PUMP LAB - PART 1
by James W. Daily
Reminiscences of Days at the Caltech Pump-labby James W. Daily
The Hydraulic Machinery Laboratory at Caltech had its genesis with some graduate student thesis projects in the early 1930s under Dr. Robert T. Knapp. Bob Knapp had spent 1929-30 as a Freeman Fellow in Europe visiting hydraulic laboratories and other hydraulic facilities and factories with a special interest in hydraulic machinery. He came back with much enthusiasm and many ideas. The sharpest was his conviction that there was a need in this country for more definitive research into the hydraulics of centrifugal pumps, that it should be done in a laboratory independent of all manufacturers, and it should be located at Caltech.
The time was ripe for many reasons. First, the Europeans, and especially the Germans, had a lead in pump research and technology. Second, Southern California needed water and there was a project underway to build a great aqueduct from the Colorado River to Los Angeles and adjacent cities and districts. It was to be some 300 miles long and a series of five pumping plants were necessary to lift the water some 1600 feet over the mountains. The pumps were to be among the largest and most powerful installed anywhere in the world. There was a fledgling organization, the Metropolitan Water District of Southern California (MWD), which had designed the project and was busy initiating the construction.
The beginning of the pump research program was in an "initial" pump lab, as George Wislicenus, one of Knapp’s early students, has called it. This was in the North end of the Caltech boiler building where there was some space for experiments. The first experiments were performed as thesis projects by some of Knapp’s graduate students. Wislicenus was one of these. "Wis," as he was soon dubbed, was a young pump designer from Worthington Pump Company in Harrison, N.J., who had come to Caltech for graduate study. In 1931 or 1932 he carried out various pump tests which figured importantly in the efforts under way to induce the MWD to support a more sophisticated laboratory effort aimed, of course, at the problems that might arise with the exceptional pumps that were required for the new aqueduct.
The "final" pump lab came into being when a contract was signed on November 14, 1933, for Caltech to build a laboratory and to conduct experiments and tests for MWD. (Incidentally both "initial" and "final" are names that Wis used to distinguish between the sites of the early tests and the later MWD program.)
The MWD Pump Program
The exact steps leading to the contract with MWD are not clear but the agreement seems to have been the result of an orchestrated effort that began in 1930. In addition to Knapp, Professors Robert L. Daugherty (M.E.), Franklin Thomas (C.E. and Chairman of the Engineering Division) and Theodore von Kármán (Aero and Director of GALCIT) were involved as well as R. A. Millikan. In 1930 Daugherty and Thomas were active in Pasadena affairs. They had been or were City Directors. At some time in that era Daugherty was mayor. Thomas was also Pasadena’s representative on the MWD Board of Directors. Kármán, who was a new and widely respected professor from Aachen, Germany, became interested in seeing the development of a program in hydraulics. Moreover, Millikan was showing interest.
As early as December, 1930, a conference with MWD officials that was attended by Knapp, Kármán, Daugherty and Millikan disclosed MWD’s interest in a pump testing facility, and that MWD foresaw many hydraulic problems of scientific importance. It was concluded that Caltech should make a careful study of the possibilities. In brief, the studies were made, there was some iteration and, following the inevitable delays, the contract was finally signed on November 14, 1933.
The laboratory was planned to occupy the tall three-story bay and sub-basement at the West end of Guggenheim and part of an East-West channel along the Northern face of the Guggenheim basement and sub-basement. The three-story bay was intended originally to house a large universal testing machine. The channel originally was to have been a towing tank.
Design of both instruments and equipment began immediately after the contract was signed. Wislicenus designed much of the main system and instrumentation under Knapp's supervision. The lab and its instrumentation was designed to provide accuracy (0.1% to an individual test point), to eliminate the personal equation in taking data, to use primary standard type of instruments and to have the utmost flexibility. Special features included weighing type pressure gauges and venturi manometers, a quartz crystal time standard, a feedback device for comparing the dynamometer speed with a standard speed and automatically providing corrections to overcome instantaneous differences, and a system for "freezing" the instrument readings when conditions were stable. This last was accomplished by both of two operators simultaneously exercising their individual controls to interrupt the current flow to the instrument readout bank.
The laboratory was calibrated and in full operation by September 1934. Today Wis assumes a modest attitude about his role. Actually his instruments used in this laboratory and in the later CIT Water Tunnels constructed during and after WWII were exceptional and excellent and later were copied extensively.
It should be said also that Professor Knapp’s vision in the whole project including laboratory plan and most of its details led to an accurate and efficient data collection system which was the forerunner of the systems now used worldwide in modern laboratories. The latter made use of solid state devices and computers, neither of which existed in the 1930’s. The pump lab instrumentation was a combination of mechanical and electrical devices and vacuum tube electronics; it made up an advanced and sophisticated system for its day.
Incidentally, the laboratory instruments (as well as many later precision machined devices for the lab) were made by Fred C. Henson’s shop on East Colorado.
Among those who took a deep interest in the pump lab development was Aladar Hollander who later was a professor of M.E. at Caltech. A.H. (as he was known affectionately) was chief engineer for the Byron Jackson Co. Pump Division and in one sense a bystander. However, he was well known for his expertise and his sage advice and his support of the project was invaluable. < A.H. was a student at the Technische Hochschule in Budapest during von Kármán’s time. (Kármán graduated in 1902. I believe A.H. graduated a year or so later.) A.H. had been at the Byron Jackson Co. since the 1920’s. Until 1936 or 1937 when the Pump Division moved to Huntington Park, his headquarters were in Berkeley at the manufacturing plant. (My first engineering job was working for A.H. in Berkeley in 1935 after graduating from Stanford in March of that year. Incidentally, I felt fortunate in landing a job. They were scarce! I was paid $10 per month more than I earned in 1930-31 as messenger boy in B.J.’s Oil Tool Division in Huntington Park. The salary in Berkeley was about $80 per month.)
For a long time preceding the move there had been a B.J. pump staff for sales and service in Huntington Park which brought A.H. to Southern California frequently and facilitated his contacts with the MWD project, with Caltech’s Robert Knapp, Robert Daugherty and, of course, his old friend von Kármán. About 1938 or 1939 he was designated as B.J.’s consulting engineer and passed the chief engineer’s duties to Carl Bloom (formerly of Bingham Pump Co. in Portland, Oregon). A few years after WWII A.H. came to Caltech as a faculty member.
I should add that Aladar Hollander was a friend of all young aspiring engineers. In my case he encouraged me to go back to graduate school. Later he wrote me (I was back at Stanford) about a young man, Prof. R.T. Knapp, who was conducting some interesting pump experiments at Caltech and I should hurry down to Pasadena. As a result I came to CIT in December of 1935 and moved into the Old Dorm without realizing that this was the start of an eleven year stay!
A few years later Hollander was commissioned to hire me back to B.J., but instead he visited us (I was married by that time) and advised me to stay in graduate school and complete the Ph.D. program. I have often wondered what A.H. told E.S. Dulin who had sent him. Dulin was President and took active charge of all of B.J. for many years.
In the beginning of the MWD program the pump lab experiments looked into a variety of phenomena as well as the basic behavior of centrifugal pumps. The effect of cavitation on the performance of pumps was examined in a detail not done before. The suitability of single-stage, single-suction, volute pumps for the MWD application was investigated. In addition radial thrust measurements were made with the results leading to specifications for heavier shafting than normally used by the manufacturers. The basic impeller flows and volute effects were studied by measuring the impeller out-flow velocities and directions versus vane tip position. Ray Binder conducted these impeller studies as his Ph.D. dissertation under Knapp. Professor A.L. (Maj.) Kline suggested a special sampling valve which was constructed for these measurements.
The first phase of the MWD program also included determination of Complete Characteristics, i.e. a four quadrant performance plot of lines of constant head and power on a chart of discharge versus speed giving alternate modes of "pumping" and "turbining." Such data was used in calculations of pressure surges (water hammer) during startups and shutdowns.
The result of this preliminary work was to "tighten up" the final specifications for the pump purchases in various ways from what MWD might have used if the prevailing "wisdom" of many pump engineers’ recommendations had been followed.
When I arrived at Caltech Ralph M. Watson was in charge of the Hydraulic Machinery Laboratory under the triumvirate of Kármán, Daugherty and Knapp. Wislicenus, who had finished his Ph.D. program in 1934, had gone back to his firm (Worthington Pump & Machinery Co.) in the summer of 1935. On arrival I learned that a man named Frank Wattendorf had preceded Watson. Frank was from Aeronautics and Kármán had arranged that he go to Peking to build a wind tunnel at Tsing Hua University. This, I was told (by the boys), created a vacancy, at the lowest level of course, which I filled. This job allowed me to continue graduate studies on a part-time basis. (Later such positions came to be called research assistantships.)
By December 1935 four pump manufacturers had submitted "bidders models" for testing to decide which ones, if any, were to be awarded contracts. The awards were to Byron Jackson (Intake and Gene Plants), Worthington (Eagle Mountain and Hayfield) and Allis Chalmers (Iron Mountain). These contracts covered three pumps of the total of nine planned for the ultimate installation in each of the five pumping plants.
At this stage the active lab staff was small and included some on temporary duty from MWD as well as the Caltech employees. The people from the MWD office were A.W. (Bill) Atwood, an E.E. graduate of Caltech doing hydraulics temporarily (who in due course became my brother-in-law) and two other E.E.s, Paul Winn and Harold Levinton. The Caltech group included, as I remember, Ralph (Pop) Baker (older of course, maybe 35!), an Aero graduate student, John Konecnik, an M.E. undergraduate, Ray Binder, an M.E. graduate student, Ed Simmons (of strain gage fame) and Rudi von Huene (Geology and Minerology). Later, Pop Baker completed his Ph.D. program and returned to his home state, Utah, where he was on the faculty of Utah State University until he retired. Ray Binder left after his Ph.D. in 1936 to teach at Cornell, Purdue and finally at USC.
It should be added that a mainstay of the lab was a mechanic, Ray Kingan. Ray is deceased but his son Jack is employed at Caltech currently.
Soon after my arrival at Caltech I met Bert Fenner who was both purchasing agent for the Institute and in charge of the "wiring shop" which handled all the electrical maintenance for the Institute and all its labs. I had many pleasant dealings with him throughout my stay at Caltech.
Later I had similarly satisfying contacts with Wesley (Herky) Hertenstein who was in charge of buildings and grounds.
All of the bidders models submitted for the MWD tests of December 1935 were designed to operate at speeds which would give the full prototype head. This gave water velocities inside the pump equal to those in the prototype machine. There was some disagreement among engineers at the time as this differed from the practice with turbines of using straight Froude number modeling and hence very low speeds and internal velocities. The Caltech method, however, gave a high model Reynolds number, a compromise with the even higher and unattainable Reynolds number inherent in the prototype.
One fact that helped both the initial bidders and the awardees was that the requirements for Intake and Gene plants were nearly the same. Consequently, one model was required for the pair. The same held for Eagle Mountain and Hayfield.
As the next step each successful bidder submitted a "contractors model" which was to a larger scale than the bidders model but also designed to operate at a speed giving the prototype head. These were thoroughly tested to verify that any changes from the bidders models did indeed satisfy the specifications.
Simultaneously with the testing of the contractors models some other investigations were made for MWD. One of particular interest at that date was measurement of valve pressure drop versus opening. Such data were especially useful for combining with the previously mentioned Complete Characteristics of pumps when calculating hydraulic transients. To make a comparison of gate valve and plug valve behavior Arthur T. Ippen (Ph.D. 1936) was attached to the project in the spring and early summer of 1936. This resulted in plug valves being specified for the discharge cutoff valves which followed each pump in each plant.
Incidentally, Art Ippen, who was from Germany, went to Lehigh as a faculty member in 1938. Following WWII he went to MIT where he became head of the Hydrodynamics Laboratory. After a very successful career he died suddenly in 1974.
The project ended in the summer of 1936. The MWD personnel returned to their Los Angeles office and the Caltech students either graduated or went back to school full time. Watson went to New Jersey to employment with Worthington, Kingan became the mechanic in the old M.E. shop and laboratory. Daily became the "curator" of the suddenly quiet pump lab as it awaited its next challenge. I believe it was at this time that MWD passed title to its interest in the laboratory to Caltech.
As "curator" my duties were to look after the equipment and the instruments. This included operating every now and then, oiling all parts and instruments and so on. One day when entering the laboratory there was a strong odor of gases such as rocket exhausts. There was also the remains of a pendulum supported from the ceiling two floors above. It seems that Frank Malina’s crew had used the dormant lab for their rocket experiments. Objections were raised and they were banished to an outside site at the East end of Guggenheim (where Firestone now stands). I didn’t know (nor did they for that matter) that this early experiment was one of the first steps leading to today’s giant Aerojet Corporation.
I should add that the MWD tests did result in large overall savings and avoidance of future problems such as deflections due to radial thrust as well as avoiding excess vibration and materials damage which cavitation would have caused. Moreover, many of the questions investigated affected the first cost of the pumping plants as a whole in addition to affecting the costs of the pumps themselves.
The field tests of the completed initial three units which took place later (1939) proceeded smoothly and the results followed the predictions based on the laboratory tests.
I should add that there was one perturbation to interrupt the "smoothness" of the field test procedures. These tests used the "salt velocity" method to measure the water flow rates in the 10 ft. diameter penstock following each pair of pumps. In this method a salt solution is injected to be entrained by the flow and the water velocity is then measured by timing the salt’s passage between two downstream sets of electrodes. The method depends on thorough mixing of the salt solution with the water flow. The first determinations gave too low a value, as everyone agreed. It was surmised that better radial mixing of the salt cloud was necessary. Baffle turbulators were added and the measured flow rate increased! Agreement was reached on the acceptable amount of baffling to use and the tests proceeded. These tests were supervised by Prof. L.J. Hooper of Worcester Poly.
Bill Atwood tells the following anecdote about the field tests. "The tests on model pumps proved that a single stage centrifugal pump would meet the requirements and an efficiency of 88% or better could be achieved. Therefore the final specifications (MWD 116) required a guarantee of 88% efficiency and provided a bonus for each percent achieved above that. Failure to meet the guarantee required the manufacturer to modify or replace the pump."
"Paul Winn and I, having spent several years as MWD representatives at the Caltech pump lab, were both deeply involved in the final field tests. The tests at intake and Gene were witnessed by A. Hollander. He would look over our shoulders at the readings as we recorded them and retire to a corner of the pump house and work his little slide rule. We would glance over and if he was smiling all was going well, but if he was frowning we would check our instruments. After the run was over Paul and I would retire to the conference room and with the calculator apply the various corrections and grind out the results. However, from Hollander’s smiles we already knew they would be good. The final results showed an excellent 90.9% for the five pumps that were tested. That evening A.H. treated the entire test crew to a case of special German wine he had brought along in anticipation of such excellent results."
"The low head pumps at Iron Mt. did even better, obtaining 91.3% efficiency. At Hayfield, which had the highest head of all the plants, the efficiency tested slightly below the required 88%. Therefore the MWD engineer directed the manufacturer to modify the inlet vanes and impeller. This the manufacturer did, spending many weeks of slow grinding in the field. By this means he was able to bring the efficiency up to the guaranteed 88%."
As another postscript, the subsequent careers of the principals of the early pump lab staff are interesting. Frank Wattendorf remained close to Kármán after returning from China and was his right hand man until Kármán’s death in 1963. I believe Wattendorf is still associated with AGARD-NATO.
Wislicenus went with Packard Motor Car Company on defense work during WWII. Then he taught at Johns Hopkins until the early 1950’s. Wis then went to Pennsylvania State University where he headed the Naval Ordinance Research Laboratory and Garfield Thomas water tunnel and simultaneously was Head of the Aeronautics Department until his retirement.
Ralph Watson remained with Worthington until the mid 1950’s when he went to Syracuse University as Professor and Associate Dean of Engineering until his retirement.
The USBR Grand Coulee Pumps
There were rumors during the MWD days of some interest by the U.S. Bureau of Reclamation (USBR) in pump studies for the Grand Coulee irrigation project. Of course, we hoped it would come Caltech’s way. Grand Coulee Dam was under construction and the USBR was designing a pumping plant to take impounded Columbia River water to irrigate the higher Grand Coulee plateau. There was need to investigate not only pump problems but other hydraulic problems in connection with the planned pumping plant.
By late 1937 the rumors became solid gossip. Something important was under way. Finally, on December 15, 1937, a contract was signed with USBR for the studies to be done in the Caltech pump lab. This contract was for one year originally but was amended later for a total of two and a half years.
The Grand Coulee irrigation needs were large. Each one of 12 pumps was to deliver 1600 cubic feet per second (cfs) at 295 ft. head. This discharge was eight times the capacity of each MWD pump. One pump was to be driven by a 65,000 hp motor, exceeding the previous record size of 27,000 hp motors used to drive the single-stage pumps in a pump-storage plant in France. There were 12 pumps planned and they were to be powered by the output from 6 of the Grand Coulee hydraulic turbines. Moreover, each Grand Coulee pump was to operate over a range in head from about 275 ft. to 367 ft. (corresponding to about 1100 and 1650 cfs respectively) depending on the reservoir seasonal elevation. Flow instabilities and cavitation were possible problems with such a wide range of operating conditions.
The experimental program began in January 1938 and continued until the summer of 1940. Donald P. Barnes (MS ’30) was representative and resident engineer for the USBR. I became laboratory manager for Caltech. In between the duties of this exalted position I continued graduate studies on a part time basis. (Courses from Epstein, Kármán, Clark Millikan, Daugherty and several others kept me very busy.)
I should note an important change that the USBR contract made possible. During the MWD tests operation of the main dynamometer was restricted to overnight scheduling because the only powerful source of D.C. current was the motor-generator used in the daytime by the large GALCIT wind tunnel. This serious handicap was removed by the installation of a separate pump lab motor-generator set. USBR advanced the costs of equipment and its installation and took repayment as a monthly credit against the laboratory’s rental fee.
The laboratory was still directed by the same three persons, Kármán, Daugherty and Knapp. Following the earlier tradition established during the MWD days a series of weekly meetings were started to discuss all phases of the pump lab’s program. These meetings continued for the two and a half years’ duration of the USBR program and were attended by the triumvirate and by Barnes and Daily. As well as considering the usual administrative problems many technical questions about the dynamics and fluid mechanics of centrifugal pumps and related equipment were discussed. Everyone entered freely into the discussions (as well as the "debates" that sometimes developed). These conferences provided much intellectual stimulation.
The staff employed on the project at one time or another included in addition to Barnes and Daily:
In addition Hugh Bell, who was on the CIT Soil Conservation Laboratory staff, was our photographer and consultant on related matters. Also Prof. H.F. Gauss of the University of Idaho, a visitor on his sabbatical, relieved his boredom by working with us.
- Lillian Price (secretary)
- John Bigelow (mechanic)
- Bob Brumfield (ME undergraduate)
- Phil Deverian
- Muriel Duer (calculator)
- Herb Ellis (ME graduate student)
- Andy Fejer (Aero graduate student, a Hungarian from Czechoslovakia)
- Cal Gongwer (ME graduate student from Columbia University)
- Tom Harper (applied physicist)
- John Konecnik (later John Kaye) (ME undergraduate)
- Joe Levy (ME from UC Berkeley)
- Carr Liang (ME graduate student from China who returned home in 1940 to join in the war against Japan’s invasion)
- Walt Moore (CE graduate student)
- George Morikawa (ME graduate student)
- Brooks Morris (CE from Stanford)
- Rolf Sabersky (ME undergraduate; currently on ME faculty)
- Sander Sheff (draftsman)
- Ed Simmons (EE from Caltech)
- Ed van Driest (CE graduate student)
- Dave Vasquez (calculations)
- Dave Wood (ME undergraduate; currently on ME faculty)
- Dave Young (ME undergraduate)
There were also others who participated in the USBR program but their identities escape me now.
The program objectives were to conduct laboratory research and tests on model centrifugal pumps and related works for the Grand Coulee installation with the aim of establishing the bases of sound specifications for the required equipment. The studies were to provide general information rather than to judge the relative merits of units produced by different manufacturers.
These objectives directed attention to ways of obtaining high efficiency over the head range expected, the optimum operating speed, the best type and design of pump casings and inlets, and the best impeller behavior. Freedom from hydraulic instabilities and cavitation problems was essential.
Pumps with five different casings were investigated. All were single suction with overhung impellers. The casing types (as illustrated in Figure 1 on the following page) included
All of the above permitted tests of many combinations with the appropriate impellers.
- Single volutes, double volutes, and fixed vane diffusers. There were two different designs of each for the alternate proposed prototype speeds of 150 and 180 rpm.
- A three vane volute designed for 180 rpm prototype speed.
- An adjustable wicket gate pump designed for 150 rpm prototype speed.
The program objectives and the several lines of investigation were so thoroughly interrelated that it was expedient to work on them simultaneously. Thus when a pump unit was on the stand all the experiments necessary for each of the investigations were made. For example, a series of experiments often included Complete Characteristics (as explained previously) for the combination, cavitation behavior of the particular impeller over a large range of discharges, and auxiliary items such as velocity profiles at pump inlet or discharge or pressure measurements along the casing walls and on the casing vanes.
The interest in various casing types stemmed from both hydraulic and structural requirements. An important question for the large pumps being planned was how to strengthen the prototype pump’s relatively thin walled casing against internal pressures and external loads. The usual solution was to provide external ribs, but this was undesirable for such large pumps because ribs increased the overall size and weight, and hence the cost of both equipment and excavation.
Thus, the casings examined were all investigated from the above viewpoint as well as their hydraulic features. The double volute increased the casing’s strength as well as providing better radial balancing of pressures at the impeller discharge and reduced radial thrusts, all with only a small increase in size. The other options to the single volute were all larger and heavier. It is interesting to note that when the final design of the Grand Coulee pumps was completed some years after WWII a double volute was used but modified by four extra partial vanes to assure sufficient strength.
I should add that the idea for a double volute was not new. It was described originally in the European engineering literature in the early 1900’s but had never been developed beyond the idea. The concept lay fallow for about 35 years before being rediscovered for the USBR program.
In addition USBR needed information on the effects of inlet structure design on pump performance. Again, costs of equipment and excavation were concerns. Thus the early experiments included tests of various inlet options. The USBR submitted designs for a short and a long radius reducing elbow. A special "Bowl type" inlet was designed by A. Hollander (and constructed by Byron Jackson as I remember). It was patterned after inlets used for double-suction pumps and was smaller than a conventional elbow. These inlets were investigated for effects on overall performance and cavitation and compared with straight axial inlets.
During the MWD program there was reluctance by some manufacturers to participate in studies where information sharing might disclose trade secrets. This reluctance persisted when the USBR program started. However, the project was able to purchase equipment from two manufacturers (Byron Jackson and Worthington). A third manufacturer (Newport News ) by arrangement with USBR submitted a model pump-turbine for testing. In addition, USBR submitted designs for pump cases and inlets that were then built for the studies.
As during the MWD tests Hollander was the chief contact with Byron Jackson. During the USBR program George Wislicenus came back to Pasadena from New Jersey for several months in late 1938 and early 1939 to be a sort of "resident" representative of Worthington for that company’s several models and versions. Another Worthington representative during the entire program was H.P. Henderson (BS ’26) of the Los Angeles office. Jack Reilly came from Newport News to witness the tests of their adjustable wicket gate pump-turbine model.
I first met George Wislicenus in Cambridge, Massachusetts, at the 5th International Congress of Applied Mechanics in September 1938. I believe that Ralph Watson introduced us. (Ralph had been at Worthington since he left Caltech in 1936 and worked in the same group as Wis.) Very soon after the Congress Wis and his family returned to Pasadena (as previously noted) for a several months’ stay. He came to observe tests and to supervise further development of Worthington pumps and impellers, the designs of which might be used later for Worthington’s Grand Coulee proposal.
To avoid delays in the excavation process for the Grand Coulee pumping plant early information was needed about the pump inlet design. Thus the first experiments were tests of the various inlets mentioned above. As each was tested for its possible effects on pump performance and impeller cavitation, velocity distributions were measured as the water entered the pumps. These experiments showed that the overall performance was little affected by any of the inlets, the cavitation behavior was best with the "Bowl type" unit, and the "Bowl type" required the least excavation, with the short radius elbow the next best.
It is interesting to note that ultimately an elbow was used at Grand Coulee. Apparently excavation proved less of a problem than thought originally and possibly the multicurved casting required for the "Bowl type" design proved too costly.
The inlet studies provided the core of the program through the summer of 1938. Incidentally, the new casings previously described were not yet available and all the inlet studies used one of the contractors’ models from the MWD program.
After the concentration on inlet designs, the studies of the several pump casing types and impellers were undertaken. These tests of many combinations of casings and impellers consumed the next two years. There was an interruption, however, for some experiments for another government bureau. Although the USBR had contracted to be the primary user of the Caltech pump lab some time was relinquished to the Bureau of Indian Service during late 1938 for development work and model tests of pumps for the Flathead project on the Flathead Indian Reservation in Northwestern Montana. This was another irrigation project similar to the Grand Coulee job. The program involved the suitability of the proposed pumps and especially an impeller that would meet the Flathead project’s cavitation requirements.
This was an interesting situation because the tests followed the submission of bids, but the Indian Service wanted to test the models of both bidders (Worthington, with the lower price, and Byron Jackson). The manufacturers accepted the idea and tests were made.
On completion of the Flathead tests in mid-December of 1938 attention was given to the adjustable wicket gate pump submitted by Newport News. The Newport News pump arrived about a week before Christmas 1938. It went on the test stand in the next few days and when the New Year of 1939 dawned tests were under way. As already noted, Jack Reilly accompanied this pump to witness its testing and handle emergencies (which a bearing soon provided). These studies and their analysis lasted 6 weeks or more. This type of pump offered certain advantages, especially under reverse flow conditions when the wicket gates can be used to help control pressure surges and runaway speeds. However, its design was not was simple as a volute or fixed vane diffuser.
During the remainder of 1939 and the early part of 1940 investigations continued of the various pump designs, particularly those with fixed vane diffusers, and of the more than thirty variations of impellers. During this period several auxiliary investigations were undertaken to support aspects of the basic studies.
One interesting investigation was peripheral to the specific Grand coulee pump requirements but was of USBR interest because it involved the "salt velocity" test method (see p. 5) that might be used for field testing the Grand Coulee pumps. In the experiment a transparent pipe was used to observe and measure the diffusion and settling of dyed salt solution clouds. This topic grew out of the 1939 MWD field tests where the special mixing needs for low velocity flows in large pipes were found.
There were several capable and interesting staff members during the USBR days. One was Cal Gongwer who was with the USBR program until he got his MS in 1939 and went to General Motors Labs in Detroit.
Cal came to CIT in 1937 fresh out of Columbia University (and incidentally never ceased ragging me about the 1934 Rose Bowl game when underdog Columbia beat Stanford). He was a young "inventor" with lots of refreshing ideas and wrote many internal reports with his ideas supported by calculations and experimental data. Included were several ingenious treatments on cavitation in pumps explaining some of the Caltech findings during the USBR program.
Later Cal presented this latter material in a paper "A Cavitation Theory for Centrifugal Pumps" which was published in the 1942 ASME Transactions. It proved to be a widely quoted reference.
Cal stayed at GM until 1941 when he joined the war effort at the Naval Under Water Sound Laboratory (at the behest of Knapp who was a consultant to NUWSL). In 1945 he returned to California to work at Aerojet in Azusa (this time at Kármán’s urging) where he stayed until 1966 when he left to establish his present consulting business in Glendora to design hydropropulsion machinery and other exotic devices.
One of Cal Gongwer’s pump lab assignments led to a social gathering. He was asked to design and build a set of controls to keep the suction tank water level constant when the pump operated on "open circuit," for example during venturi calibrations. On the job’s successful completion the entire lab staff (including our secretary, Lillian) celebrated the event with a swimming party in the renovated tank (with swim suits, of course).
Another mainstay during the USBR program and very successful in his later career was Brooks Morris (Stanford 1934). He came to CIT in 1938 after a two year stint with the Bureau of Reclamation in Denver (where he did some of the preliminary work on the Grand Coulee Pump Plant layout) and subsequent graduate studies (for 2 years) in Civil Engineering at Stanford. I had known Brooks casually at Stanford although he was a year ahead of me.
Brooks moved to the Caltech Soil Conservation Laboratory in mid 1939 where he stayed until late 1942. In 1943 he joined Aerojet (arranged by Kármán). There he transferred his talents from CE to ME applications, becoming an expert on non-rotary jet engines of the pulse jet and ram jet types. After some years Brooks went to JPL where he remained until he retired recently.
Joe Levy was a knowledgeable and capable engineer. He was born in Jerusalem of American parents, came to the U.S. in the late 20’s and did his undergraduate studies in ME at MIT and the University of California (Berkeley). Joe came to CIT in the spring of 1939 after graduate study in ME at Berkeley. In 1940 after the USBR program was over Joe went to Portland, Oregon, with the U.S. Corps of Engineers until 1943 when he returned to CIT to engage in research on underwater ballistics problems for the next 13 years. Finally, at Cal Gongwer’s urging, he joined the Aerojet staff in Azusa where he remained until his retirement. He died untimely in 1972.
Ed Simmons was a quiet "loner" and interesting person who contributed much to the pump lab over its MWD and USBR days. He was ingenious and capable. An EE with even broader capabilities, Ed produced not only exotic electronic instruments but applied his craftsmanship to develop and build other devices as well. We could always rely on Ed to produce a working answer to our urgent need of the moment. It might come at the last minute (like at the stroke of midnight on the due date) but he always came through. Ed is still active professionally and still inventive. He appears often at Caltech seminars and other events (sometimes in Elizabethan costume).
Incidentally, Ed is one of the inventors of the S-R "strain gauge." He is the "S." The "R" is for Ruge of MIT who was developing and using strain gauges in the 1930’s at the same time as Ed’s work at Caltech.
Every laboratory depends heavily on its secretary and its mechanic. The two key people for the USBR program were our secretary, Lillian Price, and our chief mechanic, John Bigelow.
The lab staff worked hard and played hard. There was a real social life beyond the work place. There were dinners and picnics of families and singles (with or without dates). Now and then a wedding brought everyone together. There were lab Christmas parties also, of course. The Christmas parties were well attended with the usual awards of a few humorous gifts, carol singing, and so on. There were some limitations, at least for any on-campus party that had a large attendance. It was still in the days when alcoholic beverages were not appreciated for student gatherings. However, there were the usual exceptions and the Christmas punch did get spiked occasionally.
The pump lab’s social gatherings had the cooperation of everyone. Don Barnes and his wife, Thedia, were generous with their hospitality. Kármán and his sister, Pipo (pronounced Pipa in Americanese) were gracious hosts and their door was always open to students and young friends. Professor Daugherty and his wife, Marguerite, liked to entertain student groups. Professor Knapp and his wife, Pearl, entertained copiously and always had their then newly-built house and patio available for pump lab functions. (Incidentally, this hosting of student groups was responsible for many of us, especially those who entered academia, to actively entertain students and colleagues in our later careers.)
At the pump lab parties it was customary to have someone, often one of the "bosses," offer a program or speak informally. On one occasion that I remember well Kármán was asked to speak. He was terribly busy so he begged off preparing something even though he accepted the invitation to attend. I guess the committee misunderstood so when the time came he was "introduced" as the speaker. Kármán was chagrined but accepted gracefully and proceeded to ad lib. For the next 10 or 15 minutes he regaled us with stories of his student days and particularly about the Saturday night wine binges and the inebriated and peace-disturbing students who found themselves in the city jail for the weekend.
A big event that evolved into a social "happening" was a gathering to "soundproof" the "office." Our office space was adjacent to the lab where the sound level sometimes reached to 90 or 100 decibels. Only a large folding door separated this office from the noise, so plans developed to rectify the problem by covering the door with a "gyptile" layer on the lab side. The job went slowly so on an appointed warm October evening a large portion of the lab staff appeared and did the job, complete with a buried time capsule with information about the event (in case posterity had any interest). If this gyp-tile covering has not been removed the time capsule is still there.
Of course, all this was without the proper approval of buildings and grounds, and I soon learned that we had overstepped our authority. However, the deed was done and accepted. Sometime later I learned from Hertenstein that he had passed by the building the evening of our handiwork and saw the job in progress, but he was sympathetic and decided to overlook the formalities. (Incidentally, the soundproofing worked!)
I should add that the lab was opposite R.A. Millikan’s office across the alley in Throop Hall and the noise upset the secretaries. However, RAM called it the "sounds of progress" and let it go on.
The USBR program officially ended on June 30, 1940. The staff scattered to other pursuits. I looked forward to full time graduate studies, but this "elysian dream" didn’t last long. Instead I became one of Caltech’s instructors and was busier than ever teaching in special programs and to undergraduates. My own graduate work was delayed again and it was another five years before a diploma for a Ph.D. became a reality.
By 1941 it was apparent that the U.S. probably would be swept into the European war, then raging, and/or the fighting in the Orient. Kármán, by this time, was devoting more and more time to research and development of high speed aircraft. Meanwhile Knapp was spending an increasing amount of time on the East Coast as a consultant on underwater ballistic problems. In this period Knapp took part in the "later-to-become-famous" launching of model projectiles into the MIT swimming pool.
These early model launching experiments were sponsored by the National Defense Research Committee (NDRC). The NDRC had been organized in 1940 and by 1941 much new work was started. In late 1941 (I think it was November) Caltech entered into a contract with Division 6 of NDRC for an underwater research program. The contract provided for the construction of a high speed water tunnel and Knapp became the director. The water tunnel was built in the pump lab using the dynamometer as its drive, the large standing tanks as part of its flow and pressure control circuits, and any of the pump lab’s instruments.
The water tunnel’s birth ended the laboratory as a pump test facility for the duration of the war.
Later, during the war, Knapp promoted a separate building to house several new and existing (but scattered) hydrodynamic facilities, with space reserved for later relocation of the water tunnel from the pump lab. This new building (behind GALCIT) was named the Hydrodynamics Laboratory.
In 1957 Bob Knapp died unexpectedly. Eventually the Hydrodynamics Laboratory’s supervision was transferred to the Aeronautics group, a second story was added and the building was renamed the von Kármán Laboratory.
The pump lab was revived in mid-1946, this time for a program sponsored by Byron Jackson Co. B.J. had teamed up with Pelton Water Wheel Co. of San Francisco to win the post-war contract for the Grand Coulee pumps because Pelton had the shop facilities (as B.J. did not) for building such large units. B.J. wanted some more development tests. I participated in this series of studies until September, 1946, when I left Caltech to take an appointment on the MIT faculty.
Reminiscences of Days at the Caltech "Pump Lab"
Some explanation seems necessary as to why I set down these reminiscences and bits of history of the Caltech Hydraulic Machinery Laboratory. The truth is I am nostalgic about my exciting time as a graduate student and my experiences in the "pump lab."
When I arrived in Pasadena the Caltech as we have come to know it was quite new, only in its second decade under R.A. Millikan. The Institute was known already for doing the exceptional and was busy expanding its areas of expertise and influence. The pump lab’s research was part of this expansion. It injected a scientific attitude and more science and mathematics into a "down-to-earth" problem area. In the short period of less than ten years the pump lab developed a world-wide reputation for excellence and thus added to Caltech’s luster.
It all was an experience not to be forgotten.
James W. Daily
Last updated 3/9/07.
Christopher E. Brennen