Anthony R. Kerr, Interviewed by Kenneth I. Kellermann, 3 December 2025

Bouton: 00:03

And I started to record on the other.

Kellermann: 00:06

Okay. So Tony, thank you for doing this. It's December 3rd, Wednesday, and we're--

Bouton: 00:14

2025.

Kellermann: 00:15

Thank you. I often forget that. It's Ken Kellermann and Ellen Bouton, and we're talking with Tony Kerr, a longtime engineer at NRAO, and now a retired emeritus scientist. So, Tony, let's start back in the beginning, just in the way of background. Your early life, where you were born, where you grew up, where you were educated.

Kerr: 00:44

Okay. I'll--

Bouton: 00:46

What does your father do that you spent part of--?

Kerr: 00:49

Hold on. Let me share my screen. It says I'm screen sharing.

Kellermann: 00:58

Yes. We see it.

Kerr: 00:58

You see it?

Kellermann: 00:59

Yeah.

Kerr: 01:00

Okay. You don't need to see my email. Okay. Is that reasonably clear?

Kellermann: 01:05

Yeah.

Bouton: 01:07

Now we see your companion.

Kerr: 01:10

Yes. So this was Guinness who, as I noted there, he used to visit the CDL in the 1970s until Dick Thompson banned dogs as being unclean.

Kellermann: 01:24

What was Dick's role? Because I don't recall him ever being a CDL director or--

Kerr: 01:30

No, he wasn't director. He was in the CDL for a while. I can't remember exactly the dates, but I used to work at weekends sometimes, and I would take one or other of my dogs in until Dick banned them. And yeah, I don't remember exactly. Well, it was certainly before 1974 because that was the year I went to New York. Okay. So next, well, those are your questions [see accompanying PowerPoint]. And then, okay. You're asking what my father did. My father was an Australian, went to England on a Rhodes scholarship and stayed in England until the war. He was working at the Royal Aircraft Establishment in Farnborough. And when the war started, he was an aeronautical engineer. They sent him to Washington during the war as a scientific liaison person. So we joined him in Washington, I think, in 1944, my sister and my mother and myself. And we stayed in Washington until 1949. So I had my first couple of years of school in Washington, which was great.

Kerr: 03:10

And then we returned to Farnborough for a few years until 1953. My father hadn't seen any of his Australian family for many years. And England was pretty depressed after the war. Food rationing was still in. So we decided to go to Australia. And he got a job at the-- first of all, at the-- I can't remember the name, the organization that did missile research in Australia, the Woomera rocket range, but he was based in Melbourne. They were called Weapons Research Establishment. But then he changed to the Aeronautical Research Lab, which was also in Melbourne. He was there for the rest of his professional life.

Kerr: 04:11

I went to Melbourne University for 9 or 10 years. After my first degree, I worked as what was then called a senior demonstrator, which was basically running lab stuff with students and low-level teaching. And while I was there, I had the good luck to go on a course at CSIRO in Sydney, which was organized in those days, I think, by Taffy Bowen, Dick McGee, and people like that. That's where I first made contact with radio astronomy. I think I had a master's degree at that point and was working on my PhD. So I spent a month or so there in the summer. They gave us a pretty good introduction to all the different aspects of radio astronomy and radio astronomy instrumentation, which was the most interesting part to me.

Kerr: 05:36

Brian Robinson was there. He'd done a lot of work on receivers originally in Holland and then back in Australia. My immediate boss was Brian Cooper, who was also here at NRAO once upon a time. Brian and Brian recommended or suggested that I might want to come to NRAO or Bell Labs, where there was some cutting-edge instrumentation work going on. So in 1969, I did that. Bell Labs wasn't interested in having me, and Sandy Weinreb thought it was a reasonable thing because he could pay me a very low wage. He admitted after the first year that I was underpaid, so they cranked me up to a normal pay level. I'm being far too verbose.

Kerr: 06:59

So anyway, in '74, I went to NASA's Goddard Institute for Space Studies, which was at Columbia University, and worked for Pat Thaddeus. At that point, I got married to an aspiring musician, and she needed somewhere that she could study at the same time as I worked. We ended up going to Thaddeus' group at the Goddard Institute. And that turned out to be a really good decision for both of us. Let's see. When I was at Radiophysics from 1969 to 1971, the first Gallium Arsenide Field Effect Transistors were just appearing at an experimental level. Plessey in the UK had developed them. And also, there was some talk at Berkeley about using Josephson junctions as mixers, not as SIS mixers as we use them now, but using the superconducting pair (Josephson) characteristics that SIS mixers suppress with a magnetic field.

Kellermann: 08:25

Tony?

Kerr: 08:26

Yeah.

Kellermann: 08:27

Plessey, is that a person or a company, or?

Kerr: 08:29

Plessey is a big-- or it used to be, I think it probably still is, a big UK electronics company. Yeah, they've probably become-- changed their name or being taken over by something else now. Yeah, they were a big deal in UK electronics in those days. Let's see. You want me to go into details of question two, what inspired me to come to NRAO, and why?

Kellermann: 09:07

Well, you've explained that a little bit.

Kerr: 09:10

I think I've probably done much of that. Yeah.

Kellermann: 09:12

But then you decided to come back.

Kerr: 09:15

Yeah. After 10 years in New York-- let's see if there's anything important I've missed here. Yeah, I guess this bottom paragraph, at the time when I was first at NRAO in 1971, Jochen Edrich, German guy, was making parametric amplifiers for the Green Bank, 140 foot telescope. And Sandy had this collaboration with Bob Mattauch at UVA to make gallium arsenide Schottky diodes for parametric amplifiers. And that was using a method-- well, Mattauch used a method that was originally developed by Charlie Burrus at Bell Labs. And the UVA Schottky diode work for parametric amplifiers, directly led into the cooled mixer work we later did. And the UVA lab has continued, from then until the present day, being highly involved in our receiver development work. So let's see. Okay. Any other questions on that, or should we go on to question three [inaudible] the priorities of NRAO changing over time?

Kellermann: 10:44

So that was the reason you came back because of the opportunities to continue the work you were interested in.

Kerr: 10:53

That's right. Yes. We'd been in New York for 10 years. We'd used up all our money because we spent a lot of time going to concerts and operas until we ran out of money. And I was on a civil service salary, which was, by general standards, pretty generous, but by New York standards was tight. We lived in Columbia housing, which was great because there were a lot of other interesting people there. We’d bump into people like Margaret Mead in the elevator and that sort of thing. But yeah, so after 10 years there, we-- oh, yeah, the other thing was Thaddeus had been casting around for his next level, his next stage of life. And he'd been negotiating with UMass and Columbia and Harvard for a position. I think he sort of played them off against each other and ended up going to Harvard to the Smithsonian Astrophysical Observatory or Center for Astrophysics, whichever they call it nowadays. And so we agreed to split the lab equipment. While he was at Columbia, Thaddeus had built this small radio telescope on the top of the physics building at Columbia. And we had-- sort of getting ahead of myself again, aren't I? Yeah. Maybe we should stick on question three for a while.

Kellermann: 12:50

Okay. So yeah. So why don't you tell us-- you've been in NRAO a long time, two phases, and you've seen a lot of changes in NRAO and the management in NRAO. And so if you could talk a little bit about how you saw this impacting the CDL and any impressions about--

Kerr: 13:19

Yeah. Well, I think what I said in my notes about Dave Hogg's son commenting that, "Oh, you're just one of the engineers," was-- sort of said something about the attitude of some of the astronomers to some of the engineers. Initially, in the '70s, we were in the Edgemont Road Building still. And then we needed to expand. So we moved over to the first Ivy Road building where we had a bit more room. And that made more of a break with the astronomers. But there were times when-- hold on. I've got a small doggy problem. Hey, Tristan. I'm in a meeting. I'm going to close the door.

Kerr’s son: 14:37

Oh, I'm leaving.

Kerr: 14:38

Okay. Yeah. So let's see. Where was I?

Bouton: 14:53

Division between astronomers and engineers.

Kerr: 14:55

Oh, yeah. After we moved to 2015 Ivy Road, the gap between us and the astronomers became even more pronounced, I think. And we had our own branch library, so we didn't go up to the library at Edgemont Road anymore.

Kellermann: 15:25

And of course, it's even worse now with the difficulty in getting--

Kerr: 15:30

It's not much different. I mean, it's a little bit further, but it doesn't take that much longer. Are you thinking because of moving out to where the CDL is now?

Kellermann: 15:43

Yeah. Yeah.

Kerr: 15:44

Yeah. That didn't make that much difference. I think just the fact that it was more than a short walk, more than just going from one building to another.

Kellermann: 15:56

I think you're right because that culture was very different in Green Bank--

Kerr: 16:01

That's right.

Kellermann: 16:03

--where we could-- not only were in the same building, you could walk downstairs and discuss things, or as any of your colleagues would describe it, I would come down and complain. I'd like to think that had some positive impact.

Kerr: 16:21

Yeah, it did. So after we moved to 2015 Ivy Road, the one person after he became Director, one person who did come and visit fairly often was-- oh, come on.

Kellermann: 16:41

Not Mort.

Kerr: 16:43

No, after Mort. He became Director.

Kellermann: 16:46

Paul. Paul Vanden Bout.

Kerr: 16:49

Thank you. Yes. Terrific, terrific aging brain working here.

Kellermann: 16:53

Yeah, I know.

Kerr: 16:55

It was Paul. Paul, yeah, he would come over periodically. And in those days, we used to have lunch in the conference room, the engineers and some of the technicians too. And he'd come over and have lunch with us. And that was really good. I don't think it would work so well nowadays because people tend not to gather for a formal lunch, although that could be organized.

Kellermann: 17:19

Yeah. I came over a lot for the-- you had the lunch talks.

Kerr: 17:26

That's right. Yeah. And yeah, I certainly remember you coming over. And in fact, I think I remember you from Radiophysics. Were you at Radiophysics at that same time as I was?

Kellermann: 17:38

No, I was there '63 to '65.

Kerr: 17:44

Well, that was probably when I did the summer course, so I probably encountered you briefly at that point. Yeah.

Kellermann: 17:55

I lived a lot at Parkes. Anyway.

Kerr: 18:03

Okay. So back to question three. You asked about work for others?

Kellermann: 18:12

Yes.

Kerr: 18:13

I think that's generally been a very positive thing. It helps to build connections, if you're talking about working for other observatories; it helps to build connections there. And as long as it doesn't hold up NRAO projects, then I think it's a good thing. So that's obviously up to the CDL director to maintain that balance. If someone asked us at the moment to make some SIS mixes for them, we'd probably refuse because we have an almost impossible task doing this Band 6 version 2 project on an almost unrealistic schedule. Okay. So, any other question three things? Question [crosstalk] processes you asked about.

Kellermann: 19:16

You want to comment on the impact of having Hein Hvatum as Associate Director as sort of an interface between the technical divisions and the Director. And of course, that stopped when Hein retired in --.

Bouton: 19:36

End of '87, Hein retired. And until then, there was this-- the sort of hierarchy was the CDL director and then Hein and then the NRAO director.

Kerr: 19:49

Yeah. And I think--

Bouton: 19:51

There was no longer that interface. And I'm curious about--

Kerr: 19:56

I think when Paul was Director, he sort of took care of establishing-  I think he had a good understanding of the engineering approach to life, and he would talk to us a lot. And so I think the absence of Hein didn't manifest itself until really after Paul's tenure. I think it was Fred who became Director next, Fred Lo. And Fred thought he would-- I shouldn't be saying this, I suppose, even though Fred's not with us anymore. But I mean, Fred thought he knew much more about instrumentation than he did. He'd had a student working on SIS junctions, it was at Illinois.

Bouton: 20:57

Illinois.

Kerr: 20:59

And so anyway, he was difficult to work with, and he didn't pay that much detailed attention to what was going on in the CDL. So my impression of Fred was that he was probably a good astronomer's director, but not so good for technical development. Yeah. What else?

Kellermann: 21:33

Yeah, let's go on. So during your tenure, you've worked for a lot of different CDL directors, Weinreb, Ballister, Webber, Dickman, and Bert. My impression is they all had very different, well, both backgrounds and styles. Do you want to comment on that?

Kerr: 21:58

Yeah, that's question four. I was just saying you want me to comment on project management while we're on question three?

Kellermann: 22:07

Okay. Yes, please do.

Kerr: 22:08

Because I think that's important. And yeah, my comment on current project management globally is summed up in that story about our neighbor in New York who was a professor in the Columbia Business School. And he and his partner would make fortunes giving these seminars to Fortune 500 company management. And they used to laugh about it. And they took acting lessons and things to polish their presentation, which may have helped a bit, but. Anyway, I think nowadays, what we saw at NRAO for a while, maybe a bit better now, perhaps as a result of Bert's interaction with upper management. For a while, we had these completely nontechnical project managers who were trying to manage everything we were doing.

Kerr: 23:25

And with engineering development work, you just can't predict what's going to happen. It's R&D. You don't know what the outcomes are going to be. Now, I added a footnote there that Mike Shannon is currently the ALMA Band 6 version 2 project manager. He's good. He was trained by us on ALMA Band 6 version 1, essentially, working under John Effland. And so he has a good understanding, not of the details of engineering, but of how engineers work. And so he's been very good. Yeah. And I think Bert has done a really good job sort of helping in that way, improving the way management of the engineering projects is done. And he's also done a great job finding recruits from around the world.

Kerr: 24:33

Alessandro Navarrini, who basically took over my job, is really just the right person for the job. He's very, very good. And Bert has found other people. He interacts with the local technical college. And so he has done a good job trying to find technician recruits, amongst others. But it's very, very difficult to find appropriate technicians and engineers to work in radio astronomy because it's quite different from mainstream electronics nowadays. I mean, it overlaps quite a lot, but there are quite great differences. And so [crosstalk]--

Kellermann: 25:35

Are our salary's competitive? [crosstalk].

Kerr: 25:37

What has affected the CDL’s mm receiver work greatly is the inability to come back to speed quickly after the loss of people like Neil and Frankie, Frankie being Françoise [Johnson], Neil being Neil Horner, who were long-term technicians. And well, Neil was technically an engineer, but he was essentially a very high-level technician. And Frankie was a technician and very good. So we're still struggling to recover from the loss of those two several years ago.

Kellermann: 26:15

Are the NRAO salaries competitive with industry? I mean, engineers and astronomers will come work for cheap because they like what they're doing. But what about the technicians?

Kerr: 26:27

Yeah. I don't know about salaries. I really can't comment on that. The people who've responded to our job advertisements, well, there haven't been that many responses. We have had some turnover of people. You get a more relevant comment on that from Bert or John Effland, probably. Yeah, so.

Kellermann: 27:06

Any comment about the other CDL directors?

Kerr: 27:11

Yeah. That's question four. Yeah. Yeah. The thing about Sandy, as I noted here, was he maintained really good connections with industry and both microwave and digital. I mean, he, of course, originated the astronomical correlator, the work that was taken over later by--

Kellermann: 27:45

Shalloway.

Kerr: 27:47

Yeah. Art Shalloway, yes. Yeah. And Sandy was good at picking technology at just the right point in its development. So for the early VLA, parametric amplifiers were a big thing commercially for satellite communications in those days. And he tapped into that. So the original VLA had parametric amplifiers and Schottky diode mixers and parametric up converters, which all worked together. And his relationship with upper management was good. And I think that was-- well, I think probably Heeschen had good confidence in him. But certainly, Hein, who was Sandy's boss, had good confidence in his judgment. Why did Sandy leave? Two main things. Probably part of it was just wanting to move on to something new. And then there was diminishing support for R&D at NRAO.

Kerr: 29:07

And I noted that there was a talk that Sandy gave in the Edgemont Road conference room before I left in 1974. It would have been probably 1973, or something, when he gave this talk. I'm sorry, 1980-something. Those dates, 1980-something. There was a talk he gave, which I remember, where he showed this graph of how the funding for CDL development work was declining over time. At the same time, the cost of supporting the CDL, supporting existing operations, was increasing. But as I noted, I asked Sandy about that recently, and he didn't remember it. So okay. Balister, he was a temporary CDL head after Sandy left. And I think he'd probably been interim head while Sandy was coming and going. I think he went on sabbatical at one point and then came back. And Mike was interim head then, as I recall. But there was never a permanent head appointed.

Kerr: 30:29

And basically, Balister was sort of a stand-in until eventually John Webber came aboard. And John was a very brilliant engineer, but he was difficult for most people to work with. I never had any problem with him, but I think he had a difficult relationship with upper management. And I got the feeling that when he finally retired, he was happy to do so. Bob Dickman was very good to work with. He paid attention to his engineers and understood them. He'd had a reasonable amount of engineering background over the years. But I think he also had problems with upper management. I'm not completely sure about that, but that was my perception. And Bert Hawkins seems to do really well with people. His background is not directly in astronomy, but he listens to engineers, and I think he understands a lot about astronomy instrumentation now. And so he's been very good, especially as I mentioned earlier, at finding new engineers and technicians.

Kellermann: 32:02

My impression is that Sandy and probably Mike, during their time as head of the CDL, were also practicing engineers and were in the lab developing things. And that was less true for the people after that. Is that correct? And is this because of the personalities of the people and their skills, or because of the changing requirements, and it's particularly administrative requirements at NRAO and the NSF by implication?

Kerr: 32:48

I think mostly the focus was on more sensitive receivers for the existing instruments and instrumentation for new things like the VLA and ALMA, MMA as it was initially. And I always had the feeling that-- well, I always just worked on what I thought was appropriate for the main projects. And I was never given any flack for that. So I don't think--

Kellermann: 33:34

Yeah, I was talking more about the CDL directors or heads. My impression, and correct me if I'm wrong, but that starting with John Webber and Bob Dickman and now Bert, they don't actually do any development work themselves, whereas Sandy and Mike did while they were CDL director.

Kerr: 33:58

I see, yes. Yeah, I think that's right. But Mike was sort of in the middle. I mean, initially in his Green Bank days, he'd been a real engineer. And then he became more of a manager. But even I remember quite late in the piece, I think he was doing the 23 Gigahertz mixers for the initial VLA, wasn't he, if I remember correctly? Not sure about that.

Kellermann: 34:32

And now it's a full-time administrative job. Is that correct?

Kerr: 34:37

Yes, it is now. Yeah. Definitely. Yeah, that's right. It's changed character quite considerably.

Kellermann: 34:47

And that's because of the demands of the job, not because of the individuals. Is that right?

Kerr: 34:58

Yeah, I suppose so. I wonder what would happen if you had someone like a Bob Wilson put in as a CDL head.

Kellermann: 35:07

You know he was offered the job once.

Kerr: 35:10

Yes, yes. Yeah.

Kellermann: 35:13

And I think he would have taken it if something could be found for his wife, Betsy.

Kerr: 35:18

Yeah. She had a degree that wasn't acceptable in Virginia, I think, was the problem.

Kellermann: 35:25

Medical degree, yes.

Kerr: 35:28

Yeah. And so they stayed in New Jersey, but then they moved to Smithsonian.

Kellermann: 35:34

Part-time. They still live in New Jersey.

Kerr: 35:39

Oh, do they? Oh, okay. Yes. Yeah. I sometimes see him at the Smithsonian meetings we do online. And occasionally, he shows up, but.

Kellermann: 35:50

Yeah. I think he's completely retired now, but for a while, he was going up there one day a week or something like that. Okay. So let's move on. The end result is the support of the observatories in Green Bank and New Mexico. And up to a point, Tucson. And I lived in Greenbank. I lived for a while in Tucson and in New Mexico. So I got to see it from the other end and sense that there were times when there was, shall we say, friction between the CDL and the local electronics people as to who should do what and who is responsible for what when it was successful and who was responsible when it wasn't successful.

Kerr: 36:54

Yes, I think that's right. I think to some degree, they felt that we were in an ivory tower in Charlottesville. And every place I've worked, there's always been some degree of tension between the receiver builders and the telescope operators or telescope users or telescope engineers if they're separate. And I think that was certainly true. But it doesn't seem to be a problem now with ALMA, but that's being done on such a huge scale. It's quite different. So yeah, I'm trying to think. I think the main cause of trouble in Tucson was these two technicians, Jack Cochran and Dewey Ross, who were basically troublemakers. They liked to stir up stuff. And so there were these fictitious, or was this one fictitious tale of misappropriation of components by the CDL, which supposedly were being sold to other observatories when Tucson needed them.

Kellermann: 38:36

I remember that.

Kerr: 38:38

Yeah, you do. I believe they actually got our business manager off the golf course. What was his name?

Kellermann: 38:47

Ted Riffe at the time?

Kerr: 38:49

No.

Bouton: 38:52

Jim Desmond.

Kerr: 38:54

Desmond, yes. I believe they-- yeah, Desmond, who had NRAO golf balls.

Kellermann: 38:59

Oh, yes, of course.

Kerr: 39:01

Yeah. I think one of the complaints, I think Jack Cochran had been on a supposedly work trip back to the CDL, at which I think he spent most of his time on social activities. And then he went back and told the guy… --who was the good guy, who was the head of business in Tucson?

Bouton: 39:29

Dale Webb.

Kerr: 39:30

Dale Webb. Thank you. Somebody's got a good memory. And told Dale, and Dale got on the phone and actually got Jim Desmond off a golf course, supposedly complaining about us selling stuff that was supposed to be.

Bouton: 39:48

Dale could be very forceful at times.

Kerr: 39:50

Dale was very good, yeah. He was terrific. Anyway, so that was the-  And then there's that problem we had with electrostatic discharge problems. I mean, it's well-known when you're around sensitive equipment. You ground yourself, you ground the equipment, and for sensitive equipment, there's a whole lot of specialized things like wrist straps and static testers and things that you need to have. And when SIS mixers became normal, they were extremely sensitive to static. And we had quite a long period when everything we sent to Tucson from Charlottesville was declared dead on arrival. And it turned out to be the fact that no one was using wrist straps or grounding the various pieces of equipment properly. And it took a while for that to sink in, but eventually it worked.

Kellermann: 41:05

Even I knew that you were supposed to do that. Yeah.

Kerr: 41:10

[Irrelevant comment redacted by Kerr.]

Kellermann: 41:23

Okay. Well, thanks, Tony. So I think you already commented on this a bit. You seem to have been privileged in the sense of being able to choose your own direction of work as opposed to being assigned by whatever CDL head it was. Correct me if that's incorrect, and also what motivated the specific work that you've done over the years?

Kerr: 42:04

Okay. Should we consider this as question seven, I guess?

Kellermann: 42:07

Yeah.

Kerr: 42:08

Yeah. Okay. Yeah. No, you're right. However, I think probably the reason that I made the choices I did was because they were the things that obviously needed to be done, which the director would have been asking me to do anyway. But yes, certainly during my New York phase, I was completely in charge. All Thaddeus wanted was initially just Schottky diode receivers for his telescope. And we ended up-- I should probably talk more about that phase because that was, in some ways, the most productive 10 years of my career. But yeah, so I think I've been very lucky to work on the stuff I really wanted to, and partly that was governed by what seemed to be needed. But also, it was good because it was the things that-- I had been basically extending what I had been doing and was interesting to me.

Kerr: 43:19

So it was the cryogenic Schottky mixers, work that was suggested initially by Sandy, and then SIS mixers and receivers and millimeter component design. I've done quite a lot of work over the years on what seems like a trivial thing, definitions of noise temperatures. But for quite a long time, it was a problem, just whether you include the zero-point fluctuations and whether you use the Rayleigh-Jeans or Planck or Callen and Welton laws for the noise temperatures of hot loads and cold loads and things. And I think that's all sort of more or less standardized in the radio astronomy community now.

Kellermann: 44:09

You mentioned somewhere--

Kerr: 44:11

I'm working on, have worked on for many years on two IEEE standards committees for waveguide interfaces, which-- basically flanges and waveguide dimensions. Seems like a very simple thing. Why would it need a lot of work over many years? But the tolerances on waveguide interfaces keep needing to be tightened, and waveguide dimensions. And there are a number of companies who all would like to have standards. And there continue to be new ideas around-- just at the moment, we're just working-- one of these groups, we're working on an idea that was originally based on something that Wes Grammer suggested, a miniature flange. And that's been extended. We're planning to, one of these groups is thinking of using that as a new standard. And so that's interesting work. Doesn't take too much time. Why? Because I enjoy it, yes.

Kellermann: 45:37

You mentioned somewhere you invented a cold pad. Is that right?

Kerr: 45:52

A cold pad? Oh, yes. Back in the early days of FET amplifier or cryogenic amplifier work, it was very difficult to do an accurate noise temperature measurement of a low noise amplifier at 15 degrees or 4 degrees. The losses of the input waveguide and the temperature distribution along it would make it very hard to calibrate. And I suggested to Sandy, back when I was in New York, I suggested using, and something we were using in New York, we put a cold attenuator at the cold temperature, which would define pretty well the noise temperature the amplifier was looking at at its input. This is pretty much standard procedure nowadays. It never got written up as a sort of formal paper or anything. We just, Sandy and I, were always in communication on and off, and it just arose as a suggestion. I think Sandy gives a reference to our conversation in one of his papers.

Kellermann: 47:20

Well, looking at it from the outside, yes, I always thought that was standard.

Kerr: 47:26

Yeah. Yeah. Was the question what I worked on for others?

Kellermann: 47:32

Yeah. So you already commented a little on that. Work for others. Is there anything else you want to add or?

Kerr: 47:39

Not really. I think it's basically a beneficial and--

Kellermann: 47:43

but clearly, as you said, when we're doing things for our radio astronomy colleagues, yes. But what about doing things for industry or the military? Do you want to comment on that?

Kerr: 48:11

I don't think I have a very strong opinion. If it seems to be helpful or useful to our general interest. I mean, obviously, the military collaboration has other factors. I mean, it brings in a lot of money.

Kellermann: 48:27

Well, that was going to be my next question. Work for others had a significant impact. Well, first of all, does all that money go to the CDL or does it get dispersed around the observatory, like paying astronomers or- ?

Kerr: 48:44

I have no idea. You'd have to ask Bert about that. Yeah. I don't allow that to bother me. [laughter]

Kellermann: 48:54

Okay. But you're not aware that the work for others had a direct financial return to the CDL. Okay. That's interesting.

Kerr: 49:07

Well, I mean, no. It may be in connection with the radar work for Green Bank, Bert's taken on one or two new people, which is presumably being covered by what we're paid for doing that radar work.

Kellermann: 49:27

Yeah. Well, of course, we're not allowed to earn a profit, but we can charge a fee, I guess. My impression is that-- well, we always charge more than, I think, our really direct cost. And I thought that those funds went back into supporting work that could otherwise not be done.

Kerr: 49:52

You're thinking of something like an overhead?

Kellermann: 49:56

Well, the overhead is a real cost in a way, but the--

Kerr: 50:02

I don't think we charge more than what it actually-- at least with the mixes and amplifiers and things we've sold. We've always budgeted them on basically a very simple idea of the man hours and material costs involved. And somebody may add on an overhead. I don't remember that, though.

Kellermann: 50:33

Oh, I'm sure they do now. I don't know about--

Kerr: 50:35

Oh, yes. Yeah. I'm sure they would.

Kellermann: 50:37

You can't sneeze here without getting charged overhead. [laughter]

Kerr: 50:42

Yeah, okay.

Kellermann: 50:48

So let's see. I was going to ask about the separation of engineers from the scientists, but you've already commented on that. I don't know if there's anything else you want to add.

Kerr: 51:05

No, I don't think so.

Kellermann: 51:07

And of course, Socorro was intermediate in that engineers were in the same building. Didn't quite have-- oh, the other special thing about Green Bank, of course, was that everybody lived together and they were neighbors and friends. And I think--

Kerr: 51:25

Yes. And they all ate in the same cafeteria.

Kellermann: 51:28

Yes.

Kerr: 51:29

Yeah.

Kellermann: 51:30

And that was less true in New Mexico. And I believe that the door to the basement now is locked unless you have a need to be there.

Kerr: 51:43

Oh, really?

Kellermann: 51:45

I think that-- with all the electronic keys and everything, maybe the scientists still have access.

Kerr: 51:53

Yeah. That would be unfortunate.

Kellermann: 51:55

I mean, certainly, in the early years, I used to wander down there and ask what they were doing.

Kerr: 52:03

Yeah. I do understand the need to prevent people who aren't trained on using particular equipment from trying to use it. But that probably wouldn't happen if the basement doors in Socorro were unlocked, I wouldn't think.

Kellermann: 52:22

Well, no, I think it's all the foreigners that are running around.

Kerr: 52:31

Oh, okay. Yeah.

Kellermann: 52:33

Anyway, let's get back to technology. One of the big things at the CDL has been low noise amplifiers. There's a variety of approaches. And in particular, I associate both Sandy and Marian as pioneers in this area, but have trouble distinguishing exactly how their work complemented each other and how it differs.

Kerr: 53:05

Yeah. Well, as I said in the notes, I think Sandy was more empirical in his approach. And Marian was very concerned with getting an electronic model that would allow him to do improved designs, and particularly the non-thermal noise sources in field effect transistors. And both those approaches, as we know, have been very successful. I think Marian's noise model is now pretty much universally used. And Marian's designed amplifiers covering the full spectrum from 1 Gigahertz up to 120 Gigahertz or so at which point SIS mixers tend to take over for astronomy. And they've been sold for many of the, probably most of the world's radio telescopes. Yeah, there's just one company now making low noise amplifiers that really are competitive with our own, probably based on Marian's theory, I suppose. Low Noise Factory in Sweden, with whom we've worked. I've worked with them. They're very good people to work with. But yeah, the fact is that radio astronomy instrumentation has, especially for ALMA, has pretty strict power dissipation limits, and also sometimes some input return loss specifications that are difficult to meet. So it's certainly something where we need to keep a dedicated amplifier group in the CDL. And we have that. The guy who's taken over from Marian is very good.

Kellermann: 55:21

Is the power dissipation--

Kerr: 55:23

Working closely with the [crosstalk]--??

Kellermann: 55:26

Is the power dissipation concern about ALMA having to do with the altitude or just the amount of power consumption?

Kerr: 55:36

No, it's in a vacuum, so the altitude doesn't matter. It's just the refrigerator capacity on the ALMA receiver package. You've got these 10 receivers, all of which have to be cooled. And there's a limit of I think it's seven and a half milliwatts per amplifier, which is extremely low by industry standards.

Kellermann: 56:07

Do I understand correctly that the initiative for cooled amplifiers came originally from radio astronomy and specifically at NRAO. Had it not been done previously by the military or by industry?

Kerr: 56:29

I think that's probably correct. Yeah. I think Sandy was basically the-- it may not be correct, but my impression is that Sandy was the first person to think about using them for radio astronomy. And he was at Berkeley for a while. There was this group at Berkeley that started making amplifiers, I think, using a similar design to Sandy's. I'm sure they had a lot of communication. Trying to remember the guy's name, Dave --

Kellermann: 57:10

Woody? No, that's Caltech.

Kerr: 57:11

No. No, that was Caltech, yeah.

Kellermann: 57:13

Yeah, that was Caltech. Yeah, I know who you mean.

Kerr: 57:16

There was an engineer. You might not know this guy's name because he wasn't an astronomer, basically. I don't think-- 

Kellermann: 57:23

But his name was on papers. I do know him.

Kerr: 57:26

Yeah. Yeah. So for a while, though, you could buy a Berkeley low-noise amplifier at L-band. But as far as I know, they're not in business anymore. There were two of them. Gosh, I wish my memory was a little bit better, that--  [Note added post-interview: Tap Lum & Dave Williams]

Bouton: 57:45

We could figure it out and insert it.

Kellermann: 57:46

Yeah.

Kerr: 57:47

Yeah. Yeah. Yeah. But that was always a friendly relationship with the Berkeley people. But basically, that was a transient thing as far as radio astronomy was concerned because they were just-- I think they only did L-band, which was sort of at the most 1 to 2 Gigahertz, and probably somewhat less than that.

Kellermann: 58:15

So very early on, we bought a lot of commercial equipment, paramps from Airborne Instruments Laboratory.

Kerr: 58:23

And Comtech.

Kellermann: 58:27

Yeah. I mean, they were very expensive, $100,000, which was real money in those days.

Kerr: 58:35

Yeah. They were basically modified telecommunications paramps. And they were used on the early VLA.

Kellermann: 58:44

Well, even earlier in Green Bank, and they were a disaster.

Kerr: 58:49

The commercial ones? Okay. Yeah. I remember the ones that Jo Edrich made. Oh, there were some Micromega paramps, one of my first jobs-- one of my first jobs at NRAO in 1971 was to go to Green Bank and fix some Micromega paramps that had got out of adjustment. And they were room temperature amplifiers, yes.

Kellermann: 59:26

And they were very unstable, especially for continuum work.

Kerr: 59:30

Yeah. Yeah, like all parametric amplifiers [inaudible].

Kellermann: 59:34

And then the cooled ones, the gain was modulated by the pump frequency, not RF frequency, but the pump cycle. [inaudible].

Kerr: 59:47

The refrigerator.

Kellermann: 59:48

Yeah.

Kerr: 59:48

Yeah. Yes. Yes. I think people overcame that by putting large slabs of, believe it or not, stainless steel in as a sort of thermal capacitor.

Kellermann: 59:58

Oh, okay.

Kerr: 01:00:00

I'm not sure whether that was a permanent solution. Nowadays, those refrigerators have a … well, a 4-degree refrigerator but you're talking about 15-degree refrigerators. Yeah. Okay. What I'm thinking is not appropriate.

Kellermann: 01:00:16

Yeah.

Kerr: 01:00:18

Yeah. Okay. All right.

Bouton: 01:00:22

Just going back to sort of the culture of CDL, you've talked a little bit about Françoise, Frankie as a technician, but how about the culture? What effect did she have on the culture there? You've had--

Kerr: 01:00:43

Yeah, I think--

Bouton: 01:00:43

--[crosstalk] engineers over the years, Nancyjane Bailey, and Priyanka now. But François was long-term, kind of.

Kerr: 01:00:55

Yes. Yeah. Yeah. I think she had a very-- she was a strong personality. She took no nonsense from engineers. And I think she had a very positive effect. She sometimes would seem to be a little bit recalcitrant in meetings. If you asked her to do something, she would [laughter] sometimes say no. But yeah, I think she was always, I think, a very positive effect. It was a little bit like having a resident psychotherapist, perhaps, or just the knowledge that she was there, maybe made people behave a bit better. It definitely was positive.

Kellermann: 01:01:52

Okay. So Tony, when we did the interviews with Sandy and Mike and distributed the transcripts, you had some comments about what they said, or maybe you had a different view on things which weren't appropriate to change in their interviews, but if you wanted to comment on that now, it would be appropriate.

Kerr: 01:02:22

Sorry. I'm afraid I haven't been back over those yet. I was leaving that till later. I will go over the ones you sent me and comment on that. I don't remember what my comments were before.

Kellermann: 01:02:40

Okay. Is there anything else you want to comment on, the whole history of the evolution of low noise amplifiers from paramps to Schottky diodes and SIS mixers, HEMTs and FETs, and so on?

Kerr: 01:02:59

Yeah. I haven't commented on TKIPs yet. I'm not sure whether that was on a different--

Kellermann: 01:03:07

Okay. Well, first of all, explain what-- I only know the K, kinetic. Explain what they are.

Kerr: 01:03:15

It's traveling-wave kinetic inductance parametric amplifiers, TKIP.

Kellermann: 01:03:30

Ah.

Kerr: 01:03:31

Yes, that's right. Yeah. The concept was invented by, at Caltech, Jonas Zmuidzinas. And it uses the nonlinearity of the kinetic inductance of a superconductor with magnetic field. So like an old-fashioned parametric amplifier, where you have a nonlinear capacitance in a diode whose capacitance depends on the voltage. With a superconductor, the kinetic inductance, which is just the inductance of the superconducting electrons, that depends on the current level. It's nonlinear with current. So by pumping a superconductor at a relatively high frequency, you're pumping a nonlinear inductor, and you can therefore make it work as a parametric amplifier at a fraction of that frequency.

Kerr: 01:04:57

Either at half the pump frequency, which in old-fashioned parametric amplifier terms you'd call a degenerate amplifier, or at some fraction of the pump frequency. And that is potentially a very low noise mode of amplification if you can make everything behave ideally, which means-- at the moment, it means cooling things below 4 Kelvin. The work that we're doing with UVA - I'm not involved in it, but what NRAO is doing with UVA - is trying to get sufficiently good superconductors at four degrees to make TKIP amplifiers out of. But it's a long-term project. I think it's something we definitely should have an involvement in. And Patricio Mena, who's leading that effort with us, and Arthur Lichtenberger's group, I think they're doing just the right thing at the moment.

Kerr: 01:06:16

We did have one engineer who was here for a couple of years. He came from Caltech with high ratings and all the usual Caltech glitz. Unfortunate story there. I don't know whether it's relevant for this, but yeah. His wife was a high-level medical doctor. She specialized in high-risk pregnancies and she was hoping to get a job here. There was a job exactly in her field advertised at UVA. And so she and Omid were talking to our Head of Human Resources. [Paragraph slightly edited by Kerr.]

Kerr: 01:07:30

And she told Omid's wife that she had a good contacts at UVA. So she should not do anything. She would take care of making sure that an application for that job was put in. Nothing happened. I think she forgot about it. Our head of HR forgot about it, probably. And as a result, the position was filled. Omid's wife was never in the list of candidates for it. And reasonably enough was thoroughly pissed about that. And she ended up getting a job in Washington, DC, and he was working here. And they had a kid. So they had a difficult problem. But apart from that, yeah, Omid was not the right person for this job. He's working at Goddard Space Flight Center now.

Kellermann: 01:08:37

Was this Sandy's student from Morocco or Turkey or something?

Kerr: 01:08:41

No, no, no, no. No. Omid was one of Jonas Zmuidzinas students from Caltech. And I think he'd done a postdoc either at Caltech or somewhere else. Should remember that. Probably at Caltech. And then he came to us with what seemed to be good recommendations to work on TKIPs. And basically, he was one of these people who felt he could be a kind of PR interface between the actual researchers and the funding agencies or the astronomers or whoever needed to be connected to the work. And so yeah, he basically didn't do much while he was here for two years or so. But that was-- I don't know. I suppose I was as responsible as anyone for him being hired for that job, and he turned out to be mismatched to the job.  [Paragraph slightly edited by Kerr.]

Kellermann: 01:10:00

Okay. Anything else you want to cover, Tony?

Kerr: 01:10:06

Oh, well, let's see. That's Sandy and Marian. We've done that. 10, 11 on the history of who developed cool mixers amplifiers. Well, I think there's no question the cooled Schottky mixers were done by Sandy and me; the first ones, because when I came to NRAO in 1971, the general argument was that cooling a Schottky diode would not reduce its noise because the noise was mostly shot noise. But that turned out to be only a part of the story. The fact that a cooled Schottky mixer can work at a much lower current level meant that it would have much lower noise. And Sandy, I'm not sure that he realized that at the beginning, but his intuition was very strong. And well, I think he probably did realize that. And so he asked me to work on Schottky mixers. I was prepared actually to work on Josephson effect mixers. That was just getting started. But he wanted to work on cool Schottkys, which was absolutely the right thing to do at that point.

Kellermann: 01:11:34

Yeah. So it's interesting you brought up Josephson. That's fizzled as far as I know, not just here but everywhere.

Kerr: 01:11:45

Yeah. I think it may be having a rebirth for quantum computing, where the signal levels are so low that you can probably make just some amplifiers that are useful. I'm not too sure about that. But in New York, for part of my 10 years there, I had a postdoc whose name was Yuan Taur, T-A-U-R. He was from Berkeley. He'd been working on Josephson amplifiers at Berkeley, or Josephson mixers, sorry, at Berkeley. And I wanted to keep working on that. And Yuan made a number of Josephson mixers, I think around 115 Gigahertz, 110 Gigahertz, but then he realized that they would never be competitive, even probably with cool Schottky mixers, because of their extremely limited dynamic range and the fact that they tended to be saturated even by their own noise. They were just very nonlinear mixers. And so he published a paper that was really the death knell of Schottky, sorry--

Kellermann: 01:13:31

Josephson.

Kerr: 01:13:32

--of Josephson mixers for radio astronomy. And about 1979, Tom Phillips and Dave Woody at Bell Labs and Paul Richards at Berkeley published back-to-back papers on  SIS receivers, which were very primitive initially. But we started working on that. And I had this student called S.-K. Pan, who we all know. He got working on that with me. Also, at that point, John Tucker had published an article on the limits of SIS mixers in which he claimed that you could have gain in a resistive mixer, and that struck me as heresy. And I called him on the phone, and we talked about it. And of course, he was right. And he ended up coming and working with us for a couple of years at GISS in New York, and that was an extremely fortunate thing.

Kerr: 01:15:06

And then Mark Feldman-- I don't know if you remember Mark. He came with Pan and me and also Peter Siegel from New York. In 1984, he came to Charlottesville. But. this little brain blocks for a minute. Anyway, Mark ended up going to UVA. He worked for Bob Mattauch's group about when Mattauch was starting to move -to VCU. And so he and Arthur Lichtenberger were closely linked during Arthur's PhD work. Anyway, where were we?

Bouton: 01:16:06

I'm interested in the--

Kerr: 01:16:08

Yeah. So that was the beginning of--

Bouton: 01:16:11

[crosstalk].

Kerr: 01:16:12

--SIS mixes. I think the first ones were made by Bell Labs and the Berkeley groups, very primitive ones. And they were developed after that by us at GISS. And then when Pan and I came down here to Charlottesville, we continued on that using UVA superconducting devices. One thing-- let's see you're on page 11. One thing I should do is show you a list of, if I can find it, all the different people who make SIS devices. I don't know where it is. Is that it? No. It's somewhere. No.

Bouton: 01:17:30

I'm interested-- stop at this slide for a--

Kerr: 01:17:32

There it is.

Bouton: 01:17:33

I'm interested in the sort of the-- you have a group that works together, and then they kind of go out in the community, but there's just this kind of ferment of this particular group that then goes out and goes to various places and has a great deal of influence. I mean, I see this with your group at Goddard there. A bunch of you then came back to NRAO, went to UVA, went to various places, and carried stuff forward. I was reminded a bit of the Bok's group at Harvard, very early radio astronomy group that had Heeschen and Roberts and Howard. I mean, so it's like this time of a certain group coming-- of people coming together at a certain time with the science or the technology, and things happen.

Kerr: 01:18:39

Yeah, that's right. And Thaddeus was aware of-- although he wasn't an engineer by any means, he had a very good appreciation for supporting engineering work. And he basically gave me a free hand at GISS. And then for part of the time, he was away on sabbatical. And I used that time to purchase expensive things like an electron microscope. He never disapproved of it. But it was just a very, very productive 10 years. And as you say, we had this group of people that came together and then continued later on. So yeah, this slide I'm showing now is all the superconducting people that I've collaborated with over the years. And generally, people aren't aware of how many people have made SIS devices or tried to. And this shows this huge list of which nowadays in the US, there are only UVA and JPL, when they want to, if they have a mission that requires it, they can re-instrument and do SIS work, but they can't maintain an SIS foundry doing continuous development work like you need and like we have at UVA. And JPL, I know because we've worked with them before, they are extremely expensive. More than twice the cost of having equivalent projects done either at UVA or at--

Kellermann: 01:20:46

They have big salaries and overhead.

Kerr: 01:20:49

--SUNY, State University of New York. Yes. They were in the SIS business briefly, not very successfully, but we worked with them for two or three years. We funded them at an equal level to UVA to do pre-production for ALMA. Well, I guess it was-- not sure whether it was MMA or ALMA at that point. I guess it was ALMA because there was money flowing to be able to support two groups. It seemed important that the most critical elements of the receivers should not have a single source of fabrication. So we were initially funding UVA and SUNY to do SIS work. But the SUNY work petered out, and we've continued with UVA ever since with great success. I mean, people point to all the delays and things, but that's true with every one of these foundries, even IBM with all their money, who we had a very successful collaboration with. Sometimes they have successful runs. Other times, things don't work out right. And so we're very lucky to have the UVA collaboration.

Kellermann: 01:22:33

Tony, one final question. You commented that you've been relatively free to do the things that you thought were important and found interesting. You had tenure from, I think, the time you came back. To what extent do you think that was important in providing you that flexibility? And to what extent, in general, do you think the concept of tenure is or is not useful for engineering staff?

Kerr: 01:23:16

I think it probably is important. If I hadn't been offered tenure, I wouldn't have come back. I was offered tenure at UVA at the same time. And I didn't look at other places. But also, well, Thaddeus was asking me if I'd go to Smithsonian with him, and I didn't want to do that. So I don't know whether they would have offered me tenure. I was on a civil service position at Goddard Institute, which was pretty close to being tenured. So I think probably I would have had a civil service position at the Smithsonian if I'd gone there, but I don't know. So yeah, I think to get people who you really want to stay, I think tenure is probably important. And I think that would be a very good thing to make sure we offer to Alessandro Navarrini, who's my successor, who is really good, and we should try and keep. And I think people like Matt Morgan, who are quite brilliant engineers, should certainly be offered tenure. So, yeah.

Kellermann: 01:24:50

Okay. Well, thank you, Tony. I think it's lunchtime. Appreciate this, all the effort you put into preparation. That was very useful.

Kerr: 01:25:02

Well, yeah. It was actually very enjoyable going through and putting it all together. It's almost all from other slides and talks and things all jumbled up together. But just going through and reviewing the past has been very interesting for me.

Kellermann: 01:25:25

Okay. I've just shut off the recording.

Kerr: 01:25:30

Sorry. I lost your sound for a second because my phone decided to ping. Okay.

Kellermann: 01:25:38

I just said I just shut off the recording.

Kerr: 01:25:42

Oh, okay. Are we finished?

Kellermann: 01:25:48

Pretty much, unless you had some more.

Kerr: 01:25:52

Oh, okay. No. If we've answered all your questions, that's fine.

[silence]

Kellermann: 01:26:10

I can't tell if we're recording now or not.

Bouton: 01:26:14

Well, I think you're-- I can't tell. Well, we'll stop or--

Kellermann: 01:26:19

Okay. Recording. There we go. Very good.

Bouton: 01:26:24

And now, you're apparently still recording.