THE MECHANICS AND DISTRUMENTATION OF A TEST PROGRAM FOR GOPMERCIALQISHWAmgfi by Paul Jay waning A Thesis Submitted to the Graduate School of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MECHANICAL ENGINEER Department of mechanical Engineering 1948 THEIiS TABLE OF comgrrs Page 1. Introduction.................... 1 2. Laboratory Arrangement.......... 2 3. Hot Water Supply System......... 7 4. Temperature Measurements. . . . . . . . 10 5. Discussion of Soiling Machine... 17 6. Discussion of Surface Analyzer.. 3} 7. Rinse'Line Flog 8tudy........... 44 8. Wash System Investigation....... 52 9. Practical use of Instruments.... Developed....................... 63 10. Conclusions..................... 65 11. References...OOOOOOO....;...O... 66 "1‘ ",,‘ -5 1.111,- 2.5. ACKNOWLEDGMENT The author wishes to express his appreciation to Mr. Pat Davis Jr., of the Champion Dishwashing Machine Company, and to Mr. G3 B. Fox of the Hobart Manufacturing Company. Each of these men was very helpful in supplying illustrative material used in this report. ' Thanks also must be given to Mr. Donald Clark and mr. Roger Cessna who helped in securing some of the data for the tests on water flow. Acknowledgment must be given also to the Elec- trical Engineering School of Michigan State College for the use of the Weston light meter in the develop- ment of the Surface Analyzer. ~ m. cc. lee. o. o. 0 .‘.~ no. N“ "fluflufl .00... Io.\.e 0.0... TEST LABORATORY Located at Cell age? East Lansing Michigan State INTRODUCTION The development of commercial type dishwashing equipment'has reached the stage where it seems that manu- facturers must secure more pertinent information about the effectiveness or their basic designs. .It is also true that Public Health officials are demanding basic information about the process of dishwashing so that pub- lic health codes can be formulated and enforced more intel- ligently. The testing of dishwashing equipment was done there- fore not only to help manufacturers improve their product, but also to learn as much as possible about the fundamental problems of machine dishwashing. This type of test work brought into harmonious cooperation, workers from the mechanical and bacteriological fields. While it may occasionally be necessary to refer to ‘ bacteriological phases of the work, it will be primarily the purpose of this paper to present the various mechanical features involved in the testing of commercial dishwashing machinery. This work was sponsored by the National Sanitation Foundation which provided necessary funds and brought to- gether the interested parties. The project was under the supervision of Dr. W.‘L. Mailman of the Bacteriological and Public Health School of Michigan State College. The author of this paper was charged with the direct responsi- bility for the installation and operation of all mechanical -2— apparatus, and the development of all instrumentation as well as a large number of the test procedures used. LABORATORX PROVISIONS Machines The arrangement of laboratory facilities for testing dishwashing machines was made to accomplish two important things. First, it was desired to test existing equipment under ordinary field conditions. Second, it was essential to establish laboratory controls over all the machines and procedures. The first purpose was realized by using typical dish- washing,machines furnished by five leading,manufacturers; The Anstice Company, the G. W. Blakeslee Company, The Champion Dishwashing machine Company, The Colt Patent Fire Arms Company, and.The Hobart Manufacturing Company. Each of these companies loaned two machines and were very gen- arena in their cooperation. The kind of machine used for the tests was the single tank manually operated dishwasher, illustrations of which are to be seen in Figs. 1 and la. Two systems are represented both of which were used. The one system‘has a set cflrewelving wash arms that rotate by the inertia effect of flowing water Just as does a lawn sprinkler, while the other system employs stationary wash Jets. The rotary wash arms had specially shaped rectan- gular nozzles, while the stationary wash arms were merely brass pipes with slots milled at intervals. The rinse «fleeces-o eon-elvothduv-Q e Ice... I. Ooo‘.... '0. .' E? Fig. 1. Single Tank Manually Operated Dishwashing Machine FT SMOOTH TOP __)L :i . ID WASH AND RINSE CONTROL HANDLE HAND NUT FOR EASY REMOVAL OF WASH ARM LOWER REVOLVING - WASH ARM ...... SELF-CLOSING RINSE VALVE -1“... . WEI“ BALL BEARING REVOLVING WASH WING 'TYPE RINSE SPRAYERS is LARGE NON-CLOGGING WASH NOZZLES FILLER PIPE / PROTECTED MOTOR SWITCH ESPECIALLY DESIGNED PUMP FRONT ENCLOSURE ..... , ~~--sELL-TOP / LOWER REVOLVING RINSE ARM lOCKS ATRE/fl' ANGLES 70 WASH ARM MM WASH OPE'RAW COUNTERBALANCE FOR DOORS REMOVABLE STRAINER PANS AUTOMATIC OVER FLOW GAS BURNER < -— FLUE THERMOSTATICALLY CONTROLLED -STEAM INJECTOR ’OR GAS BURNERS [Xi/M (WA/1’65] FOR 307/! ADJUSTABLE LEGS PANEL DRAIN CONTROL HANDLE )A A.( — 1'13. 130 Phantom View of Single Tank Machine -5- F18. laPIIANTOM VIEW OF SINGLE TANK CHAMPION . t No obstructions on top or back l ‘ Brass Door Pulleys and Brackets Bronze Door Chain Upper Rinse Nozzles Upper Wash SRpray Plglcs Counterweight (:ym'o‘zo‘df for Doors Counterbolonced Easy Operating Doors at Each End Water Inlet- " . . - ~. I I. , . . _ 3. . Above Overflow I i‘ A . - - _ .. - _ ’ ' ‘ _ Level . . ' - . ‘l ' ' ' Z ._ Lower Wash Spray Pipes Removable b Hand) Single Lever Y Control for Wash and Lower Rinse Sprays Rinse Nozzles Large Automatic Overflow Removable Refuse Screens if '- ‘ . Cast Iron I . Tank . Standard ‘ . ' Moire , l Motor Scientifically ‘~: Designed Pump u, Cast Iron Bose I Adiustable Legs Drain Valve .592? base—e mm Shea =e ho mum—338.com was EMS—@874 25. deuce—O Esme—0.8.: a; 3 9:5 mm 2.3er 3592.8 :3 science: a swsoEu mammmma 5&1 Iris macaw .3o28 whoafiafiv 2: .«o Oaxaca was Saoa of .3 52 a: meim pie :cssbnsfi 95:.» BE. .Bacfiteafioo 0.5 of E magmas—383 moor—co 53:85 3 henna E .3an me 333 ll- 0 V R . .t.. n. s I O H . . L. 3:. J. . . 23c? mammar— wad wnEmaB 05 Sandman 3 “ashamed 3 Laura mm Z .825 a»: 23. nausea: comma ems mafia—awe: 2: Om? OHOZ 42.8 no: on macaw? 95 23 .223 man5 no}? ESE—ca 8QO c5 3.02 .93an mamas? ceafiaa 9:..— .do. of S ”macaw 9.3ch “5955 SE can Emma 23 on. demeaning 0.3 ncEScnO x53 cinema e Mo £355 .9: if was mama—me? can EEO—9:8 Bo: macaw define—$2: ”WEE / . r . ZOHEMU MZANH ”HAM—DOG rmO gamut? EOHEEAH‘ -5- system in each machine is separate from the wash, and both rotary and fixed nozzles were employed in.the various machines tested. From the bacteriOlOgical standpoint it was desirable to keep the wash and rinse process separate. ‘Hence two washers of each make were set up side by side, but this would not be done in a restaurant and one such machine is sufficient for commercial work. Parenthetically, it may be well to give a brief description of the way in which a machine is used. The operator loads a square rack with dishes, places it in the machine, closes the doors, and opens the wash valve. ‘wash water at the proper temperature and with the cor- rect detergent concentration is then.flushed through the nozzles onto the dishes, by means of a centrifugal pump that continually recirculates this wash water over the dishes until the operator is satisfied that the dish- ware is clean. Next he shuts off the wash water and opens the rinse valve which provides a finer spray of fresh hot water to flush and sanitize the dishes. After unis, he removes the dishes from.the machine and allows them to dry themselves by the heat absorbed from the hot water rinse. Having set up the dishmachines in the laboratory to simulate field conditions as far as actual Operation was concerned, it became necessary also to provide accurate and flexible laboratory controls for carrying on the test work e Hat Water ggpply One of the first requirements was to provide an ade- quate supply of hot water. Little or no difficulty was encountered in maintaining closely regulated wash water in the tanks of each machine. It was essential, however, to obtain rinse water at various temperatures between 140° and 1900?. and to control it to within 1 l6 for any specified temperature desired. The amount of water re- quired for the various machines was from 8-18 gallons per minute. The maximum used at any given time was that needed to supply one machine, so the demand on the hot water sys- tem was placed around 25 gpm. To supply this, a 180 gallon tank was set up in the basement directly below the dishwashing machines. It was heated by an immersion steam coil capable of using steam pressure of 9O#/in2. The tank and coil was purchased from the Kewanee Boiler Com- pany. To aid in more uniform heating of this water, a circulating pump was installed to keep the tank water in motion. Accurate regulation of the water temperature was accomplished through a system of air operated mixing valves furnished'by the Johnson Service Company. Fig. 3 shows a schematic arrangement of the piping and controls for the system. The tank of water is heated by steam at about 90# pressure to insure rapid recovery. Steam is controlled by an air valve (A) so that water temperature can be held fairly close to a pre-set level. This hot water is then h..lU\<\\.\U~u\%/\ A: l I. .7 w is.» at «as use I - .III A 91$me _. - I - II c. Noe .__ III II. (tank VQU Q%\ llo l III. _ .QWQ RW\\ f"I . I 'l 'i i: EUR WNW; WWK «CS k0} .W\(\.V U15. . -9... mixed with cold line water to produce a tempered supply. This first mix valve (B) is controlled by a thermostat feeler bulb in the water line Just following the valve. This tempered water is then fed to a second mixing valve (C) which combines it with hot tank water to produce the exact temperature required. The feeler bulb for the operating,thermostat of the last mixing valve is set in the water line close to a header (D) from.which each dish machine drawn its supply. The three thermostat controls are set together on a board near the machines. Any de- sired temperature can be set on the control dials and hot water will be available at all machines. Successful Operation of this system depends on sev- eral factors. The proper behaviour of the mixing valves demands a flow of water past the thermostat feeler bulbs, so in order to maintain an accurate rinse line tempera- ture it is necessary to let the system Operate contin- uously during any given test run. This is done'by opening one fill pipe valve in some machine while using the rinse system of an adjacent washer. This method is used be- cause of the extreme case of intermittent flow used by the rinse systems. It was necessary to use the water for periods as short as five seconds, or as long as thirty seconds. Since so many variables are present in the entire system, some difficulty was experienced in attempting to set the thermostats during the period of initial adjust- ment, but once the controls were set, temperature regulation -10- of rinse water to within ileF. was obtained at any particular valve desired. The system was very responsive in that it quickly produced any temperature change re- quired, even as low as 1°F. A recording thermometer chart is shown in Fig. 4 from which the desired close regulatiOn is evident. Rapid cycling of the temperature is due to the action of the final mixing valve, but this was not objection- able since the amplitude of the wave was not greater than 20F. This variation was experienced at the fill pipe outlet of the machines, but was unnoticeable in the temperature of the rinse spray itself. Temperatggg Measuremgnts One of the important features of the entire test procedure was the determination of optimum temperatures for wash and rinse water. It was essential therefore that some careful attention be given to their accurate measurement. Instruments chosen for this Job were se- cured from the Foxboro Instrument Company; two six inch dials, and one recorder. Both the dials and the recorder were of the same type, liquid filled capillary tubes. . Feeler bulbs were at the end of six foot flexible capil- lary tubes so they could be inserted anywhere in the mach- ines. All three thermometers were mounted together on a small table which could be moved at will to any machine under test. This arrangement is shown in Fig. 5. The recording thermometer has three pens, one for recording I ' w nminiiiiiii""inn."-=' IIIIIIu '“ ll IIIIIIIIN‘III‘I , \\ g: Fig. 4. ' \ VNNNNNNNNNNNRA Segment Of Recording \\ WW Thermometer Chart Show- ing Close Temperature \\\\\\\\ Regulation of Hot Water, Fig. 5. Portable Table Mounting for Thermometers -12- time units and two for temperatures. This proved to be very convenient as it permitted both the supply line and the rinse spray itself to be checked at the same time, and recorded against the actual time interval used for a given test. The recorder was equipped with a two speed spring wound clock, which could be set for one revolution every twenty four hours or one revolution every twenty four minutes. The latter speed was used in all the tests since each chart division was then equal to one minute, thus making it possible to read time in seconds. Accuracy of the thermometers was checked several times against very sensitive and accurate laboratory mercury thermometers, and it was found that the instru- ments gave excellent results during the entire test period. Incidentally, it may be added that an accident occurred I during the early stages of the project which caused some concern, but which proved ultimately to be of no conse- quence. The author came into the laboratory one morning to find that the table upon which the thermometers were mounted had been carelessly bumped by one of the Janitor crew, and had fallen over on the floor with sufficient force to break the clamps and bracket used for fastening the instruments to the table. After the damage was repaired, the thermometers were checked for accuracy, and it was found that the dial thermometers were both reading about 13° low through their entire range. The recording instru- ment for some'reason was not damaged. Repair of the dial -13- thermometers was accomplished simply by shifting the pointer to the correct reading. Apparently the pointer pinions had each slipped over one tooth when the instruments fell. Tables No. 1, 2, and 3 show the results of calibration test on the thermometers before and.after repair. For- tunatsly there was no irreparable injury. Other periOdic calibrations indicated the thermometers were performing satisfactorily. Table No. 1 Showing Calibration of Dial Thermometers after Accident Mercury Dial No. 1 Dial No. 2 Thermometer 90 79 78 100 87 86 120 106 107 140 127 127 160 147 147 170 157 ' 157 180 167 167 190 177 178 Table No. 2 Showing Calibration of Dial Thermometers After Repair Mercury Dial No. 1 Dial No. 2 Thermometer Up Down Up Down 100 100 100 120 120 119.5 140 140 1A0 160 159.5 159.5 170 169.5 171 ' 169.5 171 180 179 180 179 180 190 187.5 :39 189.5 190 200 197 1 198 201 Table No. 3 Showing Calibration of Recording Thermometers After Accident Mercury Pen No. 1 Pen No. 2 Thermometer 100 ‘ 100 100 120 120 120 140 1#O 1‘0 160 , 160 160 170 170 170 190 180 180 190 19° 19° -14- The dial instruments were used primarily for explora- tion work within the machines while the recorder was used to make records of the test conditions. A convenient way of using the bulbs was to insert them in very small sheet metal trays which were then placed in the dish racks to catch the rinse or wash water as it cascaded over the dimh- ware. This method prevented too rapid change in the ther- mometer readings which would have occurred had the bulbs simply been laid.bare in the racks. Fig. 6 shows the record of a typical rinse test. Supply line temperature is shown in blue, rinse spray in red, and time in green at the edge. Another problem of temperature measurement and control was found in the method used for drying freshly soiled test plates. This work was done in a desicating oven wherein were installed an electric fan to circulate the air, a couple Of resistance heating elements and a number of glass trays filled with flake calcium chloride. During use the fan was kept running continuously to prevent spotty heating. Temperature was controlled by connecting the two elements directly to a Minneapolis Honeywell mercury switch thermo- stat controller. This arrangement besides providing auto- matic Operation, gave excellent results in standardizing the curing process. Uniformity was obtained with minimum effort and attention. A third problem in temperature measurement was encoun~ tered in the investigation of heat movement through a china dinner plate such as was used in the tests. This study was \ ‘Fig. 6. Recording Thermometer Chart Showing‘Use of Three Pens During 3 Rinse Test Red.1ine - Temperature of supply water Blue line - Temperature of rinse spray Green line - Time intervals in seconds _..._ -___.l, I "--—.—u—. _- . .. h... Fig.a7. Thermocouple on Plate -16- of interest because the rate and amount of heat acceptance by a dish has some bearing on the time rate of bacteria re- duction by hot water as it pours over a dish. Some diffi- culty was experienced in getting an exact picture of the rise in temperature of a china dinner plate, but the method used proved adequate. During the operation of a dish machine a plate is struck by water on every side so that heat pene- trates to the center from front to back. This complicates the determination of the exact behavior of heat flow. How- ever, since a plate is usually placed in the rack on edge at a slight angle from the vertical with face up, it was decided to observe the temperature conditions on the back surface by placing a thermocouple on the plate in the manner shown in Fig. 7. The couple was a covered with a round rub- ber pad about an inch in diameter and a small glass dome filled with glass wool was clamped over it. This insulated the couple from hot water, but allowed it to lay in direct contact with the plate. During the tests the whole unit, plate with attached couple, was placed in a rack, p33 as shown in Fig. 7, but normally as it would be during a washing operation, 1. e., at an angle slightly off the vertical, face up. Use was made of an iron-constantan couple directly connected to a very sensitive millivoltmeter made by the Rawson Electric Instrument Company. Before installation, it was calibrated by immersing it with a Foxboro bulb in a tank of water which could be heated at will. Millivolt meter readings were taken for a series of water temperatures. -17- Leads for the couple were rubber insulated and a thermos bottle filled with ice was used as a cold junction both A during calibration and subsequent tests. Results of the test are shown in.the curves of Fig. 8. DEVELOPMENT AND CALIBRATION 0F SOILING APPARATUS In order to measure the bactericidal activity of hot water it was necessary to develop a film or coating to hold the bacteria in place on the test plate during the rinsing period. Also in testing the washing cycle it was necessary to develOp a soil to be spread evenly over test plates. The soil in both instances cited had to be spread evenly and in constant volume. The reason is quite obvious. If unevenly coated plates were passed through a dish- washing machine and came out with soil spots remaining on their surfaces, it would be difficult to determine whether the machine was faulty or whether the varying soil thickness prevented thorough washing. Further, in checking rinse temperature, it is essential that the bacteria be spread evenly over the test plate surface and that the number on each plate be approximately the same so that quantitative bacterial tests can be made. In order to eliminate the soil distribution factor it was found necessary to develop a method for soiling dishes uniformly and consistently. As soon as consideration was given to the problem of coating a dinner plate with an even film of soil, it was evident at once that ordinary methods such as brushing, smearing, or dipping are entirely ~18- \WQZOUWW\ m:\.h or. 00. On. 0V. 0»... 0d. a: no. 00 0% on 00 om 0V. Om. O. h. OQ MLTSQ Qm2>tQ I35 ‘35? v IOSONIL WLZNZUxoxx wkckxwultmk \ IIIIIIIIIII IlwlllllIIllmImflIIlll \ OP. Om: Om“. 0w 00. ON. 06‘. 09. 00. 38/7LVEIER7’W34 :J- -19- unsuited for the reason that in every case there is no con- trol of the amount placed on the plate, the thickness of the film, or the distribution over the surface. It was decided, therefore, to spray the soil on the plate. The soiling apparatus (Fig. 9) in its final form consists of a fixed spray nozzle and soil reservoir supplied with air under constantly controlled pressure, a turntable for moving the plate under the spray, and a mechanism for adjusting the amount of spray. The process is completely automatic except for loading and unloading and starting. Once a plate is set in position and the process started, the soil film is uniformly deposited until the plate is covered. Thus a constant amount is assured for each plate. The Spray Nozzle Reference to Fig. 9a shows the spray nozzle, reser- voir, and air valve at A. The nozzle is a piece of copper tubing bent at right angles with a moderate radius to allow for the introduction of the small fluid tube at the angle. This fluid tube, also of copper, begins-at the reservoir bottom, pierces the cap, enters the nozzle tube at the right angle, and terminates at the nozzle opening. The fluid tube .carries the soil emulsion to the nozzle tip where it is atomized.by the primary air stream flowing through the noze zle. To facilitate movement of a fairly heavy and somewhat viscous soil, it is necessary to force it through the small tube with air pressure which is built up in the sealed re- servoir by tapping the primary air line. -20- Fig. 9 Plate Soiling Machin e -21- SOILING MACHINE . CONTROL We j. FOR TEST PLATESN I QRNTABLE _,._——— v ,fiFW e77 ~—.7.~ 7» 7 ,A,_.___. Fig. 9a. Diagram of Sailing IMachine -22- This secondary air.is fed into the reservoir through a small valve which thus becomes the control unit for start- ing or stopping the spray and determining the amount. Pri- ' nary air pressure is held constant by means of a reducing valve in the main compressed.air line supplying the appara- tus. An accurate gauge is connected to the primary air line. This secondary air is fed into the reservoir through a small valve which thus becomes the control unit for starting or stopping the spray and determiningpthe amount. Primary air pressure is held constant by means of a reducing valve in the main compressed air line supplying the apparatus. An accurate gauge is connected to the primary air line to measure the air pressure. 6 ab Various attempts were made to spray the soil-on the plate from a considerable height above the plate in an ef- fort to spread the soil evenly over the surface. However, this was found to be impractical so an arrangement was made for rotating the plate while moving it laterally under the spray. This has the effect of producing a spiral which be- gins at the outer edge and ends at the center of the plate, but the effect is not objectionable because there is ample overlapping of the spray area throughout the spiral. The spray area is about 2 in. in diameter and the spiral lines are about 1 in. apart. It is, of course, necessary to have the small area be- ing sprayed move under the spray at a constant speed. This -23— is accomplished by using a friction wheel drive on the under- side of the turntable. ‘Lateral travel of the turntable is produced through a rack and pinion. As the table revolves, the pinion on its shaft engages the rack fixed to the base and thus "walks” along. The table and its bearing are fixed to a second bearing at right an- glee which is the»horizontal slide and support bearing for the table. A handle on the bearing unit permits the table to be tilted about its horizontal axis, thereby disengaging the rack and pinion, allowing the table to be moved back to its starting position. Spray Control It was previously stated.that it is necessary to have the area sprayed pass the spray at a constant speed. This is accomplished in the machine through the friction wheel drive and rack and pinion. However, in spite of this pro- vision, the amount of soil sprayed on a plate increases considerably toward the plate center when the spray control. .valve is wide cpen during full lateral travel of the plate. This is true because as the center of the dish moves toward the spray, there is an accumulation due to the fact that the same size spray area is gradually working into a smaller plate area. Therefore, to produce an evenly coated plate, it is essential that the Spray control valve be gradually shut off as the plate center approaches the spray. The valve is operated by a cam and lever system shown in Fig. 9a at A and B. Not only does the can give the correct -24- valve behavior but it also makes the process automatic since it is directly connected to the lateral movement of the turn- table, being fastened to the bearing unit. Before plotting a cam profile, the actual condition of soil deposition across the plate diameter must be known. To determine this, two cellophane strips were laid across the plate at right angles to each other and cut into smaller sections of 1/2 1 11/16 in., the section at the intersection being 11/16 in. square. These were numbered accordhng to their position on the dish. After each section was carefully weighed, it was re-~ placed on the dish and the dish sprayed with a suitable soil. After soiling, each section was again weighed, thus giving the actual weight of soil sprayed on a given section. Proper correction was made for the center square. From these data, a curve was plotted showing the weight per a section against positions of the section. Such a curve is shown in.Fig. 10. This portrays the true build up of soil across the plate under the conditions of wide open spray valve for full transverse. It is obvious that the soil film is far too heavy in the center and defin- itely uneven. The curve also indicates that the valve action closing off the spray must be definitely controlled. The next step, therefore, was to study the delivery of the spray nozzle for various positions of the rotary valve stem. The rotary stem petcock valve was closed at 0° and fully open at 90°. Actually -25- quxawnx 20 0 m 0 >\Q\K.\ WORN b 0 m. 0\ .vxbx (-0 w>N.D U XVQQMI Q .1 VVOkaouzb t3kx>a zoxkbmxwk “(Q 4\QMJ 11 a. / V N \9 SWVU9/77/W CD 9. N -25- for all useful purposes, it was fully open at 36°, so nine valve positions were used. At each of these positions the amount of soil sprayed was determined by holding small cardboard squares under the spray, each for the same length of time; one for each valve position. The curve, Fig. 11, shows the results of this study. From it can be learned the exact valve Opening in degrees for any percentage of the total soil desired. With this informa- tion about the valve behavior and machine characteristics it was possible to proceed a step further in plotting the control can. I Reference to Fig. 10 shows curve O-A which pictures the true soil deposit across the plate. Only one half is shown because the full curve is symmetrical about the plate center. The ideal condition of uniform.distribution would produce a straight line as 00 which indicates a constant amount of soil at each position on the plate. To obtain this the control valve must be made to close gradually until it is entirely closed at position 9 (center of plate). To accomplish this, a curve 0-8 was imposed on the system from which it was possible to determine the proper percentage of valve Opening for each position across the plate radius. .Curve O-B is simply the reverse of O-A and when added to it algebraically about an assumed axis, 0-0 has the effect of cancelling,O-A to give 0-0. Valve position is, therefore, determined.from.curve 048 in terms of percentage of the full open position. For example, at position number 9 on the plate~the spray without valve -27- control deposited about 3.2 mgs. when it should have depos- ited only 1.8 mgs. To do this the valve must close a little. Just how much is found as a percentage from curve OéB. The vertical distance from the curve at position 4 to the horizontal line through position 9 divided by the vertical dis- tance from O to the curve attposition 9 gives that percentage. In this case it is 85 percent open. 'Likewise the percentage of cpening can be found for each position across the plate radius. In order to interpret the various percentages in terms of actual degrees of valve opening it becomes necessary to refer again to Fig. 11. Continuing the example of position 4, it will be noted that 85 percent of the total amount de- livered by full valve opening is 14.5 mgs. Following down vertically from where this valve hits the curve it is evident that the valve should actually be open 23° for position 4 on the plate. Therefore, each position on the plate has its particular valve setting,in degrees which can be determined from.the curve of Fig. 11. Once these values are known, it is an easy matter to plot them against positions on the plate radius which actually are positions of turntable traverse and thus arrive at a proper cam profile. The shape of the cam is such that the valve is opened quickly but shuts gradually according to the demand of the curve. ,A4ll.LI TQQM. QM» so 92,50 IRIS .20Rhflkfi 9Q 30%. SWV815/77/W -32.- Te demonstrate the accuracy of the soiling device, a soil containing,Micrococcus caseolyticus was spread over four series of 50 plates each using different batches of soil. From each run, three plates were selected and exam- ined for bacterial content. The results are presented in Table 40 Table 4 The number of bacteria deposited on freshly soiled plates by the soiling apparatus Plates Percent per Plate Bacteria mean Runs run No. Count deviation 1 50 1 203.000 25 182,000 50 194,000 5.7 2 SD 1 312.000 25 301.000 3 50 1 528,000 25 512,000 50 #90,000 3.9 4 50 1 120,000 25 127,000 50 118,000 4.0 The number of bacteria on the three plates of each run are fairly constant showing the accuracy of the machine in de- positing a constant amount of soil on each plate. The successful operation of the soiling apparatus herein described can be assured if a few simple precautions are ob- served: E 1. The soil must be very smooth in texture. Even very small particles in the mix tend to clog the spray nozzle. It cannot be very viscous or else it will not flow through the small fluid tube. Viscosity can be determined by experience -33- and should be kept constant. A simple drip test can be used to facilitate adjustment. 2. Air pressure on the primary air line supplying the nozzle must be held constant. Various pressures can be tried'but once a suitable value is found, it should be con- stantly used. 3. The entire apparatus should be thoroughly cleaned after use. This is especially true of the spray nozzle which clogs quickly, if any soil is allowed to remain after use. DEVELOPMENT AND CALIBRATION 0F PLATE SURFACE ANALYZER As the testing work progressed it was found necessary to have an accurate method for measuring the physical re- moval of soil from test plates after washing in a dish machine. To be able to state that one set of conditions produces better wash-off than another, means there must exist some basis for comparison between the results of verbose tests. A simple but somewhat unhandy and inaccurate way is to set up a ser- ies of plates which have been treated to show progress from a state of no-wash to one of complete wash-off. With such a scale of cleanliness, comparisons could be made and test plates rated accordingly, but even though a careful check is made, this method is subject to variations in Judgment which makes close comparison impossible. Photolometer devices have been developed and used by several other research workers, but application of this principle to the problem of analyzing plate surfaces was made for the first time in the work done by the author. -34- Gilcreas and O'Brien (1) in 1941 used a photo electric colorimeter (Luximeter) which passed a beam of light through a glass test slide. For many tests however, it was found that the area scanned by the beam was too small. Dr. Mall- man (2) prepared a modified photo electric calorimeter 'which passed light through approximately 1 sq. in. of the glass slide. This was an improvement but it was difficult to differentiate between clean and slightly soiled slides. Kenn and Ruchhoft (3) in 1946 reported a modified instrument in which an entire glass slide (1 x 3) was scanned by measur- ing all of the light passed with a photo cell. Six soiled slides were scanned at once and improvement was evident. Other workers have developed methods of measuring soil removal by shooting a beam of light upon an opaque surface and then determining the intensity of the reflected beam. When surfaces are perfectly smooth and flat such a procedure is useful, but such conditions do not exist with china dinner plates which were used in these tests to simulate the situa- tions of practical dishwashing. Therefore, in answer to a need for accurate comparison, a better method has been developed in an instrument by which examination of plates can be made quickly and soil retention metered with scale units. -The Surface Analyzer, as it will be designated, operates with reflected light and a photo electric cell, indicating the relative amount of soil reten- tion on the plate in the apparatus Fig. 14. It consists of a motor driven turntable upon which a plate is set to revolve . At», yo" -_ -35- Fig. 14. Surface Analyzer Fig. 15. Diagram of Surface Analyzer -35- during a test; a photo electric cell and light source mounted in proper relation on top of the box in such a way as to permit radial movement with respect to the plate; a sensi- tive millivoltmeter direct-connected to the cell} and finally a set of No. 6 dry cells to power the light. A diagrammatic sketch is presented in Fig. 15. A friction driven turntable is used to revolve the test plate so that an average reading may be taken on the plate surface. In this connection it should be stated that the speed of revolution must be high enough to prevent undue flicker or oscillation of the photometer needle during the testing of a spotty-soiled plate but not so high as to make it difficult to put a plate in or out of the instrument; or cause it to fly off the turntable. A speed of approximately 200 r.p.m. was found suitable. Preper arrangement of the photocell and the activating light source is very important. After some preliminary experimentation, it was found.that the photocell should be about léln. above the plate and that the light beam should strike the plate at a 45° angle. This permits the cell to register only a spot of light on the plate rather than re- ceive a direct reflection of the light source. Greater activation of the photocell could be obtained by direct reflection but less consistent behavior would be obtained since the angle of reflection tram the plate surface is not always the same, due to irregular and uncontrollable plate contours. Better operation, therefore, results from -37- placing the photocell vertically above the spot of light which is produced on the plate by light rays impinging thereon at a 45° angle. The light source is a microscope field illuminator, containing a simple condeasing,lens and small 3.8 volt flash- light bulb. This illuminator is fixed in a definite position relative to the photocell and the two, as a unit, are mounted atop the cover of the instrument in a way which allows them to be moved at will to any of four settings along the radius of the re- volving plate beneath. This arrangement-makes it possible to ”scan” the whole plate from the center to the edge. Thus because the plate is also revolving, an average reading for the entire plate can be obtained. It should be noted that provisions have been made to tip the photocell unit to conform to the plate contour near the edge, in order that this change in surface may automatically be accommodated. The photocell and meter used with the Analyzer, are actually a unit in themselves and together comprise an il-' lumination meter built by the Weston Company -- model No. 603. The scale is calibrated in foot candles with 50 divisions in the full swing of the pointer. In measuring illumination the foot candle scale is in ascending order but in the case .of the Analyzer, a reversed scale is used. when testing a dinnerlflate for physical cleanliness the greater amount of reflected light impinging on the photo- cell, the cleaner the plate, and, conversely, the smaller the - -33- amount, the greater the amount of soil. Hence a perfectly clean plate produces a full scale swing Of the meter pointer and should therefore, read zero, while a dirty plate produces less swing of the pointer and should show some number above zero. Thus a reversed scale shows the number of soil units on a plate. As has been mentioned briefly, the light source is powered by a set Of NO. 6 dry cells. These are connected in series with the lamp and a Rheostat, the latter being in the circuit to provide adjustment Of voltage and hence light intensity so that the meter pointer may always be set at zero when I clean plates accepted as standards are inserted in the Anal- yzer just prior to testing a series Of plates. The operation of the Analyzer is not difficult, but some precautions must be exercised and one should remember there are some limitations involved. The use of the instru- ment is limited to testing 9 inch dinner plates. These plates must be white in color and free of art work or de- signs of any kind- The soil used must be black to provide the necessary contrast in light reflected into the photocell. A further limitatiOn lies in the spotty wash Off on some plates. Occasionally test plates sent through a washing machine are not washed off evenly and may have areas where considerable soil remains along with adjacent areas which are relatively. clean. Obviously the.Analyzer is limited in its power to ascertain the actual plate condition by means of a scale reading alone but it can be stated that in most cases .0 -39- the average reading determined.by revolving the plate is reasonably indicative Of the soil conditions of the plate as a whole. An extreme condition of'this nature is seen in Fig. 16, where most of the soil is washed off on the edge Of the plate. However even in this case, there is usually some 8011 still left which is uniformly spread over the entire surface of the plate and the meter will show a usable reading. A more typical wash off condition is shown in.Fig. 17. In such an instance, the meter gives a good numerical picture of the residue. The chief precaution to be noted is the adjustment Of the meter for standard plate conditions prior to its use in test work. Ideally, there would be no need for adjustment if all test plates were exactly alike. ‘Unfortunately this is not true, so it was necessary to have a standard clean plate with which to set the meter pointer at zero before using the Analyzer. The use Of such a standard eliminates the need for a fixed voltage control on the light and,iin- cidentally, permits effective cOmpensation for light bulb' deterioration. The batch of test plates used in these studies was sorted into three groups; that is, plates in_a given group produced like readings on the meter. Hence it was necessary to use three standards, one for each groupl The groups were lettered A, B, and C, with letters painted on the backs of the plates so that all plates could be used simply by ‘9 -40- Fig. 16. Extreme Case of Spotty Washoff Fig. 17. Measured Plate Conditions Other Typical and Easily -41- setting the Analyzer with the proper standard before scan- ning any given plate. Use Of the instrument to determine the actual amount of soil left on a plate, after it has been washed, is pos- sible when reference is made to the curve of meter readings 6 versus soil weights as plotted in Fig. 18. This curve is typical of the behavior of photocells, when they are used to measure amounts of material by means of light reflection .Or transmission. The readings Obtained by scanning various plates soiled with different weights Of material are not directly proportional to those weights. However, once a curve is established for the meter and a given soil formula it is a simple matter to measure the actual amount of soil for any meter reading. For example, plates soiled to a level of 44 units would have about 9.7 gms. each of soil. If when washed, an average level Of 15 units was reached, there would be about 0.025 gms. of soil present. This gives a soil reduction of 97 percent. Obviously the instrument is more sensitive to small changes in soil weight in the range Of 10 to 35 units of meter reading but this is not detrimental to its use so long as the prOportionality curve is known. Not only has the Analyzer been found useful in determining the condition of soil on washed plates, but it also performs another useful task wherein it provides a ready means for checking the condition of plates during the soiling Operation. As the soiling of test plates is begun, the desired amount of -42- V? 0.0. milk-25, Om. 4-0.0 ON 0. mu\ «0Q Ukvfwflx QWQ .30“: K0 klelflE ea Nb USU [1 “1&ku {\QW «emu 0501mm NH W) q (2?. Quit-Q3“: _ ~20 50 . 410 m0 0.0 to 31V'7d 273d 9w v2.15 -43- ‘ soil can be established by reference to the Analyzer and con- stant check can be made on the soiling procedure by periodi- cally noting,the readings on the different plates. Properly soiled plates will only show a maximum difference of about one unit on the meter. It is important to keep this difference as small as possible, especially when the reading lies between 42 and 44, because the proportionality between these readings and the actual amount of the soil deposited on a plate either in terms Of weight or number of bacteria changes very rapidly beyond the 44 reading. In other words, a plate can only be soiled to a certain'black appearance beyond which the meter will not register additional soil deposit in terms of color involved. Table 5 Random plates taken from groups of 50 to show constancy of meter reading for each group to the number of bacteria in the soil. Greatest mean Group Meter Reading Bacteria Count deviation percent I 41"? 203,000 '5e70 . 182.000 194:999 II #0‘41 312 .000 +20 95 ' 301,000 ' 298st III 42.4} 528,000 T3092 512,000 490,000 It is apparent from Table 5, that when a certain level of soil is maintained, good uniformity in the amount of soil sprayed on successive plates is assured. From groups of 50 consecutively sprayed plates, three were picked at random and -44- tested for bacterial count. The results show that the meter can be used to control the soiling process. MECHANICAL MEASUREMENTS Rinse‘Ling Flog Conditions The first series of tests conducted concerned the deter- mination of the proper temperature and time for the rinse Operation. This involved several factors not the least of which was the measurement of water flow conditions. Both the amount of water used and its distribution within a machine were matters of vital interest. Flow line conditions in the five different dishwashers were determined to find amounts of water dispersed. Typical field conditions were maintained in the setup Of the machines but facilities were provided for very accurate measurement and control of both temperature and water flow. Various rinse line pressures were obtained through the use of a pressure re- ducing valve in the hot water system so that any flow pressure from 0 to 35 lbs. could be set at will on any machine. In this phase Of the studies, it was possible to study the rinse systems Of each machine and observe the effects of various pressures and volume settings as well as the influence Of the number and arrangement of nozzles. All five rinse systems were basically alike. They all controlled the water through a quick acting spring-loaded valve and sent it through several nozzles which broke it up into a fine spray. Distribution Of the spray was attempted in two ways. One provided upper and lower sets of fixed nozzles and the other used revolving arms, one lower and -45- one upper, each carrying several nozzle tips, and rotated by the inertia effect of the flowing water. It was a relatively simple matter to ascertain the pressure flow characteristics of any given rinse system; since each machine had a tank for wash water into which rinse water naturally fell during the Operation. All that was needed was to measure the time required to fill the tank by means Of the rinse nozzles, and knowing the tank capacity, the actual flow in gallons per minute could be readily cal- culated for any given line pressure condition. These calculations were made for each machine on the basis of flow pressures measured in the supply line at a point as near as possible to the rinse valve. All machines were tested with water at a temperature of 58° F. If hot water has any bearing on the flow conditions, it would tend to improve them rather than hinder, so any possible adverse- effect of using cold water rather than hot can be discounted. Rinse flow line conditions for the five machines are presented in Fig. 19. At a glance it is evident that three Of the machines do not give as large a flow as do the other two. In fact if a horizontal line were drawn through all the curves at a flow valve which would include all of them, the maximum flow possible would be 8 gal. per minute. To produce this flow required a pressure Of 38.5 lb/in2 on Machine 0, 30 lb. on Machine D, 22 lbs. on Machine E, 8 lbs. on Machine B, and 7.5 lbs. on Machine A. The great difference between pressures on Machine A and C is striking. I” -45- -<\ 00 \hnw MNOWWMQQ \SONux on On: . 0N o- 9 v o .a. a me a ml 3 7 \ R \\ mx .uQ WNXDVOL \SONux wsz thxwx 0. ON MO7_-/ N/N 33d 9N077vg NI -47- Both machines, however, had equal numbers of spray jets and the openings in them are of the same diameter, although the nozzles are of different types. The great difference in pressures required to give the same volume in both Machines A and C was probably due, somewhat, to difference in the type of nozzle, but mostly to the piping arrangement in the mach- ines. In the case of Machine E, piping after the control valve appears to be undersize in comparison with that in Machine A. A similar situation with respect to restricted piping either in small size pipe or number of turns and fittings after the valve, is evident in Machines D and E. This con- dition in the three machines, C, D, and E is probably the cause of such low volume values. Machines A and B had rinse lines with less resistance after the valves which appears to be a distinct advantage as far as producing adequate amounts of water are concerned. A second important consideration regarding rinse systems is observed from the curves in the matter of constant pressure. If, for example, a pressure of 30 lb./in2 was set on each machine, the delivery would be as follows: 7 g.p.m. for Machine 0, 8 g.p.m. for Machine D, 9.5 g.p.m. for Machine E, 14.5 g.p.m. for Machine B and 15.5 g.p.m. for Machine A. This is significant, in view of the fact that most manufacturers re- commend a minimum of 30 lb./in2 flow pressure in the supply line coming to their machines. Furthermore, because of the possibility of poor overall installations where machines may -48- not receive water at 30 lb./in2 pressure, it would appear advisable to redesign Machines 0, D, and E to provide greater water flow. This type information was sent to each of the respective manufacturers. It was mutually understood that each machine maker would get the data pertaining to his equipment only. This proved to be a satisfactory arrangement and provided each maker with that information useful for the improvement of his own product. Generally speaking, larger amounts of rinse water do better at sanitizing than smaller quantities. In Fig. 20 are presented percentage-kill of M. caseolyticus in each of the five machines of plate in position No. 2 (see Fig. 22) at 165° for 10 seconds at a flow line pressure Of 30 1b./in2. The correlation between volume of water supplied to each machine and the réduction in bacteria in the same.machine is very significant. It will be noted that Machines A and B having compara- tively large volume flow also had the highest reduction figures, Machine A with 15.5 g.p.m. and 99.60 percentage-kill and Machine B with 14.5 g.p.m. and 99.61 percentage-kill. It will also be seen that the reduction of bacteria in Machines C, D, and E varied in the same order as their res- pective volumes; Machine E with 9.5 g.p.m. and 99.34 percen- tage-kill, Machine D with 8 g.p.m. and 93 percentage-kill, and lastly, Machine C with 7.5 g.p.m. and 78 percentage-kill. Although other variables in design also enter the picture, these data suggest that vOlume Of rinse water plays an -59- .EQmw >1 0.10. 52340; 0— ON com 9me 0m. 5- M‘QDWONNB‘ >>0Nu\ NO? 20-k-00Q. 2- Miran-RN .ouw. OT. mfoux .2. teens ”UZ\IU(§< “3Q 2s kin.“ hmz\q Om OD. 7 vza3_Lovg NOLL anasaf % -50- important role in the proper sanitization or dishes. In another series of tests, the relation of rinse pres- sure and percentage-kill of test organisms was made with Machine.B.. In tese studies temperature of 170° F. for 10 seconds was used to determine the minimum rinse flow and pressure for effective sanitization. The data are presented in Fig. 21. To obtain 95 Percentage-kill it was necessary to maintain a flow Pressure of 12 lbs./in2. Machine B at 12 lbs. flow pressure passes 10.5 g.p.m. These results on Machine B and the previous data pre- sented in Fig. 20 would indicate that the minimum flow through the machines tested would be 10 g.p.m. at 1700 r. for 10 seconds. It would be well to speak briefly about another kind of test sometimes used for rinse systems by the manufacturers themselves. Since it is fairly obvious that proper and uni- form distribution Of water over the entire tray area is very important, a few dishmachine makers have attempted to test their systems by catching water from the upper rinse arms in beakers placed in the center and four corners of a dish rack. The assumption was that if equal amounts of water fell into each beaker uniform distribution was accomplished. From the physical standpoint this is probably true, but when bacteriological tests were made this was found to be in error. This can be readily shown with the data presented in Table 6. Here it can be seen that the system apparently distributed water quite uniformly, and therefore might be ex- pected to produce equally sanitized dishes in all parts of the dish rack. -5]: - \an NQBWWMQQ \SOVK ON 0- O- Wh.\<\mus \N .W\.u\ obukaQOMh-xu N>\ u-Zm-Zd‘ON-O .mumw O- .. .m-Z-F hoor- l WUF<>> Q U>§IU<2 Zx 22-010 .ovx WQ3WWWQQ \SOQIUx .3- cix< «(\kaufm 00 lm$ oo- [VD/.1 300337 V/c‘j3_l_ 3V9 °/e -c 22" Table 6 Water Distribution over Rack Area as Tested by Beaker Method during a 10 sec. rinse period Rack Corners Rack Center 2 - it? : 42 cc. 3 .- 4 - 40e5 N However, such was not the case with dinner plates as is evident in Table 7. Actually plates no. 1 and 2 (see Fig. 22) were not as well sanitized as the others, and the dis- crepancy was due to the fact that those plates improperly treated were not thoroughly reached by the spray in the particular machine under question. This machine had rotary rinse arms, both upper and lower, and during Operation thee spray of hot water and tending to recede from the surface of plates no. 1 and 2. Hence they did not receive as much water as plates no. 3, 4, 5. Figures for bacteria reduction show this very quickly whereas the simple beaker test did not. Table 8 shows results of similar bacteria tests on same machine after manufacturer corrected the system. wash System The testing of the washing prOcess called for the use Of a test plate soiled with a specially prepared composite food soil colored black with india ink and inoculated with a heat resistant spore-former called B. subtilis. The ink was added to make possible the use of the surface analyser previously described and the bacteria was used to provide a further and still more exact means for measuring soil removal. - 53- Table 7 Data Shothinsibility at Various Positions Within.The Machine .Test was made with six plates arranged as per sketch Water Temperature 165°F Time -‘10 sec. Hater'Pressure (Flow) - 30 lb./in2 Rack - By Champion Bacteria Count on Control Plates (Average) - 169,000 Plates slanting face up. Position (see Fig. 22) Runs Average 10,100 1 7,800 12,200 ‘ 18,600 14,200 2 8,400 11,800 12,800 1,100 3 ' 800 1,000 . ‘ 1,100 . 600 ' 4 1,800 4,800 12,000 2,400 5 . 2.300 ' 2,000 1,400 - ‘ ‘2- !i‘fl ‘F‘Jhfi-u ' Fig. 22. Position of Test Plates -55- Table 8 Data Show' Rinsibility at Various Positions Within the Machine With Corrected Rinse System Test was made with six plates arranged as per sketch. Water Temperature - 165°F. Time - 10 sec. I Water Pressure (Flow) - 25 1b./in2 Rack - Steel Wire Bacteria Count on Control Plates - 170,000 Plates slanting face up. Position Runs Average 5.200 l 4.900 5.200 5.600 . '4,700 2 '4,300 4,400 4,100 3.100 3 2,900 2.900 2,600 ' . 2,400 4 2,600 2,100 1,700 ' 1,000 5 2,100 1,300 900 I ‘ w~— -_ .__..,.. I - ., ~,,__ ,_ _-.-_-v_- ‘ll __'—._.. v—«--——V Fig.u2_3. 7 TypicTal Mash Pattern Tests -56... This test plate, properly cured, was consistently used to determine the effectiveness of the wash systems of the machines. The photographic results of a typical ween test are shown in Fig. 23. Use of the plate in this manner clearly shows that the machine under test is at fault be- cause the plates are unduly streaked by the upper wash sprays. Conditions for the tests as shown in Fig. 23 were as follows: Five plates were set in a rack as shown in Fig. 22, and these placed in a machine facing toward the right of an Operator who stands in front of the machine. They were then washed with water at 140°F for 20 seconds. NO detergent was used. Photo A-R.shows results of this test. A second test was made with five other plates washed under identical conditions except that they were facing toward the front Of the machine instead of toward the right. Photo A-F shows results of this test. It is clearly evident that plates in position A-F were washed better than those in A-R. This is true because plates in photo A-R received the flat wash jets on edge so to speak. Streaking resulted. The plates in A-F were turned through 90° with respect to those in A-R and they consequently received the flat wash jets in a way which tended to "spoon ” Off the soil over a large area per jet. Hence the better wash off. However, even in this position, it is true that the jets did not reach out far enough over the rack area. This test was 02 special interest because results shown in A-R were produced when the machine was loaded in the natural way, i. e., the Operator would load Q -57- the racks most quickly and naturally if the plates faced to his right or left. Best results however, were evident when plates were loaded in the rack facing front or back, an un- natural loading position. An obvious improvement which the manufacturer should make therefore, would be to turn the wash arms through 90° so the flat jets would produce the results shown in A-F; when an operator loaded the racks in the nor- m81 way. Tests of this nature gave useful information to the dish machine makers and resulted in many improvements. as a r 4 0 An attempt was also made to study the wash systems with respect to jet velocities and water flow conditions. To facil- itate this study a special manometer was designed and used to explore these conditions. . Pitgt 2gb; and Manomgtgr . The construction of the manometer called for an accom- modation of certain vital factors.‘ Reference to Fig. 24 . BhOws the actual instrument and Fig. 25 gives a diagrammatic sketch to show the principles involved. A special feature Of the manometer is the provision of an air release valve shown at A in Fig. 25. This permits water from the pitot tube to make its way right up to the U tube and does away .With the air column, otherwise trapped in the long lead tube. This makes it possible to use a long tube for more convenient exploration within a machine. Operation is as follows: 88 the water impinges on the pitot tip which is fastened ‘0 a wash arm directly over a~particular jet, 1‘5 moves along the 7 V M” 9.. Fig. 24. Jet-Velocity Manometer Used forgChecking Wash Water Flow D/A GPA/W OF MERCY/DY MANOMETER _4\ V k \X X \XX\ \\\ \X\\\Y‘ X\\\ T "' Fear] p/ror 72/05 . r1 6—: 5—55: 14—; —: N O llmlllll FIG.3_5_ h-- -59- tube forcing air to exhaust through the small valve cpening at 'A. As water fills the chamber the float rises to the top and shuts off the air vent. When this occurs the pressure from the pitot tube then actuates the column of mercury in the ~ U tube which gives a reading of the Jet velocity in term of feet of water. The surface of the water in the float chamber when it is full Just at the instant the air valve shuts, is at the same level as the zero point on the mercury column. Thus since the pitot tube entrance is also at this same level, there is no water column influence on the instrument. This air vent arrangement also eliminates any reverse movement of the mercury column so that when Jet velocity pressure drops to zero the mercury also returns to zero because the air valve at A Opens immediately as soon as the water in the float chamber begins to flow out. There is no question therefore but what the manometer starts initially at zero whenever Jet velocity reading is to be taken. The manometer was mounted on a stand 30 that it could be moved up or down to allow its being set at a level with either upper or lower wash arms. Fig. 26 shows a typical study made with the Jet velocity mB-nomet er . Water flog Volume Qetggmination Not only was the study of Jet velocities interesting and “Bérul, but it was also necessary to learn something about the V°lume of water delivered through the entire wash system. This was determined by the following method which yielded re- Bults that were sufficiently accurate for the problem at hand: -5;- JET VELOCITY STUDY P25551125: —— FEET or were: ”"L— 7 LOWER 5:7“ _9 L—qs p—l Inn-u ._ II- 2.‘ - - _ l‘ ‘6 _ 3 l 3.8- 7 5.4 :1 E. Z 2- .514 5-5.6 :31 I: :3 3-54 9~36 .._ —- a 4- 5.6 w-4.z ——1 —8 .__4 if j :4 5-3.6 “-4.8 1 2 ~ : 5-5.4 Iz~ .526 __J L__J t4 . I’— 6"—+—5 —“| AVEAeAcsE 4.67 -:* UPPER 557 *I—nl G. : '9 ‘0 HI. —- .. has 7-J.4_ N -I 3 — 7 .._1 . "‘ "" “'4 2-38 8 was 0 fiL —-1 L— K -+ h—0 fl" 3~3o8 9-2.2 ‘ 4 -——b ~— _‘ “-1 -—3. 0- . t] :8 ___1 4 G l 28 -:,- —-r : 5- 3.6 u ‘24 _>5 b—J ~L'1 -- j H 6— 2.8 02-23 —.J L—A l 9 ~ I 9" I AV£RA6£ \3. 09 onw TEST I P_UI_\_4_P—- S‘fAT/c 7.65/51: 'FLow 3.753044 WA 752 DELI VER v a PM. A VE. JET _ FLOW No. FLOW FLOW V54 P2. PER .247 .5407: P5! IBM ARA/s PE: jar Z OWE: 487 2.46 IZ 29.5 3 66. 5’ ukpee 3.09 2.06 / 6 342.9 3 9g. 7 TOTAL ~ /87.2' -62- A water meter was first checked to see how accurate were its readings. Having found its calibration curve it was then used to find the amount of water flowing through one slot in a particular wash arm. The arm was removed from the machine and connected to a water supply and the water meter. By use of the manometer previously discussed, the average Jet tel- ocity of a given wash system, i. e., 4.87 ft. of water, for the lower system (Fig. 26) was attained by adJusting the valve of the water supply pipe. Once this Jet velocity was established the flow of water through the nozzle was measured over a period of time by the water meter. When this valve was determined it was a simple matter to get the total flow for the system by multiplying it by the number of slots in use than adding values for the two systems. I The volume of water flowing through a wash system will have some bearing on the ability of the machine to wash dish- ware. This coupled.with the actual Jet velocity determines how effective will be the wash system. There was some attempt made to.correlate the wash ability of the various machines with the values of Jet velocity and water flow. However, nothing conclusive was discovered and little or no correlation was evident. There are many factors which bear on this prob- lea and a solution was not forthcoming in the use of the machines as they were set up. This study will require a more flexible system wherein the volume, pressure, Jet velocity and type of nozzle can each be controlled carefully and varied at will independently. Unfortunately it was not possible to -6} .. continue this investigation. It is felt however, that the method used.for measuring flow volumes can be used effectively in studies of this nature. PRACTICAL USE OF THE SOILING MACHINE AND SURFACE ANALYZER This paper has presented two special instruments. It has sought to show how they were conceived and made. Obviously the soiling machine was indispensable to the successful con- duct of the testwork, but it may not have been quite so evident as to how the Surface Analyzer was used. Actually this instrument was not found to be invaluable in the exact- ing work of determining soil removal or bacterial reduction quantitatively. more satisfactory results were obtained by using the bacteriological method. However, the Surface Analyzer is a useful tool and both it and the Soiler can be used effectively in testing commer- cial machines. An example of this is shown in Fig. 27. Plates first prepared with the soiling apparatus were run through a restaurant machine during a regular run period. Three plates were used. One was placed in the front row of a rackg another in the very center, and the third in the back row; all on a diagonal. Results were quickly evident. Photo- graphs were made of the plates for future reference, and Analyzer readings were taken. When meter readings are taken, they provide a simple but effective index to the behavior of a machine. Table 9 shows such a series of readings taken from.various restaurants in a large community. It is obvious at once that machine No. 12 is definitely unacceptable. -54. Machine No. 5 was doing an excellent Job. Thus, machines can be rated very quickly by a sanitarian. Table 9 Field Tests Using Test Plate for Checking the Efficienty of Dishwashing Machines Soil Removal in Soil Time in Seconds Units Position in Tray No. Type of Machine Wash Rinse Front Center Back 1 Single tank so 20 44* 4o 44 2 " " 30 2O 44 38 4O 3 " " 3o 20 24 12 20 4 " " 30 20 41 33 26 5 " " 3O 20 8 4 . ' 3 6 " " 30 2O 44 8 8 7 " ” 30. 20 44 24 10 8 Conveyor type 55 15 14 13 9 " ” 45 26 25 43 10 " " 60 20 11 17 ll " " 45 7 15 12 Immersion type 60 42 44 44 * A reading of 44 soil units is the initial soil reading. Readings of 10 or less indicate good performance. . 4.. ; 7‘! .. . ‘ r ‘ I . - ’3 ‘ ‘ fl . .f . - _ e. _ ‘ , g . i .2 m 4‘ vi'fi' 3‘s . ‘~-" - ~ 7 ‘ ‘ '_ ‘A ‘ ‘. 4‘ f - . 5'". Fig. 27. Typical Results of Restaurant Test .65 - It might be suggested that any meter reading above 15 is unacceptable. At least it is evident that some standard could easily be established by using this method and the two instruments. Thus the Public Health.official has at his disposal useful tools for checking dishwashing machinery in restaurants anywhere. anglusion The work.herein described demonstrates quite clearly .how the mechanical engineer can serve other branches of science. In this case the science of bacteriology was aided. As workers in.that field sought to furnish public health authorities with useful data on the problem of commercial machine dishwashing. Instruments and technics developed for this test work should be found useful to sanitarians .and research workers alike. It was also true that results Obtained by these tests were such as would benefit the manufacturers of dishwashdng equipment. -55.. REFERENCES (l) Gilcreas, F. W. andJ. E. O'Brien. Laboratory studies of eating utensils and evaluating deZergents. Am. Jour. Pub. Health 31: 143, 19 1. (2) Maiimsnn, w. L. Unpublished data. (3) Mann, Edward H., and C. C. Ruchhoft. A per- formance test for rating dishwashing detergents. Pub. Health Rep. 61: (No. 24) 847, 1946. ‘ . _ i. ‘\ -.‘ «. ,-. ~. ~~ __ ( I I - U :s 'y \ . I . I . . ‘ 4 ‘ - ’ » t " ' ‘ l‘ I .- ' - AP z I 53 . ' ' a ‘. IDA/I ' Mai/27’.“ ‘ I A I I It ’ u 1 l | ‘ i a x ‘ l ' ‘ ‘. . \ 2;: - . ’.4‘ e . ‘. \\ i! e \- I u- . 5 ‘ 7 \ a I v ' I I I I I J I l O ‘I I I! - {I . [I I _ ' f . s I I . _ l . . .. J a ' \ - ,1 ~ I . k . I I' 5 "- . i ' \ III | . . i J I ' a \ . h I If - _. k". - - ’ ‘ | _ I I l . . v ‘ , ‘ . ‘ I . ‘ u. '. . . ~ I I. _ ' ‘I I I - _’ l , \ o x I -/, ' ‘ ., x ‘ ‘ ' . - O" . | f 'K ‘ v . . ~' V ‘ I K ‘, ’ Y n ’ ' . .I 't- I I ' | MICHIGAN STATE UNIVERSITY LIB I III IIIIIIIIIII‘I‘II‘I’“ 3 1293 030710010