AVS Historical Persons | John L. Farrant - 1993

John L. Farrant - 1993

Oral History Interview with John L. Farrant

Interviewed By Bruce Kendall, November 18, 1993
KENDALL: I am here in Orlando at the 40th Symposium of the AVS with John Farrant1 to discuss the development of the Oil-free Mechanical Vacuum Pump. John, can you tell us how this came about.

FARRANT: In 1945, CSIRO2 imported the first electron microscope in Australia from RCA in the USA. In fact, it was the second one they built. Until the war, electron microscopes were made in Germany and Belgium. RCA had asked Professor Burton at the University of Toronto for students who had experience on microscopes and James Hillier went to RCA. The Model A, designed by Martin, was too researchy. RCA had excellent electronics and designed feedback circuitry for stability. The Model B was the most successful unit so far and Australia got an even better designed unit. I was appointed to introduce electron microscopy to CSIRO which then had 30 divisions covering all areas of science except medicine. There were about 7000 employees, including 2000 PhDs. 

The microscope was in the unit to which I was appointed, and that was one of the sections in those days of the Division of Industrial Chemistry. The overall Laboratory of Industrial Chemistry was directed by a man named Walk3, and it was on his initiative that a microscope was purchased. He appointed a man named Lloyd Rees, who was in England when he was appointed in '44. He was the head of the section of Chemical Physics. That consisted of subsections of which one was Electron Microscopy, to which I was appointed. Another was Electron Diffraction, to which John Cawley was appointed., For some years, he has been professor of physics of what Australians boast is the leading school of electron microscopy in the United States, at the University of Arizona. It also had subsections which dealt with spectroscopy. There was Light Spectroscopy, Mass Spectroscopy, and Theoretical Chemistry. It was directed by this very ambitious young man, Lloyd Rees. Well, Rees was determined to make his laboratory one of the most notable. Well, we would have been content if it was the most notable in Australia, but he wanted to do rather better than that. And in those days, most of the universities were rather remiss as far as research was concerned. They were largely undergraduate institutions. University of Sydney and Melbourne were big enough to have small research schools, but the smaller universities were lucky if they had one or two research students in physics or chemistry in any particular year. 

Well, because of the problems of the contamination of the specimen, I pondered the mystery of how this contamination was being formed. But in 1949, CSIRO sent me on a year-long mission. I had a roving commission to go anywhere in the world where I could learn about electron microscopes and the possible applications of electron microscopes. So, naturally, the first place I went to was the USA, and then off to Princeton. RCA's laboratory is in Princeton. James Hillier and his colleagues, who had designed the EMB and the EMU, were still hard at work trying to perfect the microscope. Jim Hillier, in view of the fact that the German school had been dissipated by the war, was clearly the leading man in the field anywhere. Hillier had investigated the sources of this contamination, and one of the things he pointed out to me was the evidence that he had to show was that the oil in the backing pump was a major source of the contamination, because the backing pump was used to rough the microscope and to dry out the photographic plates, so it ran quite some time before one switched on the electron beam. And during that interval, oil from the backing pump contaminated the low-vapor-pressure oil in the diffusion pump, which promptly boiled that low-vapor fraction up into the column of the microscope, so all the surfaces were covered in a thin layer. No matter what people did, it seemed to make very little difference to the rate of contamination.

After talking to Hillier, it seemed to me that absolutely oil-free pumps were needed. Nobody else was able to able to conceive a method of producing an oil-free pump. I got nowhere with the project, but several years later, I did notice that we had some large glass hypodermic syringes. Both the barrel and the plunger were glass, and they were ground to a rather precise fit. I suppose the ones we had were about an inch in diameter. We got them from one of the animal divisions. RCA is all over. One could put one's thumb over the end, having pushed the plunger down to the far. And then if one withdrew the plunger halfway down the barrel and held it for several minutes, and then slowly lowered it back, one found that a very small quantity of air had leaked through this long, narrow gap. It was important to lower it slower because if you let the plunger go, it went straight through the far end of the syringe and destroyed the barrel. Anyhow, it was clear that a glass piston operating in a glass cylinder was scarcely an appropriate way to develop an oil-free pump because particles of gritty dust would soon destroy the surface. 

It was not until 1968 that I heard that ways had been found by greatly increasing the wear resistance of Teflon, the material that is used in cooking. Well, they use it to line fry pans. It's a very inert substance, and has a very low coefficient of friction around about…Well, the coefficient of friction is of the order of .1, whereas all other substances in a dry condition exhibit much higher coefficients of friction. But it seemed impossible to construct a pump using a Teflon coating on the piston because Teflon is so soft. But a method had been found of increasing the wear resistance of the order of 1,000-fold by incorporating into it graphite, glass fiber, molybdenum disulfide, various carbon blacks, even bronze powder. Quite a number of substances we used.

Well, in Chemical Physics, we had lived for a long time on the reputation gained by the work carried out by a man named Alan Walsh, who invented atomic absorption analysis in the division. That led to the development of a small manufacturing company named Techtron, which produced the spectrophotometers and the special labs used to exploit Walsh's invention. It soon built up into a world supplier of this equipment. But the owner found that he didn't have the resources to provide himself with marketing and service and sales facilities throughout the world, so he sold the controlling interest to Varian. Varian still operates what it called Varian Techtron in Melbourne, and today it employs about 450 people, selling about $2 million worth of equipment every week, most of which is exported. In recent years, it has even transferred some of the spectrophotometers it used to produce in California to Melbourne. Believe me, very few overseas companies operating in Australia transfer something that they made formerly in their home countries to Australia, but Varian has done so. Well, this made a great impact on the controlling body, which ordered the affairs of CSIRO, and we lived on it. We got quite large grants to pursue our work wherever we chose. CSIRO had a very broad charter. It could spend money on virtually anything to do with science and industry and its promotion in Australia, and to assist scientific education, that sort of thing.

By the late ‘60s, when I heard of this filled Teflon, as they were called, the impact of atomic absorption analysis was starting to wear off. Walsh had conceived the idea as early as '53. I could see that Chemical Physics was going to need another commercial success. Little did I know how long it takes to create a commercial success! Anyhow, I was very fortunate in that as a member of my group, I had a young man named Eckhard Bez, who had been brought up in the German tradition, having been of German extraction, and he'd been brought up in the "School of Thorough"; in the old German tradition, if a thing was worth doing, it was worth doing properly. He was quite a bright young fellow. As a matter of fact, he was apprenticed in our workshop as an instrument maker, and he won the Apprentice of the Year competition in Australia, which enabled us to send him for two years to various friends of CSIRO, such as Perkin Elmer and Varian in the United States, and to Siemens in Europe. So he had a wealth of experience. As soon he returned to Australia, I set to work to extract him from the workshop into my group. We soon found that with the sort of background my colleagues and I were able to provide him, he became more valuable, I used to say, than many of the Ph.D. scientists we had. 

In 1970, I said to him, "Well, I'd heard about this low-wear Teflon, and I had this demonstration of how low the leakage was in glass hypodermic syringes." I said, "Let's see if we can pump down a mercury barometer just by hand with this four-inch stroke hypodermic syringe, using as valves the bending of a rubber tube connecting the barometer with the hypodermic." And we found that in a few minutes, we could get down to below one centimeter." Well, Eckhard Bez had a little boy for whom he had made an oscillating steam engine. I knew this because I had seen him producing it in the workshop in his spare time. So I said, "Bring in that oscillating steam engine." So, running the oscillating steam engine piston-cylinder dry as a backing pump, we found we could achieve pressures of about three centimeters of mercury even though it only had about three-quarters of an inch bore and stroke.

KENDALL: But it was probably made to very high tolerance, wasn't it?

FARRANT: Oh, yes. So, I then drew a sketch. Now, in my childhood, I was one of those children who had a father who liked steam trains, so I had no great trouble in persuading my father, when I was nine years old, to import from England a Bassett-Lowke steam train. At the age of nine, I had a number of books on steam engines. Steam engines were quite interesting to me, so it's not unexpected, perhaps, when I say that I drew a sketch rather like a double-acting steam engine piston and cylinder with a four-inch bore, eight-inch stroke, and a two-inch-long piston. The induction port was four inches from one end, so that when the piston uncovered the port, that opened the port to the vacuum vessel. There was an exhaust valve in that end of the cylinder, in the cylinder head, and another exhaust valve in the crank case end. In addition, in order to start against atmospheric pressure over a diameter of four inches, I had pressure-operated induction valves in series with the ports. Well, Bez built one of these with a four-inch stroke, and rather than spend time making a mechanical system to oscillate this piston, we connected it up on a bench with a lever.

We called Rees, the chief of the division, for an exhibition of what we could do, and I was foolish enough to volunteer to operate the lever. Well, I was completely red in the face and out of breath by the time I achieved a millimeter or so, but we managed. After that, Rees's cooperation was such that it was embarrassing. Some days, he would come and visit us three times in a day to see what progress we made because he was determined to make Australian science much more productive than it had been previously. Having had the great success of atomic absorption, he felt it was time we had another industrial success. By 1974, we were able to publish our first little paper that was on an opposed cylinder model. The piston rod with a gland on it off my first model had been changed as it had been evolved into a T-shaped piston with a large head about four inches in diameter, and a smaller waist about two inches in diameter, and a stroke of about 19 millimeters. With that, we were able to achieve pressures all up - that's including the water vapor - of about 80 microns of mercury. That year, 1974, I had brought the International Congress. These congresses are held every four years in electron microscopy, and I had brought that one to Australia. I ran the Congress, the eighth of the series, in Canberra in '74. We exhibited this model to those physicists who came to Melbourne, and I put a small publication into the conference proceedings. That had the advantage that we didn't have to say very much about the technical details because it seemed to me to be likely to turn up quite a valuable invention. Little did I know.

By 1977, we looked for a licensee. Seeing that our duty was to cultivate the development of Australian industry, and there being no vacuum company of any real accomplishment in Australia, we were trying to establish a vacuum industry in Australia. The license was offered to a company called Repco, which was the biggest Australian-owned manufacturing company at that time. There were bigger companies like General Motor's Australian offshoot and Ford's Australian offshoot.

KENDALL: But Repco is known here in the United States as a manufacturer of brake pads and brake lights.

FARRANT: Yes, indeed. "Repco" is a contraction of "Replacement Parts Company," and it makes replacement parts in the automobile industry. It made O-rings - even a Viton, which we needed. It made brake cylinders. It had a concession to make cylinders of both cast iron and of aluminum for Volkswagens. We soon found that rather than turn out cylinders, especially a machine with fins for cooling, it was far better to use Repco Volkswagen cylinders. This resulted in some of our early pumps being sold by my two colleagues because I not only-- Once we had achieved a reasonable vacuum, I found that I had no difficulty extracting another man from the workshop who was also of German extraction and had served his apprenticeship in the old school in Germany before the war. His name was Balkau. Well, I couldn't have done any better than have those two working on the project, because one of the main difficulties we encountered was coping with the very high thermal expansion of Teflon because, well, the pump was cold. One didn't want too large a leakage past the piston rings, so one made them a reasonably close fit. But in a few minutes, as they warmed up, the fit became too good, and we'd often run a pump for an hour, and it'd still be okay. Two hours, oh, we'd be even more optimistic. But in another ten minutes or so, it would rapidly rise in temperature and then cease. That must have happened hundreds of times. 

Balkau and Bez must have made at least 200 different styles of piston ring and seal. The amusing thing is that ultimately, we found that only the seal facing the atmospheric pressure in the crankcase wore slowly. All the others operating in the vacuum in these cylinders wore much more rapidly because of the drier conditions. There was no water film to reduce the friction. So, we ultimately reverted to the sleeves that we had used first of all, and we had one…well, I think they were known as cup seal, on the small diameter of the piston, facing the atmosphere. Both diameters of the piston were then covered with a thin layer of Teflon only half a millimeter thick, and this solved our problem because we could make that layer only half a millimeter thick. So even though its thermal expansion rate was high, it wasn't any great thickness to expand and seize.

We had numerous critics in the division and elsewhere. They all maintained that it was crazy to try to compete with an oil-lubricated machine with a dry machine. And they all predicted that they'd run very hot, that they'd have a very short life, but it turned out they were all wrong. The pumps that are on the market today, they're devoid of any lubrication in the cylinders; run cooler than the traditional, vane-type mechanical backing pumps, which operate with oil lubrication; and the oil not only lubricates, but it also conducts away the heat generated in the cylinders from the piston to the walls of the cylinders where, of course, it can be easily dissipated. Well, we had Repco working on the pump under license to CSIRO, but Repco really didn't have its heart in the work because they weren't really a vacuum company.

It wasn't until I went overseas in '78 and visited Varian, Perkin Elmer, Leybold, and Balzers, trying to seek information from them concerning the marketability of an oil-free pump. I found general skepticism everywhere. I even went to Japan to electron microscope companies. There was general skepticism everywhere. Very few people thought a dry pump had any real future. Except at Leybold, where the director of research, a man named Hansen Pfaff, who was also the second man on the board. After I talked to him in Cologne for about an hour and a half, to my utter surprise, he said, "I think I ought to go to Australia and see this for myself." And he was as good as his word. I went to Australia, I said nothing to anybody, and then I had a ring from Germany saying Pfaff would be arriving next week. So I went to the acting chief of the division, Rees having retired. His name is Matheson4, and he's a Scot. He still retains his original Scottish accent. He was also my closest friend. But I can still hear him say, "Leybold. You cannot do much better than that!" [Laughs]

Well, Hansen Pfaff turned up, and he said, "I've come to Australia to seek a license to manufacture these pumps." I could scarcely believe my ears. Well, we explained to him that Repco was licensed because CSIRO's obligation was to try to develop industry in Australia. He visited Repco and said to them, "Yes, you can produce these pumps. I see you have the ability. But you won't be able to sell them because you're unknown in the vacuum world, whereas I have agencies, sales, service facilities in 43 countries. I'll offer you a deal. In return for an exclusive marketing license, I'll give you access to my technical ability and to my marketing skills." Well, of course, Repco accepted. And Pfaff immediately said, "Send me pumps by Christmas." This is in October '78. Well, Repco demurred, but I persuaded them to send one of their engineers and one of their pumps and one of our pumps, and off we went to Germany in early February 1979. They conducted extensive tests of every aspect of the pump. They then set out what they called their protocol, saying, yes, it was very encouraging. They thought it should be developed to a commercial degree by overcoming certain difficulties and improving the pumping speed at low pressures. They also provided us with a lot of market surveys they had made. But unfortunately, Repco was one of those companies run by accountants, and they weren't over-impressed by the potential markets, which were largely seen by Leybold as being in the instrument field. By the end of '79, Repco decided it would do better to concentrate on its car part industry rather than dabble in things which were really somewhat outside of its experience.

Matheson, my friend, was replaced as chief by a man named Chellerton5 from England, and he had very little faith in this pump. But I had sufficient influence in CSIRO, having been there since 1945, to be able to keep the project going. And by the middle of 1980, I said to him, "Well, I can reach Leybold's specifications for a commercially useful pump. I'm going to take pumps to Leybold for a second assessment." I took Bez with me at my own expense, but anyhow, off we went to Cologne. Repco and Leybold were very encouraging, but unfortunately, they were having a difficult time financially; had to cut some of their own projects. And although they were enthusiastic for the development to continue and still wanted the marketing license if we were able to produce them, it was not able at that time to continue to help. So, to encourage Hansen Pfaff to make a greater effort, I said, "Well, CSIRO will advertise them again. And I'm sure that this time, Varian Techtron, who have large works and laboratory only a mile and a half from our division, one which grew out of our division, I'm sure they'll apply." And to my astonishment, instead of producing the effect I wanted, Pfaff's reaction was to say, "I have to admit, Varian would be a very good company to develop these pumps." So I ran to Chellerton and said, "Well, I'm going to Varian again to see if I could persuade them."

In January 1981, I arrived at the Varian division in Lexington, Massachusetts, where they conducted most of their pump-making and vacuum activities. There, Marsbed Hablanian, who is well known in the AVS, had said to me, "There's a meeting of Varian people next weekend. I'll do what I can, and I'll let you know what they think." Well, I went off to see my brother, who settled in Canada, and to my delight, when I rang Hablanian a few days later, he had persuaded them to ask me to go to their works in Santa Clara. They called that the Industrial Products Division. That was run by a man named John Vossen, who answered to a vice-president named Richard Scholl.

KENDALL: By this time, of course, Varian had made a very large amount of money out of the atomic absorption, hadn't it?

FARRANT: Oh, yes. Yes.

KENDALL: So it's probably ready for another…

FARRANT: Yes. Varian Techtron was one of their most profitable divisions.

KENDALL: So they're ready for a repeat of that process, perhaps.

FARRANT: Well, that's what we thought, anyhow. [Laughter] And we hoped. Well, I went to them in mid-January, and after I talked to John Vossen and to the marketing manager, an Italian named Alessandrini they had imported from the factory in Torino, I was asked to arrange with CSIRO to send pumps for testing and assessment by Repco engineers. So I went back to Australia and did that. The following January (that's January of 1982), I took two of our pumps to Varian in Santa Clara, the Industrial Products Division, and there we ran those pumps for…well, the agreement was to run them for three months, and then they'd make up their minds. Well, in two months, I had persuaded them to apply for a manufacturing license from CSIRO, and that agreement was signed early in '83. They asked me and my right-hand man, Eckhard Bez, to go and talk to their consultants together with some of their staff people, like Marsbed Hablanian, at their consultants' establishment up in Albany, New York.

The name of the company was Mechanical Technology Incorporated, which Varian saw as having the advantage of not being involved in compressors or vacuum pumps, so they felt safe to have them learn all the secrets of these dry pumps because they were sure that they wouldn't find that they'd lost the project to their consultants. Although they're very competent engineers, they had no experience whatsoever in vacuum. Bez and I found that we had to teach their engineers the elements of vacuum physics. But they designed the first pump, which Varian produced. Now, this was a four-cylinder model, a VAT-4. Two pairs of cylinders on either side of the axle, on which there were large, eccentric bearings mounted, and four of these T-shaped pistons. Two cylinders were running parallel as the high vacuum side, the next cylinder was the intermediate, and the last cylinder, the backing cylinder. Well, the pump they produced looked somewhat like an engine - much more like an engine than a vacuum pump. I returned to Australia, and I must have written them many reports recommending that they get rid of the external conduits they had encumbered the whole thing with. We had gone back to Australia, and we had built two short-stroke pumps, such as they wanted. They wanted the pumps to run at full motor speed because the oil-sealed pump people had mostly introduced models which dispensed with the belts, which had become regarded as old-fashioned.

KENDALL: So this would be about 1800 RPM or so.

FARRANT: Yes. So, they wanted to run at 1800, but in order to get down to a micron pressure, which was always our aim, we had to increase the stroke to 45 millimeters, which is near enough to 1.8 inches. In order to meet the specification they wanted of full motor speed in North America - full motor speed in Australia is somewhat slower in that we operate on 50 cycles; that meant about 1400 revolutions a minute, whereas here, it's a matter of about 1730 - so we had to reduce the stroke, and I warned them that this would result in a poorer ultimate pressure. But they wanted to dispense with any belt, so that was the compromise.

Varian needed some help in putting these pumps together, so I suggested to Richard Scholl that they ask CSIRO to lend them the services of Eckhard Bez. He was sent over to Varian and to their Vacuum Products Division, where the project had been transferred, in Lexington, and he supervised the assembly of the first model, which wasn't that satisfactory. Anyhow, they realized that they would have been wiser to listen more to us with our experience than to their consultants, but they had exhibited the NIH syndrome - the "not invented here" syndrome. So they didn't believe these two young Australians, or one young Australian and one aging Australian, and were inclined to take their consultants' word above ours. However, by that stage, they realized it would have been better. So they gave Bez the job of redesigning various aspects of the one designed by Mechanical Technology, and they incorporated the features we'd built into the two 1730-revolution models we built of short-stroke pump we had built in Australia in the interim.

Varian decided to announce in 1986 at the International Vacuum Congress, which was held that year in Baltimore. They didn't exhibit the pumps because they hadn't accumulated enough to sell. They had them tested by various friends of Varian, and they were finding them satisfactory. So, the first models were actually sold in '87. If you remember, I mentioned that their initial interest had developed in Santa Clara, in the heart of the semiconductor industry. The market surveys had been conducted amongst the customers in the semiconductor industry, so they had the semiconductor industry in their minds as the potential market. The initial pump had a nominal speed of about 500 liters a minute, and actual speed of about 430 liters a minute. So a fairly big pump. 

They sold slowly at first, but numerous other industries found they had advantages. Lawrence Livermore, for example, bought quite a large number, much to our delight, because one of the offshoots of atomic absorption in chemical physics was that it had occurred to Walsh and to Rees that the same processes which were used in atomic absorption could be used to separate isotopes. That is, if you wanted to extract uranium-235 from uranium-238 (most ordinary uranium is 238, with less than 1% of 235), if you were able to illuminate a vapor of uranium hexafluoride with the appropriate wavelength to ionize uranium-235 hexafluoride molecules and not the 238, you could then extract the 235 and achieve the separation. At the Lawrence Livermore, they had set up a demonstration of this isotope separate method, and they had plants in which the lasers, which supplied the monochromatic radiation that was needed to excite the uranium-235, were rare gas lasers. When the gas was circulated by oil-sealed pumps, the oil vapor resulted in this same contaminating layer depositing on the windows and absorbing a good deal of the light. By changing to our oil-free pumps, well, these plants were quite large, and each plant was ultimately fitted with 12 of these pumps being made by Varian under the name of DVP, the dry vacuum pump. By fitting 12 to one plant, they were able to increase the value of the isotope turned out by a million dollars at the expense of $100,000.

KENDALL: Which would pay for quite a few pumps.

FARRANT: Yes. So they were one of our best customers. That laboratory must have bought something in the vicinity of 100 altogether. People producing channel plates and all sorts of things realized that dry pumps had their advantages - much to our delight, because those critics in Chemical Physics were very much confounded. And even people eminent in Varian! Perhaps I may be so bold as to say that one of the leading men in Varian, he was an engineer by training, and he declared he couldn't see that it would be possible for any dry mechanical device to compete with an oil-lubricated one. But in the end, he had to see for himself, and he realized that as we said, dry pumps ran cooler than traditional oil-lubricated pumps.

KENDALL: Down there in the exhibit hall today, we saw a large number of different models of these pumps from different manufacturers. How many manufacturers are now selling them?

FARRANT: Well, Varian ultimately divested itself of most of its pump-manufacturing activities, and it sold their license from CSIRO to a small Pittsburgh company called Vacuum Research Corporation. And they began by producing the two models, which Varian had designed - that is, the DVP-500 and the DVP-1200. But recently, Mr. John Hartnett, the owner of the company, has introduced new models that amount to two-cylinder pumps, which he is running rather more slowly, so they run more quietly because the forces that cause vibration and so on, depend upon the square of the angular velocity of the shaft. By running at 750 revolutions rather than 1200 revolutions per minute, as Varian ultimately did, these new models run very quietly, and they are of a size suitable for many a scientific instrument.

Up until the present time, what has been lacking has been small models suitable for the scientific instrument industry. They need pumps of the order of 100, 200, 250 liters a minute rather than the 400 and 900 that Varian had initially produced. So, we hope that these pumps, which were exhibited for the first time at this conference in Orlando-- They seem to excite quite a lot of interest in the fact that they're so quiet and cool presages quite a future for them, I feel. In addition, a Japanese company that has been involved with Mr. Hartnett's company (that is, with Vacuum Research Corporation) has been able because Mr. Hartnett changed from an exclusive license to a nonexclusive license. They were able to get a license from CSIRO to produce a new version of the pump in which, instead of just having one step in the piston so that there are just two working spaces, they are now producing a pump with three working spaces in that there are two steps in the piston. That pump is only of recent origin. It was designed by Mr. Bez, who is linked by CSIRO in succession to Varian, then to Mr. Harnett's company, Vacuum Research Corporation, and now to the Japanese company, Fuji Seiki,, which is a Yokohama factory run by a very enterprising man named Mr. Taniguchi. They are aiming to invade the oil-sealed pump's territory in the instrument area. So these pumps have two versions: a single-cylinder pump - and of course, a single-cylinder pump cannot, without elaborate arrangements, be perfectly balanced, but there is another version of it, which has two cylinders opposed, and they can be run either in series, in which case they get down to about seven microns of mercury even though they are running with a short stroke, or in parallel, where they're up around 30 microns. But they have twice the pumping speed, of course.

KENDALL: Looking a little further ahead, what do you see as the future of this pump?

FARRANT: Well, there have been quite a number of other, shall we say, so-called dry pumps developed, and these are in the form of Roots blowers, which have two globes geared together so that they each run at the same speed - they're like a two-toothed gear wheel, engaging one another in a housing, or there are three-globed versions, and they are being sold as dry pumps. Another derivative known as the claw pump, where the rotors are modified so one has a projection which is in the shape of a claw, which descends into a cavity of corresponding shape in the other. But these two require to be run in synchronism to a very high degree because the tolerances are small, which means that they have to be geared together. And those gear wheels have to run in oil. So there's quite a reservoir of oil. If anything happens to the shaft seal, the oil is in the high vacuum chamber. In addition, provided the shaft seals don't leak, they have not much oil in the working chamber. We have no oil whatsoever in the working chambers, but they must have at least a narrow annulus on each shaft, each side of the rotor, where there's a shaft seal, which will only run provided the heat generated in the lip of the lip seal is conducted to the shaft by a film of oil. So there's a narrow annulus of oil, which, of course, involved vapor. So even though you start off with a fluorinated oil, such as Fomblin, very low vapor pressure, gradually some of the Fomblin decomposes, liberating low molecular weight vapor straight into the high vacuum. And we have none whatsoever. Well, of course, whereas an oil-sealed pump requires traps to trap the oil vapor, many of these operated at liquid nitrogen temperatures. Liquid nitrogen doesn't come cheaply, even in the United States. And furthermore, oil is a hazard.

KENDALL: Well, thank you for being with us tonight. This has been an interview with John L. Farrant, a pioneer developer of the modern dry piston pump.

1. For more information, see
'Farrant, John Lascelles', in Physics in Australia to 1945, R.W. Home, with the assistance of Paula J. Needham, Australian Science Archives Project, June 1995, http://www.asap.unimelb.edu.au/bsparcs/physics/P001651p.htm. [ Details ]
2. Commonwealth Scientific and Industrial Research Organization.
3. Name uncertain
4. Spelling uncertain
5. Name uncertain

return to top