB Bewbll. Rboihald Bbaeplbt. Abtbub Oswald Sbbbbbn. BoBERT Jobb Simpson. Bobbbt Jobb Josbpb Bloab. Jobb Kibbt Bwalbs. Edwabd Wautbb Taplib. Pbbot Tatlob. Lbobabd Templb. Pbbot Thomas, B.Bo. (FWoHo.) NOBMAB ABTHUB THOMPSON, B.A. (OiMteb.) Bboibald Abthcb Walcot Thomson. Ebbest Tbobntob. Warbbbbe Beaumont Tbblawbt. Oltvbb Bowlabd Walkbt. William Watbbb. Oeobob WmoHAM. Ebbbbx Clabk "Whub, BBo. (Ffa- torta.) OhABLBS THAOKBBAT WILB'BAHAM. Josbpb Ernest Wilkes. Waltbb Oobdob Wilsob. Habbbb Bobbbt Toobo. The Candidates balloied for and dnly elected were : aa JHsmbtrs. HiatJEL DB Teitb b Aboollo. Patrick Fobstall Oohbbb. John Henby Dabbt. FbKDBHIOK BbBNABD D«EBIBfl. John William Dobman, B.A.(I)«AKn.) Abgcibald Oampbbu. Elliott, D Jo. iBdin.) Bobbbt Gillbam. Oboboe Stlvestbb Gbimbtoh. . WiLPRiD Jambs Linbham. T [If inntM of mnmoT. IKITT. lOUOM. Lawmnoi Par- Pmrw. SBAM PoWBiU BlIOHARDflOM. toBiimoii. OMNaOM. LAHDBR BaITOBKBO. ROB. L inuu r. ■IBBH. OR. B Bloak. 1. lAnot. le. (FMorio.) Tbomhoh, B.A. WaIiOOT THrniBOH. >iiT TmLAWirT. UTalkit. rmri, BBo. (Fto- AT WlUUAHAM. rooHO. ere: as nu. BtLiorr, D.Bo. B GBIMBTOlf . iIMBHAM. Prooeedlnp] ■LKOnOMB, XTO. 120 Jambi Mactbab. Ohablbs William Eablb Mabbr. Thomas Josbph Mtlbi, BE. (DubUn) Hb^bt Johx Obam. EOWUI WiLBCB BiCB, Jnn. Mmnktn — continned. William AcoBrrri Riobabomx. William Edwabo Rilbt. FbBDBBIOX HbBBT BMALb Nabob Bouami. JOHM FlXDLBY WaLLAOB. AttodaU MmiMft. HuoH Damibl Badooob, M.A. ((Taon.) Obobob JoorH Bill. Altbbd Rowb Bbllamt. Habold Bbbbidob, Btod. Inat. O.S. JoHK Bhobb. WiLLUM Bbowv, B.Bo. {Edin.), Stud. Xiut. O.B. BioHABD Jamb* Ohuboh. LlBDBIAT COLTU ClABX, H.O.B. (JW6.) Bbbbit Obay Oootti. Altbbd Dotbb Dblap, B.B. (DMbtAi.) Hbhbaob Wnnra Fibob, M.A. (Owm.) Hbbbbbt Pauxm Flbtohbb, Btnd. hat C.E. Bbnbr BuroLPH For. JosBPH Gabtiblo. Fbbdbbick WnXUM GODrBBT. WiixiamHall. Gbobob Walkbb Hbbdmax, M.A., B.80. iBdin.) ALrBBU Edwabd Hubib. William Samobl Jombs. Mabttx Hamiltor Kiloour. Bbdfobd If oNbill. Jambs Wblbt Madblbt, M.A. (Con- Uib.) WiLUAM Hbnrt Moobbt, H.Bo. (FMoKa), Btnd. Inst O.E. JohbMuib. Gbobob Axdbxw Nobtboboft. William Hbmrt Fbbttt. joibpb tobktn rodda. WiLUAM Hbrbbbt Bbiblm, B.8«. (OUu.) Datid Gcillabd Tatlob, B.80. (ObM.) Waliu ABTBim Valoh. Stud. Inst. O.E. WiLUAM HBBBKBT WrITB. Gbobob Bbahsbt Williams, Btnd. InstCB. JLtrnteiaU*. Cbablh Albrxd Gbippb, Q.O., M.P. MiocBi. Bibbido LisbOa, late Captain BratOiam Navy. JoHR Pbiob. Jambs Bailtor, Jan. Edwabd Pbbot BimpsoM, H.A. iOxon.) The diwnusion upon " The Law of Condensauon of Steam," by Meosra. Callendar and Nioolson, was oontintied and conoladeu. [tHK INST. O.R. VOL. CXXXI.] I n n f ^i i Mii i niii i i.-" !'' ■.■I ll II I I I T— — I LONDON) PMNTED BY WILUAM CL0WK8 AND SONS, Liwrm, RAMvuHD nuua Alio cu*>na caoM. Ji" I'D/I J . PLATR IfinutaA of Proossdin^ ofThe Institutini of Ctvil En|me«rs. Vol. CXXXl . Seaiion 1891-3S Arti THO* KKr^L A SOK.LrTH.LOMDON. 4'e-*S.'«£'*3«ss5- ■