qmVEBSITY OF CALIFORKIA PUBLICATIONS COLLEGE OF AGRICULTURE AGRICULTURAL EXPERIMENT STATION BERKELEY. CALIFORNIA DUST AND THE TRACTOR ENGINE BY A. H. HOFFMAN BULLETIN No. 362 May, 1923 UNIVERSITY OF CALIFORNIA PRESS BERKELEY, CALIFORNIA 1923 David P. Barrows, President of the University. EXPERIMENT STATION STAFF HEADS OF DIVISIONS Thomas Forsyth Hunt, Dean. Edward J. Wickson, Horticulture (Emeritus). , Director of Resident Instruction. C. M. Haring, Veterinary Science, Director of Agricultural Experiment Station. B. H. Crocheron, Director of Agricultural Extension. C. B. Hutchison, Plant Breeding, Director of the Branch of the College of Agriculture at Davis. H. J. Webber, Sub-tropical Horticulture, Director of Citrus Experiment Station. William A. Setchell, Botany. Myer E. Jaffa, Nutrition. Ralph E. Smith, Plant Pathology. John W. Gilmore, Agronomy. Charles F. Shaw, Soil Technology. John W. Gregg, Landscape Gardening and Floriculture. Frederic T. Bioletti, Viticulture and Fruit Products. Warren T. Clarke, Agricultural Extension. Ernest B. Babcock, Genetics. Gorpon H. True, Animal Husbandry. Walter Mulford, Forestry. James T. Barrett, Plant Pathology. W. P. Kelley, Agricultural Chemistry. H. J. Quayle, Entomology. Elwood Mead, Rural Institutions. H. S. Reed, Plant Physiology. L. D. Batchelor, Orchard Management. W. L. Howard, Pomology. IFrank Adams, Irrigation Investigations. C. L. Roadhouse, Dairy Industry. R. L. Adams, Farm Management. W. B. Herms, Entomology and Parasitology. John E. Dougherty, Poultry Husbandry. D. R. Hoagland, Plant Nutrition. G. H. Hart, Veterinary Science. L. J. Fletcher, Agricultural Engineering. Edwin C. Voorhies, Assistant to the Dean. . DIVISION OF AGRICULTURAL ENGINEERING L. J. Fletcher James Koeber A. H. Hoffman E. G. McKibben B. D. Moses J. D. Long R. R. Thomson W. L. Zink H. L. B ELTON t In cooperation with Office of Public Roads and Rural Engineering, U. S. Department of Agriculture. DUST AND THE TRACTOR ENGINE 1 By A. H. HOFFMAN CONTENTS Page Dust Causes Rapid Wear 470 Efficiency of Air Cleaners 470 Air Cleaners Not All Equally Efficient 470 Tests Made at Davis 470 The No. 1 Standard Dust 471 How the Dust Was Fed to the Cleaners 471 Finding the Efficiency of the Cleaner 471 Three Standard Efficiency Tests 471 Special Tests 472 Low Water Level Test 472 Fill-up Test 472 Vacuum Tests 472 Power Tests 472 The Cleaners Tested 474 Source 474 Classification 474 Water Type 474 Dry Type 474 Oil Type 474 A Homemade Cleaner 476 Identification Table 476 Results of the Tests 476 Tables: Identification of Cleansers Tested. Table 1 477 General Summary. Table 2 478 FiU-up Tests. Table 3 479 Chaff Tests. Table 4 480 Notes Taken from the Test Records 480 Graphs 482 Dust-Separation Efficiency. Fig. 6 482 Vacuum or Choking Effect. Fig. 7 483 Effect on Power of Engine. Fig. 8 485 Vacuum Effect Not Entirely Harmless 482 How to Get Best Results from Air Cleaners 484 No Leaky Connections 484 Breather Pipe Should Have Clean Air 484 Keep Water Level High 484 Place Air Intake High 484 Give Needed Attention 485 A Simple Efficiency Test for Air Cleaners 486 Acknowledgments 486 1 No attempt is made in this paper to illustrate or to explain apparatus and methods for testing air cleaners. For complete description see Hoffman, A. H., ' ' Efficiency of Dust Separation in Air Cleaners for Internal Combustion Engines, ' ' Transactions A. S. A. E., vol. 16 (1922). Reprints in limited numbers may be had by addressing Agricultural Engineering Division, University of California, Davis, Calif. Full data of all tests are on file and accessible at the same place. 470 UNIVERSITY OF CALIFORNIA — EXPERIMENT STATION Dust causes rapid wear. — When it is remembered that fifty-five per cent of the average dust floating around in California fields (fig. 1) is shown by test to be sand, it will not be hard to understand why a tractor engine needs an efficient air cleaner. An engineer of a prominent California tractor manufacturing company recently stated that their latest model tractor would run in dusty conditions under full load for three thousand hours with no adjustments what- ever, and for six thousand hours with only minor adjustments, pro- vided it had a satisfactory air cleaner; but that without an air cleaner the engine would be ruined in fifteen hours' use. Fig. 1. — Typical conditions under which the tractor in California must do its work. Dust entering by way of carburetor and breather pipe mingles with the lubricating oil and makes a grinding mixture which quickly wears out many parts of the engine. EFFICIENCY OF AIR CLEANERS Air cleaners not all equally efficient. — It has long been known that the numerous makes and types of air cleaners on the market differ in their ability to remove dust from the entering air, in the degree to which their use imposes vacuum or choking on the carburetor intake, and in their effect on the maximum possible power to be obtained from the engine ; yet there has been no publication of authoritative tests previous to those here given by which the user might know what service he might reasonably expect from any given air cleaner. Tests made at Davis. — To determine their dust-separation efficiency, vacuum imposition and effect on power, a long series of tests was made at the Branch of the College of Agriculture, Davis, Calif. Twenty- six cleaners (described later) were tested. A twenty-five h.p. four- cylinder engine was used on which to make the tests in order to have Bulletin 362] DUST AND THE TRACTOR ENGINE 471 conditions as nearly as possible like those encountered in actual work. Since in a field test it would be very difficult to avoid errors due to differences in kind of soil, dryness of soil, direction and strength of the wind, temperature and moistness of the air, etc., it was deemed better to make the tests in the laboratory where both load and speed could be made the same for all cleaners tested, and where a standard dust might be fed to all cleaners in the same manner, in the same amounts, and at the same rates. The No. 1 standard dust used in the efficiency tests was made up from fifty-pound soil samples taken from ten cultivated fields or orchards scattered throughout the principal tractor-using areas of this state. The ten soils were dried and pulverized and the finest parts taken out by an air-floating process. The dusts so obtained were mixed to make the No. 1 standard dust ; hence it is a composite which may be regarded as fairly representing the average dust in our fields at a height of about four feet above ground. How the dust was fed to the cleaners. — The standard quantity of dust fed in was fifty grams (1.76 ounces), and the usual standard time was about thirty minutes. The dust was fed to the cleaner under test by placing it in a tin tube from which a plunger pushed it out slowly against a revolving bristle brush wheel by which it was swept in a very fine cloud directly into the air stream entering the cleaner. Finding the efficiency of the cleaner. — The air leaving the cleaner passed next through an 'absolute cleaner' which caught the dust not caught by the cleaner under test, and then into the tractor engine. The cloths of the absolute cleaner were thoroughly dried and accurately weighed before and after the test, the difference in weight being the weight of dry dust not caught by the cleaner under test. Where this weight was five grams the efficiency would be 90 per cent. Three standard efficiency tests. — Three standard tests were run on each cleaner as follows : 1. A test at normal speed (1200 r.p.m.), normal load (20.4 h.p.), and quick dust feed (50 grams in 30 minutes). 2. A test at normal speed, light load (12.6 h.p.), and quick dust feed. 3. A service test, at quick dust feed, in which working conditions were approximated. In this run, variable load, variable speed, idling, backfiring and quick pick-up with wide open throttle, were features. At intervals load and speed were brought to normal and readings of vacuum were taken. (See table 2, p. 478, and fig. 6.) 472 UNIVERSITY OF CALIFORNIA — EXPERIMENT STATION Special tests were run when circumstances required. Among these were a low water level test, a fill-up test, and a chaff test. On water type cleaners (also on one oil type cleaner) where the level of the liquid would be lowered by use, an efficiency test was run with the liquid at the low level mark (if any). On cleaners which from their construction might presumably have their efficiency or vacuum affected by the entering dust a so-called fill-up test was run. This was not a standard efficiency test and for it the material fed in was a prepared dust known as No. 2 standard. It was uniform, but was much coarser and denser than the No. 1 standard. The purpose of the fill-up tests was to obtain a rough estimate of what might be expected of the cleaner when neglected and subjected to gross abuse. The amount of dust fed in fill-up tests was as much as seven pounds in some cases. The rate of feeding was also very rapid and for this reason little importance is to be given to those instances in the fill-up test where the cleaner did not show up well. However, where a cleaner under this very severe test was able to maintain good efficiency and low vacuum, its ability to stand abuse would in a measure seem to be established. (See table 3, p. 479.) The chaff test was designed for cleaners not provided with an adequate screen and of such interior construction as presumably to have the vacuum affected by the entrance of insects, leaves, and other organic refuse. Ten grams of dry wheat chaff passed through one- fourth inch mesh screen was allowed to be drawn into the air stream entering the cleaner, and the effect on the vacuum due to the cleaner was noted. (See table 4, p. 480.) Vacuum tests. — The vacuum or choking effect was measured by use of a U-tube manometer at intervals during the efficiency test runs and also during tests made to find the effect of the cleaners on the power of the engine. (See table 2 and figs. 7 and 8.) Power tests. — The tests to find effect on power were run as follows. After warming up thoroughly, the engine was run, cleaner off, with wide open throttle and carburetor adjusted for highest power. Next came a similar run with cleaner on, the carburetor being readjusted; then immediately another run with cleaner off. The highest power at standard speed obtained in the run with cleaner on was divided by the average of the highest powers obtained at standard speed in the first and third runs. The per cent so obtained was used as the means of comparing the power effects of the several cleaners. (See table 2, p. 478, and fig. 8.) Bulletin 362] DUST AND THE TRACTOR ENGINE 473 Fig. 2. — Water type cleaners. 1. J. I, Case Threshing Machine Company. 2. Cleveland Tractor Company. (Cletrac W.) 8. W. H. L. Donaldson. 10. Ford Motor Company (Fordson). 11. International Harvester Company. (10-20 Titan.) 14. Eoss-Wortham Company. (E. W.) 15. Stewart-Warner Speedometer Corporation. 16. Tractor Appliance Company. (Taco Siphon.) 24. Samson Tractor Company. (Model M.) 474 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION THE CLEANERS TESTED Source. — The manufacturer of each of the thirty 2 cleaners tested was invited to determine the size of cleaner to fit the engine on which the tests were to be run and to choose the particular cleaner to be tested. With a single exception manufacturers availed themselves of this privilege. Classification. — For purposes of comparison the cleaners were classified into three groups according to their mode of operation, as water type, dry type, and oil type. Fig. 3. — Dry type cleaners. • 6. Donaldson Company, Inc. (45 h.p. tractor.) 7. Donaldson Company, Inc. 13. Liljegren and Dugain, "Success" (for Fordson). 17. Copied from eiderdown cleaner shown in Chilton Tractor Jour- nal, June 1, 1919, p. 41. 18. United Manufacturing and Distributing Company. (Collector type.) 19. United Manufacturing and Distributing Company. (Ejector type.) 23. Bennett Carburetor Company. 25. Stromberg Motor Devices Company. Of the eleven water type cleaners tested, seven (nos. 1, 2, 10, 11, 14, 24, and 24a) 2 passed the dusty air under the edges of a float, two (8 and 27 ) 2 used the centrifugal principle, and two (15 and 16) used the atomizer principle. 2 Twenty -six cleaners were received in time for all the tests, four (nos. 24 a, 25 a, 25b, and 27) in time for power tests only. See tables nos. 1 and 2. Bulletin 362] DUST AND THE TRACTOR ENGINE 475 Of the ten dry type cleaners, two (13 and 17) used cloth niters, the other eight employed the centrifugal principle, the whirl being obtained by means of vanes, tangentially placed inlets, or both (6, 7, 23, 25, 25a 3 and 25b 3 ), or by a rotating member (18 and 19). Some of these (25, 25a, and 25b) had an ejector operated by exhaust gas from the engine to throw the separated dust to the outside air, while another (19) ejected the dust by the centrifugal action of the rotating member. These last cleaners therefore did not require to be emptied. Fig. 4. — Oil Type Cleaners. 3. Dailey Bros. 4. Dailey Bros. 5. Donaldson Co., Inc. 9. Fageol Motors Co. 12. A. E. Palmer (Experimental; . 20. Vortox Mfg. Co. (Pomona). 21. Vortox Mfg. Co. (Pomona). 22. Vortox Mfg. Co. (Pomona). 26. Dailey Bros. (Model 1). Of the nine oil type cleaners, one (12) employed the atomizer principle; the other eight had oiled niters, one (9) of cloth, three (3, 4, 26) of fibrous material, tAvo (21 and 22) of fibrous material combined with oil centrifugal, one (20) of wire combined with oil centrifugal, and one (5) of fibrous material combined with dry centrifugal. s For 24a, 25a, 25b, and 27 see tables nos. 1 and 2. 476 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION A homemade cleaner. — One cleaner, no. 17 (called no. 17' when outer shell is left off, as in fig. 5) was made up by a local tinsmith. It is pictured in the Chilton Tractor Journal, June 1, 1919, as an example of an obsolete type. Made up as shown and with two thick- nesses of eiderdown blanketing wrapped on, nap side out, over the cylinder of one-half inch mesh hardware cloth, its performance under all tests was very satisfactory. A cover is not essential to its oper- ation. It may be placed in any position. Fig. 5. — A homemade cleaner. Eiderdown cleaner (no. 17 and 17') cover removed, cloth laid back. A (9" X 18") cylinder of wire screen soldered to heavy tin pans. Tube (size of carburetor inlet) extending to inside. Iron bars suitable for attachment to tractor riveted to ends. Two thicknesses of eiderdown blanketing, nap side out, tied on. Identification table. — The cleaners tested are listed in table 1, the essential dimensions by which they may be identified and compared being given. A list of air cleaner manufacturers' addresses may be obtained from the Division of Agricultural Extension, University of California, Berkeley, or from the Division of Agricultural Engineer- ing, University Farm, Davis, California. The results of the tests. — Table 2 gives a general summary of the tests. Table 3 summarizes the fill-up tests, and table 4 the chaff tests. Bulletin 362] DUST AND THE TRACTOR ENGINE 477 TABLE 1 Identification of Cleaners Tested Make or name Type Weights Outlet Diam. inside, inches Inlets Size of body No. Clean dry lbs. Ready for use, lbs. No. Diam. inches or size of opening proper, Height, Length, Width, inches 1 Case Wet 18-3/4 39-1/4 2-1/2 1 2-1/2 19-1/2 x 12D 2 Cletrac W Wet 57-1/4 74-1/2 1-1/2 1 1-1/2 14 x 13 x 9 3 Dailey Oil 11-3/4 12 1-3/4 1 1-7/8 18 x 7-1/2D 4 Dailey Oil 12 12-1/4 1-7/8 1 1-7/8 18 x 10D 5 Donaldson Oil& Dry 5-3/4 5-3/4 1-1/2 15 1 -1/4 x 3/8 16 x8D 6 Donaldson Dry 5-1/4 5-1/4 2-1/4 1 2-1/4 12 x 11D 7 Donaldson Dry 2 2 2 15 1 x 1/4 11-1/4 x 8D 8 Donaldson Wet 15 10 1-7/8 1 1-3/4 17x9D 9 Fageol Oil 9-1/4 9-1/2 1-1/2 1 1-1/2 13-1/2 x 10-3/4D 10 Fordson Wet 40-1/4 53-1/4 1-3/4 1 1-1/2 x 1-1/2 12 x 14-1/2 x 6 11 I.H.C. Wet 18 34 2 1 2-1/2 18-1/2 x 10-1/2D 12 Palmer Oil 9-1/2 15 1-11/6 1 1 x 3 18 x 8-1/2D 13 Success Dry 10-1/2 10-1/2 1-7/8 1 1-7/8 15 x8D 14 R.W. Wet 8-3/4 18-1/2 1-3/8 6 1 15 x9D 15 Stewart- Warner Wet 7 16-1/4 1-3/8 2 1-3/4 x 2 13 x 9D x 6-1/2 16 Taco Siphon Wet 44-1/2 56 2 1 2-1/2 12-1/2 x 16 x 7-1/2 17 Eiderdown (closed) Dry 9 9 1-7/16 1 2 26 x 12-1/2D 17' Eiderdown (open) Dry 5 5 1-7/16 1 565 sq. in. 19 x 10D 18 United collector Dry 8-3/4 8-3/4 1-1/2 1 1-1/2 10-1/2 x 5D 19 United ejector Dry 2-1/2 2-1/2 1-7/16 1 3-1/4 7x5D 20 Pomona Oil 4 5-3/4 2 1 2 13-1/2 x 4-5/8D 21 Pomona Oil 3-1/4 5 2 1 2 13 x 4-5/8D 22 Pomona Oil 9 16-1/2 2 1 2-1/2 15 x 10D 23 Bennett Dry 4-1/8 4-1/8 1-3/4 3 1-3/16 9 x6D 24 Samson Wet 27-1/4 42-1/2 1-3/8 1 1-3/8 14 x 12-1/2 x 7 24a Samson Wet 28-1/4 37-1/4 1-1/4 1 2 x 6-1/2 15 x 10D 25 Stromberg Dry 3-1/2 3-1/2 1-5/8 12 7/8 x 3/8 14x6D 25a Stromberg Dry 3-1/2 3-1/2 1-5/8 12 7/8 x 3/8 14 x 6D 25b Stromberg Dry 4-3/4 4-3/4 1-3/4 12 1-1/8 x 3/8 16 x 7-1/2D 26 Dailey Oil 10 10-1/2 1-9/16 1 1-1/2 17-1/2 x 10D 27 Bennett Wet 8-1/4 15-1/4 1-3/4 1 1-3/4 12x9D 478 UNIVERSITY OF CALIFORNIA— EXPERIMENT STATION TABLE 2 Summary of Tests Make or trade name Dust separating efficiency per cent Vacuum Maxi- mum power with cleaner on No. At 20.4 h.p. 1200 R.P.M. inches water At max- imum h.p. and 1200 R.P.M. inches water Cleaner cleaned out after effi- ciency tests Ratio h.p. with cleaner on to Average all tests Service run Low Water or oil Cleaner clean 150 gram No. 1 Std. dust in h.p. with cleaner off. per cent 1 Case 96.0 96.1 95.2 1-3/4 1-11/16 2.6 27.0 100.0 2 Cletrac W 89.3 85.8 83.6 3-13/16 5-1/8 8.5 26.6 98.7 3 Dailey 99.5 99.9 6-5/16 22-7/16 Run lost due to error 4 Dailey 99.8 99.9 5-7/16 18 10.7 26.6 98.7 5 Donaldson 97.8 98.0 3-1/8 20-9/16 4.5 27.2 100.6 6 Donaldson 74.7 67.4 1-5/8 2-7/16 3.8- 27.1 100.3 7 Donaldson 51.0 50.3 2-1/16 2-5/16 4.2 27.1 100.2 8 Donaldson 96.5 97.3 95.0 2-5/8 4-3/4 7.4 27.0 100.0 9 Fageol 98.5 99.5 4-1/2 16 12.8 25.7 95.2 10 Fordson 88.9 89.8 85.5 7/8 1-1/2 3.3 26.8 99.3 11 I.H.C. 94.6 94.4 92.3 2-11/16 2-7/8 3.5 27.4 101.7 12 Palmer 96.6 96.9 94.7 oil 4 6 6.3 26.9 99.6 13 Success 98.7 97.8 1-1/2 27-1/4 + 2.2 27.0 99.9 14 R.W. 93.1 92.6 92.6 3-1/4 3-5/8 8.3 26.9 99.7 15 Stewart- Warner 95.7 95.8 Const, level 4-5/8 5-9/16 11.1 26.5 98.3 16 Taco Siphon 92.4 94.0 85.9 2-1/2 3-3/16 5.8 27.0 99.8 17 Eiderdown (closed) 99.4 99.5 2-5/16 2-5/8 5.6 27.2 100.7 17' Eiderdown (open) Test not run 4.8 27.0 99.9 18 United (collector) 73.3 72.5 9 9 12.7 26.5 98.2 19 United (ejector) 62.2 58.3 3-3/16 4-1/2 9.3 27.0 99.8 20 Pomona 98.4 98.2 2-3/16 2-5/8 7.2 26.7 98.8 21 Pomona 98.8 98.6 2-5/16 2-3/4 5.3 27.5 102.0 22 Pomona 97.4 97.1 1-5/8 1-11/16 4.8 26.6 98.3 23 Bennett 42.7 37.6 5-1/2 6-1/16 8.9 26.6 98.2 24 Samson 95.9 95.2 Not run 5-3/8 5-3/8 10.1 27.0 100.1 24a Samson 6.9 26.8 99.1 25 Stromberg 88.5 87.4 4-3/8 6-7/8 11.1 26.5 98.3 25a Stromberg 8.8 26.4 97.6 25b Stromberg 7.0 27.2 100.6 26 Dailey 99.7 99.6 8-1/8 9 17.4* 26.7 98.9 27 Bennett 9.3 26.8 99.1 ♦Abnormal; chaff of chaff test not all out. Bulletin 362] DUST AND THE TRACTOR ENGINE 479 TABLE 3 Summary of ' ' Fill-up Tests ' ' Dust Water Clean- Test Lgth. B.H.P. Room Dust caught or oil er No. run run Temp. No. 2 by going No. Min. °F Std. lbs. abso- lute cleaner lbs. over lbs. 1 114 48 20.4 104 9 1/16 2 102 54 20.4 91 5 1/8 .9* 5 71 31 20.4 83 50 gram 1 gram 9 22 108 20.4 16 78 99.7 0.16 5 t .05 10 28 13 16 68 8 36 16 20.4 82 67 0.16 3 t .25 15 16 50 38 20.4 85 5 .02 17 138 76 23 18 20.4 20.4 92 95 5 2.5 t X 20 21 33 30 20.4 76 3 1 22 81 36 20.4 100 7 1/8 24 38 56 20.4 86 5 1/8 Trace 26 57 62 20.4 76 5 t Vacuum due to cleaner, inches water Min. 1-7/8 5-1/8 7-3/16 6-3/8 1-1/2 11-1/4 5-7/16 2-3/4 3 2-3/4 2-3/8 1-3/8 4-7/16 8-5/8 Max. 11 7 20-9/16 26-3/4 2-5/16 24-5/8 16-1/8 5-1/16 4-17/32 3-1/8 14-5/16 3-7/16 26-3/8 10 Rate of air flow 53 53 55 53 47 51 55 53 54 58 55 54 Note 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Note 1. 2. 3. One pound sand, etc., found in horizontal inlet tube in base of cleaner at close of test. Dust of previous runs (200 grams) left in. Cleaned out before test. Dust of previous runs (100 grams) left in. dust. Not deemed necessary to use any No. 2 Standard 4. Dust of previous runs (100 grams) left in. 5. Dust of previous runs (200 grams) left in. Tested at Fordson load 6. Dust of previous runs (100 grams) left in. 7. Dust of previous runs (150 grams) left in. 8. Dust of previous runs (250 grams) left in. 9. Dust of previous runs (150 grams) left in. 10. Dust of previous runs (150 grams) left in pound dust in. 11. Dust of previous runs (200 grams) left in to spray. 12. Dust of previous runs (150 grams) left in. 13. Cleaned out before this test. 14. Dust of previous runs (150 grams) left in. * Amount abnormal since rapid rate of dust feed raised water level about 5/8 inch above high mark t Not weighable on 1/4 oz. sensitive scale. % Not weighed. § Cu. ft./min. 60°F. and 14.7 lbs./sq. in. Dust going through rather freely after 4th half Sand began passing to trap soon after oil ceased 480 UNIVERSITY OF CALIFORNIA — EXPERIMENT STATION TABLE 4 Summary of Chaff Tests (Ten grams dry wheat chaff was fed to cleaners not having adequate screen, to find effect on vacuum.) Cleaner No Run No Length run, Min B.H.P Rate of air flow cu. ft. per min. 60°F and 14.7 lbs./sq. in Vacuum due to Min. cleaner Max. Remarks 1 5 8 10 17 20 24 113 72 129 107 137 77 37 10 11 30 8 8 47 33 20.4 20.4 20.4 16 20.4 20.4 20.4 53 54 53 47 53 54 55 1-3/4 3-5/16 4-3/8 1-1/8 2-3/4 2-11/16 5-3/8 1-3/4 3-3/4 10-3/16 1-5/32 2-7/8 6-3/4 32-15/16 Note 1 Note 2 Note 3 Note 4 Note 5 Note 6 Note 7 26 58 43 20.4 54 9-5/8 28-3/4 Note 8 Note 1. Dust (150 grams) of previous runs left in. 2. Filter cleaned out and reoiled before test. 3. Drained out (but not flushed) and refilled before test. 4. Dust (200 grams) of previous runs left in. Water level low. 5. Dust (150 grams) of previous runs left in. 6. Filter cleaned out and reoiled before this test. 7. Dust (150 grams) of previous runs left in. No. 24A (received too late for test) has an adequate screen. 8. Dust (150 grams No. 1 St'd. and 5 lbs. No. 2 St'd.) from previous runs left in. NOTES TAKEN FROM THE TEST RECORDS* Cleaners nos. 3, 4, and 26, Dailey. — The intake cone furnished regularly with this cleaner was not in position during any of the tests. Its effect would be to increase efficiency and vacuum slightly and to lengthen a little the time between clean-outs. Cleaner no. 4. — The vacuum measured by 18 inches water (read after 150 grams no. 1 standard dust had been fed in) dropped to 10% inches when the cleaner was jarred with wooden mallet. Cleaner no. 6, Donaldson Dry. — This cleaner is designed for a larger engine than the 25 h.p. machine used in the tests. After the runs on this cleaner, a large quantity of dust was found in shroud of cleaner. The efficiencies found were therefore higher than they would be after runs enough to fill free space in shroud. Cleaner no. 8, Donaldson Water Type. — This cleaner should have a more obvious place for filling with water. Cleaner no. 12, Palmer. — A trace of oil in very finely divided form was found on top cloth of absolute cleaner after each test. When motor idles or stops, oil settles into air intake tube in such quantity as to cause 'choking' of carburetor. This vacuum (fluctuating about l!/4 inches of water) might cause trouble in starting a warm engine. 4 See also notes in tables of summaries. Bulletin 362] DUST AND THE TRACTOR ENGINE 481 Cleaner no. 15, Stewart-Warner. — At the end of dust feeding in fill-up test, vacuum due to cleaner was measured by 161/g inches water. After seven minutes longer run without more dust, vacuum decreased to 14 1 % 6 inches. One side of cleaner was almost completely clogged because of the rapid rate of feeding. Some iron rust was found in this cleaner, and it may have contributed, to the clogging. The later cleaners of this make are made of brass. Cleaner no. 19, United Manufacturing and Distributing Company Ejector. — A test of length of life of the bearing was run on this cleaner, twenty-five cubic feet per minute air being drawn through by vacuum cleaner. Time 216 hours before the efficiency tests, 315 hours after. Total 531 hours. Cleaner chattered a good deal toward close of test. Cleaner no. 23, Bennett Dry. — Run (no. 64) discarded as official service run, but averaged with runs 62 and 63 since dust from previous runs (nos. 62 and 63) probably dropped from interior walls of cleaner and passed on to absolute cleaner. If cleaner had been shaken during and at close of runs 62 and 63 this dust would have reduced their efficiencies. Dust (dry) removed from cleaner after runs 62, 63, and 64, totaled 61.90 grams. Cleaner no. 25, Stromberg Ejector. — With normal load and speed and with cleaner outlet and inlet disconnected from other apparatus and closed, expirator produced a vacuum inside of cleaner body measured by 12% inches water. Pressure of exhaust gas in pipe one inch from cleaner was meas- ured by 11% inches water, average, during first tests on this cleaner. During the last tests it was about 4% inches. The difference is due probably to a slight change in the point of exhaust manifold where gas for the cleaner was piped out. It is interesting to note that the cleaning efficiency is unchanged. After the seven runs, a carbon dej)Osit about %2 inch thick was found in the throat of the expirator. Graphs. — Figure 6 shows graphically the dust separation efficiency of the cleaners tested, and makes possible comparison individually and by groups. Fig. 7 makes possible a similar comparison with respect to the choking effect and how this changes as the cleaner takes in more and more dust. It will be observed that for the large majority of cleaners the vacuum is less than ten inches of water and either constant or nearly so, while for a few it runs up very rapidly. Fig 8 shows the effect of each of the several cleaners (when clean) on the power of the engine and the effect on the vacuum when the engine is giving maximum power. The curved line (used in each of 482 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION the four parts of fig. 8) was obtained by measuring the maximum power obtainable from the engine, first with the carburetor unob- structed, then with its intake closed more and more by a gate valve. The object of this test was to learn whether the effects on power of the several cleaners were due only to the vacuum they produced or to some other cause. It should be pointed out that, while the utmost care was used to avoid errors, the results are not conclusive for the reason that the effects of the cleaners (when clean) on power are so small that it is possible that slight changes in the engine itself may in some cases have had more effect than the presence or absence of the cleaner. Dust SePAeAnno frr/cic/icr. Avc^agc or all resrs. pe/?cf/rr 20 40 60 SO 100 A v3 7 HI 17 — iSXn i=i 924 97S 9SB sower 6Z4 Lot* fYATCR <5Q3 978 99S 9*4 seo H5£ .84 6- S>?6 ^SSZ OX.94, Fig. 6. — Graph of dust-separation efficiency. Vacuum effect not entirely harmless. — While it has been shown that none of the cleaners if clean has very much effect on the maxi- mum possible power of the engine, it must not be assumed that the vacuum effect is entirely harmless, especially if it increases as dust is absorbed by the cleaner. Tf at full load of the engine the vacuum be above about 10 inches of water and the needle valve of the car- buretor be set for correct mixture, it will be found (in the case of some carburetors) that at half -load the lower vacuum and hence less suction on the gasoline jet will make the mixture much too lean, possibly causing back-firing and even stoppage of. the engine. If it be adjusted correctly for light load, the mixture will be altogether too rich at full load. Thus the high vacuum will cause one or the other of two undesirable conditions: either the mixture will often Bulletin 362] DUST AND THE TRACTOR ENGINE 483 be much too rich (with consequent waste of fuel and rapid carbon- ization of spark plugs and inner walls of the combustion chambers), or the operator will be under the necessity of adjusting the needle valve too frequently. This will be the case for variable load, e.g., a tractor on drawbar work. For belt work with steady load, e.g., run- ning an irrigation pump, a vacuum measured by 10 inches or even 15 inches of water is not serious. CD -♦- O en (U n o c c D 0) u 3 O 5 20 X) o o o ,5 E u O Z. \Q 7^"' "^^Avff n "d d"r"y" / _^ ^-—\ iX W E T. I6^r-W£T A 7 PR V_ >HII KWW rri ' -WET i -WET 7^ DRV IOO I DRV /N2I- OIL oil ££ OIL IpO GPAM5 OUNCES 1 — I l 1 1 — O 1 2 3 4 5 Amount No. 1 Standard Dust Fed In. Note: Sudden increase of vacuum was due to cloth cone inside of cleaner being sucked against outlet opening. Fig. 7. — Graph of vacuum or choking effect. A line rising as it passes from left to right indicates a cleaner the vacuum of which increases as dust is absorbed. 484 UNIVERSITY OF CALIFORNIA — EXPERIMENT STATION HOW TO GET BEST RESULTS FROM AIR CLEANERS No leaky connections. — No leaky connection should be permitted between air cleaner and carburetor or between parts of the air cleaner itself. A loosely fitting slip joint or an ordinary flexible metal tube are almost sure to admit some air and a great deal of dust. A piece of radiator hose fitting tightly over the tubes to be connected is very satisfactory if the connection is short. If long and the vacuum rather high, the hose walls may be caved in by the air pressure and thus the passage be obstructed. In such a case the line may be spliced, a thin metal tube making up the larger portion of the length. Ordinary friction tape is a very satisfactory emergency material for stopping air leaks. Filters, whether of cloth or fiber, cannot be efficient if they have holes in them. The dusty air will go in largest quantity by the path of least resistance. The dust will find the hole if there be one. Breather pipe should have clean air. — A special cleaner 5 may be used or connection may be made to the carburetor air cleaner. If clean air is not provided, the breather pipe itself acts as an oil-type air cleaner taking in breaths of dusty air, collecting the dust and sand on the oily walls, then blowing out puffs of clean air. When dust goes in through the carburetor there is a chance that some of it may escape through the exhaust. Not so with dust on the walls of the breather pipe. It remains until the oil washes it down into the crank case. When it is not feasible to connect the breather to the air cleaner, a small bag of eiderdown blanketing or cotton flannel, lint side out, may be tied over the breather pipe. This will remove one leading cause of rapid wear of timing gears, front piston and cylinder and front main bearing in some tractors. Place air intake high. — Except in orchard work it is usually possible to use a 'periscope' or high vertical extension for the air intake. The advantages are that the quantity of dust to be removed by the air cleaner is greatly reduced and that the coarser dust is avoided. Some air cleaners are regularly furnished with a jointed periscope. Part or all of it should be used whenever feasible. If a periscope is provided, it should be smooth inside and free from sharp turns. The inside diameter should be amply large (not less and preferably more than 2 inches for a 10-20 tractor), or the vacuum effect will be unnecessarily increased. s The market affords breather air cleaners for certain tractors. Manufacturers ' addresses will be furnished on request. Bulletin 362] DUST AND THE TRACTOR ENGINE 485 Give needed attention. — There is not on the market, so far as the writer is aware, any air cleaner that can be put on a tractor or automobile engine and forgotten aftd yet month after month give adequate protection against dust. All require some attention and have their own peculiar troubles. Those with moving parts have troubles due to wear and to accumulation of dust and oil. Thus the float in some water type cleaners may have holes worn which would permit dusty air to pass without going through the water. Centrifugal types WET EFFECT OF AIR .EANERS ON POWER OF MOTOR I 10 20 10 20 10 20 10 20 VACUUM AT CARBURETOR INTAKE. INCI-IE5 WATER BELOW ATMOSPHERE 30' 40 50 Fig. 8. — Graph of effect of air cleaners on power of engine. The curved lines show how power falls off as carburetor intake is choked (no air cleaner on). Distances up show brake horsepower, distances to right show vacuum or choking effect measured in inches of water. Maximum power with cleaner off was 27 h.p. Each cleaner's effect on power and vacuum is shown by position of the numbered points. may become so encrusted with oil and dust that their action may almost entirely cease. Cleaners having small passages inside may clog up solid. Other kinds may so increase their vacuum effect due to accumulation of dust that the power of the engine may be greatly reduced. Nearly every air cleaner has its plate giving the manu- facturer's directions for the care required. Some tractor operators may be able to improve upon these directions, but none may safely neglect them. 486 UNIVERSITY OF CALIFORNIA EXPERIMENT STATION A SIMPLE EFFICIENCY TEST FOR AIR CLEANERS Not what the air cleaner catches, but what it lets go by does the mischief in a tractor engine. Whether any appreciable amount of dust gets past an air cleaner may usually be known by disconnecting the air cleaner from the carburetor and wiping out the inside of the connecting tube. If after a ten-hour run under dusty conditions only a trace of dust can be wiped out, the cleaner has probably done a first-rate job. This test may fail, if, as might possibly happen, so much water or oil should go over from the cleaner that the tube would be kept washed out. ACKNOWLEDGMENTS The author wishes to acknowledge his indebtedness to manufac- turers who loaned apparatus and made helpful suggestions as to test methods, to members of other divisions of the University of California for equipment placed at our service, to Professors Fletcher, Thomson, Moses and others, who furnished counsel and active assistance in the testing work.