Interview with Barry G. Clark
Description
Creator
Rights
Type
Interviewer
Interviewee
Location
Original Format of Digital Item
Duration
Interview Date
Interview Topics
Notes
The interview listed below was conducted as part of Sullivan's research for his book, Cosmic Noise: A History of Early Radio Astronomy (Cambridge University Press, 2009) and was transcribed for the NRAO Archives by Sierra E. Smith in 2016. The transcript was reviewed and edited/corrected by Kenneth I. Kellermann, Ellen N. Bouton, and by Barry G. Clark. Any notes of correction or clarification added in the 2016 reviewing/editing process have been included in brackets; places where we are uncertain about what was said are indicated with parentheses and a question mark, e.g. (?) or (possible text?). Sullivan's notes about each interview are available on Sullivan's interviewee Web page. During processing, full names of institutions and people were added in brackets when they first appear. We are grateful for the 2011 Herbert C. Pollock Award from Dudley Observatory which funded digitization of Sullivan's original cassette tapes.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons, including subsequent experiences, interactions with others, and one's feelings about an event.
Series
Unit
Transcription
Transcribed by Sierra Smith
Begin Tape 104B
Sullivan
Ok, this is talking with Barry Clark on 2nd May 1978 during a convenient power blackout at the VLA. Alright, can you tell me when you first came into contact with radio astronomy?
Clark
I was a summer student working during the summer at Owens Valley Radio Observatory summer of 1957. That was the summer after my sophomore year. I was an astronomy major and during my sophomore year I went around the astronomy department looking for a summer job. They sort of chased me over to the radio astronomy people. So I got that job.
Sullivan
What did you actually do that summer?
Clark
During the summer I lived at the Owens Valley Observatory. And at the time there was a 26 MHz interferometer there that had been built during the preceding winter, which consisted of half wave dipoles strung on 4x4s about a quarter wave above the ground. And the first thing I was put to doing that summer was enlarging this dipole covered area, which involved shoving brush and dirt out of the way with a bulldozer - figuring a telescope with a bulldozer - and digging post holes for 4x4s.
Sullivan
But this is John Bolton’s style I gather?
Clark
Well to some extent, yeah. John is a very good bulldozer operator.
Sullivan
Was this all in preparation for the occultation?
Clark
No, the idea behind this instrument was to monitor ionospheric fluctuations and it was just left running all the time. And part of my job that summer was to keep it in chart paper and complain if it wasn’t working. One of the reasons it wasn’t working usually was that it was a power supply with a bad contact in it. So a swift kick in the power supply usually restarted it.
Sullivan
Now were you looking at radio sources to do this? Was it just a transit thing?
Clark
Right. The beam was adjustable at declination in rather coarse steps by changing lengths of feeder lines and was just a transit instrument in right ascension. And so it just saw what went by, Cas, Cygnus, Taurus, and, I guess, Virgo, were immediately obvious on the records. You could see other things if you wished but you had to look for them.
Sullivan
What was the purpose of this? Just something to do…?
Clark
To a fair extent. The idea behind building it, they said, was to look for the ionospheric fluctuations. So we complied a goodly amount of data about what were days of severe fluctuations, and how severe did they get, and so forth, which never saw the light of day.
Sullivan
Yeah. What I’m wondering is here is this Owens Valley Radio Observatory. It seems like a rather strange thing to do. It’s not really astronomy.
Clark
Well, that was very much in the line of radio astronomy of the day.
Sullivan
I agree. The idea was to go to higher frequencies, certainly, eventually. It was going to be directly useful.
Clark
Yes, the real motivation was to get an instrument on the site as quickly as possible.
Sullivan
So, I guess it was the following summer, now that I look at it, in June ’58, when you did the… or was there anything else that first summer, I’m sorry.
Clark
Let’s see. When was the paper?
Sullivan
The paper was ’58 and the occultation was also June ’58. [J.G. Bolton, G.J. Stanley, B.G. Clark. A Solar Occultation of the Crab Nebula at a Wavelength of 12 Meters. PASP 70: 594, 1958. http://adsabs.harvard.edu/abs/1958PASP...70..594B]
Clark
Right. Well, there wasn’t much else going on the first summer. They were going doing… the east-west railroad track was already there and I did a little conduit screening. But as far as the eventual interferometer, mostly it was puttering with this 26 MHz machine, just keeping it going.
Sullivan
Was that the first interferometer with a railroad track? Now that I think about it.
Clark
I believe so.
Sullivan
Was there any uncertainty about whether that would work, or uncertainty about putting a telescope on a railroad track, do you remember? Or was that pretty straightforward?
Clark
Well, there was a bit of uncertainty in that the design which supported the 90 foot telescope on only eight wheels, and, not very big wheels at that, was loading the wheels far beyond normal railroad practice. The theory was that we would be saved by the fact that we were never moving very fast, and that the total amount of traffic over a given amount of rail was not very great. As a matter of fact, if you go back and look at the rails now, you will see that they are sort of squashed over at the top. But they still work. The rail itself was very nearly Civil War surplus. It was military surplus, mostly 60 pound rail; considered very light by present standards and put on a concrete sleeper.
Sullivan
What do we have on the VLA here for comparison?
Clark
This is 80.
Sullivan
80 pound.
Clark
Anyhow, some of the rails were dated and the earliest date I believe was 1881.
Sullivan
Oh, you’re not exaggerating!
Clark
Not much.
Sullivan
So you came back down the following summer and what did you work on then in the summer of ‘58?
End of Tape 104B
Beginning of Tape 105A
Sullivan
Barry Clark on 2nd May ’78.
Clark
The first instrument that Caltech radio astronomers had was a 32 foot telescope on Palomar Mountain.
Sullivan
Right, I’ve heard about that.
Clark
Which was used mostly by Tom Matthews and Dan Harris for a hydrogen line survey. Then as soon as the Owens Valley site was reasonably ready, they moved this 32 foot to Owens Valley. So Thanksgiving of '67 a group of random undergraduates bossed by Dan Harris went up and finished taking the 32 foot apart at Palomar. And the pieces were carefully labeled and carried to Owens Valley and reassembled there. And sometime in that winter or early spring, at home at Caltech campus, there were fiberglass panels made for it. The original surface had just been fairly coarse mesh and had been good enough for 21 cm. But it had been John’s idea that when it was reassembled at Owens Valley, it would have this fiberglass surface and hopefully be good to somewhere in the neighborhood of 3 cm.
Sullivan
I see.
Clark
So that summer of ’58 then, we took the fiberglass panels that had been cast at Caltech and started putting them in the dish, putting the dish together. And when we put it in and started surveying it, we discovered, of course, they didn’t fit. And for reasons that are far from clear to me, the circumference of the outside of the panels was a couple of centimeters too great. So it ended up being my lot to reduce that couple of centimeters by means of a handheld sander.
Sullivan
Once again real radio astronomy.
Clark
Yes. Incidentally, the 32 foot has never worked once it was taken off Palomar.
Sullivan
That’s what I was about to say. I’ve never seen any 3 cm publications out of Owens Valley. I have seen it over on the junk heap. It’s still there.
Clark
Yes.
Sullivan
But what’s the end of that story, just to finish that one off.
Clark
Then we tried to make the fiberglass into a reflector by rolling ordinary kitchen aluminum foil, gluing it down to the surface. And then sometime the following winter that all blew off, which is one of the reasons it never got reused.
Sullivan
And interest just had shifted after that I suppose.
Clark
Eventually the interest shifted to making that a radio link interferometer by ’62 or ’63, I guess. They had concluded that is what should be done with the dish. It was lifted off its permanent foundation and put onto a set of surplus airplane wheels. But again, the radio link electronics never had taken high enough priority to actually work.
Sullivan
How far away was it going to go roughly?
Clark
Well, its line of sight had to be from the design of the radio link, but the line of sight in the Valley there could be up to 30 or 40 kilometers, I guess, without any trouble.
Sullivan
Well, I’ve never heard about that dish except up on the mountain. That’s interesting. Well, tell me about this occultation in ’58.
Clark
Well, again I was just sort of keeping track of making sure the machine had chart paper, and so forth, in it, and wasn’t really aware of what was happening. And John Bolton, in looking over the records, rather to his surprise, noticed that the Crab wasn’t there one day.
Sullivan
Oh, you mean it was accidental?
Clark
Oh, yes.
Sullivan
Oh, I see. Well, this experiment had been done a few times before. This was accidental.
Clark
It never occurred to anybody that that was an appropriate machine for getting an additional point on the spectrum of it.
Sullivan
And you happened to be looking at Taurus A that day fortunately, I guess.
Clark
Well, as I say, the beam was pretty broad, probably parked somewhere near the zenith. That’s enough to get the Crab. So noting when it went away… well, eventually John put the paper together and put my name on it because I had been faithful about putting chart paper in…
Sullivan
I see. What it a matter that he looked the records and said what happened this day and then all of a sudden it occurred. Or was it a matter of someone said, “Gee, there was an occultation a couple of weeks ago?”
Clark
No, no, it was a matter of his looking at it and saying, “Hey, where the Crab?”.
Sullivan
Now that’s strange. What else did you work on that summer? Was it mainly still monitoring with that low frequency?
Clark
Well, that was a small part of my time. It was mainly working on the 32 foot and miscellaneous small chores. More pulling wires and putting in conduits for the 90 footers that were being built.
Sullivan
But you were still an undergraduate, so I guess you went back to school again.
Clark
Right.
Sullivan
And the following summer, was that the next time you…
Clark
Yeah, yeah. So that summer they had single dish receivers on ninety-foot one. And Dan Harris had been using, I believe, a 960 MHz receiver, measuring positions and looking at doing essentially the CTA list of sources. And Radhakrishnan was working on putting together a 21 cm receiver. The 21 cm receiver was a real beast. He was still using it to frequency switch so there wasn’t a great deal of stability as a continuum instrument. That spring John Bolton had been taking observations of Centaurus A, just drift curves. One or two a day as the object drifted through at 960 MHz. And I did some of the plotting. John did some of the plotting of that. I did the reductions on it, subtracting the point source and integrating various parts of it, finding out how much flux was in each thing.
Sullivan
Modeling it more or less?
Clark
Sort of, I guess.
Sullivan
Yes, yes.
Clark
A double with a point source in the middle and a curious extension on one side. It is not clear anymore whether that belongs to Centaurus A or our Galaxy.
Sullivan
You haven’t mention Gordon Stanley. What was his role in the development…?
Clark
He was the electronics guy. He did most of the electronics design and assembly. There was a technician there who had been there for many, many years, John Harrington, who did a fair amount of the assembly. Gordon, for the most part, designed those mixer receivers that were really a very nice piece of engineering at the time. They were narrow band but they have a factor of 2 lower noise temperature than anybody else’s mixers.
Sullivan
Really? I see. What were they? Do you remember the value?
Clark
About 250 [degrees] double-side band at 21 cm.
Sullivan
So now we are up to the summer of ’59 actually and I guess that’s when you graduated.
Clark
Yeah.
Sullivan
And obviously by that time you were already were a radio astronomer.
Clark
Well, to a very minor extent. About that time also I did a little Schmidt print looking at Dan Harris’s positions, which didn’t come to too much. Mostly pretty well known identifications, those that had identifications…
Sullivan
This wasn’t accurate enough to make new ones yet?
Clark
Yeah. It was a couple of minutes, maybe even 5 minutes, uncertainty. Anyhow, I took Polaroids of the positions and marked the coordinate systems on them.
Sullivan
What was the operational procedure to do that? Find reference stars nearby and then just offset them more of less?
Clark
Yes, yes, using the AGK2 [Astronomische Gesellschaft Katalog 2]. My procedure was to pick three stars and just solve for two plate scales and rotation. Other people did more elegant things later.
Sullivan
Well, this is pre-computers still, I guess, or did you have a small computer to work with?
Clark
I always worked it out with a slide rule. A couple of years, or a year later, Fritz Bartlet was doing the same thing. He used more stars and more plate constants and programmed a little computer.
Sullivan
So, as a graduate student, you mentioned this attempted identification work. What things did you get involved in otherwise?
Clark
Well, the interferometer worked for the first time, I guess, that winter. And then in the summer of ’60, they essentially parked Ken Kellermann and I at the observatory to run the place for the summer. They said, “Here’s all the pieces. Go ahead and hook it up as an interferometer. It’s been done before. You won’t have any trouble.” Which, of course, we did. But, by golly, that was a good education. We learned so much, so fast.
Sullivan
So you got it working that summer?
Clark
Yeah. It started out on Ken’s thesis measurements, the source spectra. And I can’t recall when it was that we first got the line interferometer going. Sometime after that summer. But there were a couple of things that needed building — the set of crystal filters, for the 30 MHz, or 10 MHZ IF rather. We just couldn’t do anything without these filters. The previous set of filters had been about 30 kilohertz wide, just too wide.
Sullivan
You needed a swept frequency LO also, I guess?
Clark
And their out-of-band rejection wasn't good enough. We also needed the swept frequency LO, which was designed and built by Bob Wilson.
Sullivan
You say that it had been done before but, in fact, you were operating at higher frequencies than people usually did interferometry and perhaps than anyone had done interferometry. Is that correct?
Clark
Well, what I mean was, that Bolton, Stanley, and Radhakrishnan had gotten the thing working at 960 MHz early that spring. And then had disassembled the interferometer and gone by to single dish for some stuff. And then Ken and I got there and they said, “Ok, hook it up.”
Sullivan
But, nevertheless, this high frequency, for that time, interferometry was a good bit trickier than low frequency stuff, I would suspect. Or is that not true?
Clark
Yeah. I don’t know. I’m only familiar with the one interferometer at that time.
Sullivan
That’s right. That’s what you were brought up on.
Clark
The baselines were considered fairly long except in comparison to the Jodrell Bank machine. But those baselines and wavelengths were still respectable except for the very longest elsewhere.
Sullivan
What was the first science that you began doing with the interferometer?
Clark
Well, it was quite a number of things. I don’t really recall the order except the first thing we did was Ken’s thesis observations, which I was serving as an observer assistant on. And Radhakrishnan was working on polarization at the time. I helped again as an observer assistant and what have you. And that eventually went into [George] Seielstad’s thesis. Because of the fact that I had done or had been involved in this Crab occultation at low frequency, when the AFCRL [Air Force Cambridge Research Laboratory] people said that they had an anomalous broadening in the Crab at about a gigahertz, I don’t recall, I was very interested in that so I did some observations with John Wyndham to check that. [J.D. Wyndham, B.G. Clark. Occupation of the Crab Nebula by the Solar Corona in June 1963. Nature 200: 766, 1963] We were able to set 20 times or 100 times smaller limit than what they said.
Sullivan
Now this was a map of the Crab or was this an occultation again?
Clark
It was an occultation comparing the same baseline when it was near and far from the Sun. Since the visibilities were the same, we were able to set a limit of the scattering that was very much smaller than the Crab itself. Despite the fact that we didn’t have enough phase to map it, we made observations of polarization at the time. We were hoping to the get the Faraday rotation out and therefore the magnetic field and electron density of the Sun’s corona.
Sullivan
The rotation would come from the comparison with the NRL [Naval Research Laboratory] observations at higher frequencies?
Clark
No, just watch the polarization angle as it gets near the Sun.
Sullivan
Oh, I see what you are saying. I see.
Clark
It was essentially defeated by the complexity of the Crab. You really need a polarization map or a shorter baseline than we had to figure out what was going on there.
Sullivan
One question which comes to mind, maybe not directly relevant to you though, in retrospect one can look back and say that the Owens Valley interferometer could have been aperture synthesis at a very early stage. Do you have any theories on why that didn’t happen? Or was there no need?
Clark
Well, we were very aware of the possibility. With the plus or minus 4 hours in hour angle, you have to have a lot of configurations to do it. So people tended to do the one dimensional synthesis as Al Moffat did, rather than to try to do the two-dimensional thing. We did tend to think in terms of aperture synthesis. If you asked would an observation at this configuration give you need data or duplicate something you already had, then we would plot it on the UV plane. And think of it in terms of a map like that.
Sullivan
But no one quite had the interest to put in the dedication that would be required?
Clark
To do a full two-dimensional synthesis would have required just very many configurations. And even then it wouldn’t have been too good because of the missing stations. That interferometer was built when much less was known about the structure of sources. So the stations were exponentially spaced rather than uniformly, or near uniformly. So while the original spacings were at 200, 400, 800, and 1600 feet, (they've added more since). It had restricted stations, restricted hour angle coverage.
Sullivan
Still it would seem like that might be of great interest, to come up with a single high resolution map. But that’s all in retrospect, of course.
Clark
It was just difficult. [2016 note added by K.I. Kellermann: Alan Moffet later claimed that Caltech, unlike Cambridge, did not have the computing power to do a full 2-D synthesis.]
Sullivan
I guess we are up to the stage where you began working on HI interferometry. Tell me about how that went.
Clark
Well, it went fairly smoothly actually. I did a little playing around with the receiver. And, as a matter of fact, cobbled up a receiver that only I used. In my life I have designed a total of two circuits. One of them was built into that, and the rest of it was miscellaneous pieces that were lying around. One of the problems with doing line interferometry with mixer receivers is rejecting the image. And the image rejection scheme we used was to offset the LOs at the two antennas. And because it was convenient, we offset them by 2 MHz. And then, we used a couple of different schemes. We had a broadband and a narrowband receiver. The narrowband one, we just mixed one IF back up. The broadband one was sort of cute in that the two IFs from the two antennas were just multiplied together to get the 2 MHz fringes. And then we did an ALC on that. It was a slightly more elegant scheme than the ALC on the IFs. And then just detected 2 MHz fringes. But that worked pretty well. The real question was how deep were the deepest absorption lines. And it was a lot of instrumental fiddling around about how much contribution was there from the wings of the filters. A number of questions like that called for a careful instrumental investigation.
Sullivan
And also I guess the confusion with how much small scale structure emission might be…?
Clark
Yes. For that reason, I looked at two or three regions of emission with reasonably long baselines. The sort of things towards the anti-center where things are really quite hot. And saw no fringes at the levels of interest so I was happy with that. Although with the more sensitive receivers now, you can see fringes at 200 feet anyway. On occasion…
Sullivan
Yeah, like you say, that is much more sensitivity. When you went into this experiment was the idea was that there might be a different component to be seen in absorption that in emission? Or was it just with…
Clark
It was already clear that the absorption lines were much narrower and much spiker-looking than emission lines that were general in the same direction. And this could be explained on either of two bases. That there were temperature segregated clumps and the absorptions lines see only the low temperature ones and emission lines see mostly the high temperature ones. The basic suggestion there was John Bolton’s. I just took it and worked it out. Another possibility, of course, was that it was not temperature segregation but spatial segregation. So that if you looked at an emission with a 2 minute beam, you might also see this.
Sullivan
This was the original idea wasn’t it of [John] Hagen and [A. Edward] Lilley and…
Clark
Well, they didn’t phrase it quite that clearly but that was…
Sullivan
It was implicit in what they had.
Clark
Right. And the thing we did that sort of convinced me was, well, just looking at enough sources to get a statistical handle on what the absorption lines look like statistically. And then you can readily add a few up and see whether you get something that looks like the emission. And the answer, of course, is that you don’t, which starts making the two component model look pretty good.
Sullivan
And where did the idea of possible pressure equilibrium come, which as I mention seems like sort of a throw-in in your 1965 paper?
Clark
Well, once again, we knew at the time that these clouds are not dense enough to be stable, gravitationally stable. So they must either be made furiously or must be in some sort of pressure equilibrium. There is no very convincing argument either way even now, but throwing in the assumption of pressure equilibrium will let you do a few things that are interesting. So why not.
Sullivan
Let’s see. Is there anything else that you worked on at Caltech that we haven’t covered?
Clark
Well, there were the Venus measurements. This must have been the winter of ’62-’63. An inferior conjunction of Venus, and I wished to resolve the planet and attempt to measure its diameter. And to do that, 10 cm and 1600 feet was just barely adequate. And it was the first time the interferometer had been used at that baseline and frequency. It had been used at each separately. And I got the calibration pretty badly loused up. And the result was that the diameter came out much too large. And this would have fit in with some of the ionospheric models at the time which called for limb brightening. But I didn’t have a great deal of confidence in it at the time. And then just before I left in the summer or spring of ’64, [Arkadii Dmitrievich] Kuz'min was visiting. He and I redid Venus.
Sullivan
Did that first result get published?
Clark
Yes, it was at a meeting on Venus at JPL [Jet Propulsion Laboratory] and it had extended abstracts or something. [Citation to full paper: B.G. Clark, A.D. Kuz'min. The Measurement of the Polarization and Brightness Distribution of Venus at 10.6-CM Wavelength. ApJ 142: 23, 1965, http://adsabs.harvard.edu/abs/1965ApJ...142...23C]
Sullivan
Yes, I have seen that report. That’s where [Frank] Drake also… he was working at JPL, I guess, looking at water vapor and things like that. Well, in conclusion, as you look back at the early days of Caltech, what would you say were the main contributions that Owens Valley made to radio astronomy of that era?
Clark
Moffet’s thesis observations put the double radio sources on a solid footing for the first time. Showed there were so many of them. Radhakrishnan’s and [G.L.] Berge’s Jupiter observations tied that down for the first time. And I’d rate the hydrogen line stuff below that. But it's still useful. And let’s see, Matthews' positions were potentially useful. They didn’t have quite the pay off one would have hoped.
Sullivan
Is there anything in the approach in the way that things were done at Owens Valley or the equipment that was there that was different or enabled one to… do you see what I’m trying to get at?
Clark
Yeah. There were a lot of ideas in that receiver, a few ideas for the first time. Many of them that we have copied shamelessly ever since. And then the development of mechanical analog computer for fringe stopping. Jodrell Bank did the same, more or less independently from Caltech. The fringe stopping, automatic delay tracking. I don't know who has priority on that. The interesting tricks you can do with mixer receivers to sort out side bands if you are so inclined. The fact that mixer receivers are insensitive to IF phase shifts. And as I said, those mixers were very good for the time.
Sullivan
What about the idea of the north-south, east-west… or was there only east-west at this time, track?
Clark
The year that the north-south was added was probably ’60 or thereabouts. I remember I was out surveying in rails. Ken Kellermann was soldering together wires.
Sullivan
Do you see that as a major step in the development of radio astronomy, just the idea of having rails and putting dishes on them? Or was that sort of…
Clark
That’s a point. It hadn’t occurred to me that was the first large railroad-mounted interferometer. I don’t recall how that Jodrell Bank remote element moves.
Sullivan
They just carry them around in the back of the truck I believe, certainly in the early days. I was just talking to Dick Thompson. And then I think also even into the early 60s, I think, they were just carting them around the countryside. They were operating at lower frequencies, of course, and didn’t have to have dishes. I don’t think they had dishes, did they?
Clark
No.
Sullivan
Well, I think that covers it pretty much. That ends the interview with Barry Clark on 2nd May ’78.