Interview with Samuel J. Goldstein

Goldstein: 00:00

Can't take history there?

Sullivan: 00:01

Oh, you can. Oh, yes. You can major in history, then minor--

Goldstein: 00:03

Well, but I was a graduate student at Stanford and took a course in Electrical Engineering. But I did my thesis at Harvard and was going to be an astronomer. [inaudible] and-- so I got into the OH scheme in a very complex way. First of all, there's an electrical engineering meeting in the Boston area, a NEREM meeting.

Sullivan: 00:37

Yeah, I'm familiar with it.

Goldstein: 00:38

--and I was a session organizer for the talks on radio astronomy, often given from a kind of engineering point of view sometimes.

Sullivan: 00:46

And what year are we talking about now? What year?

Goldstein: 00:50

Well, that was in the early '60s. And in '63, in November, one of the papers I managed to organize was Arno Penzias' description of his search for OH in emission at Bell Labs which was unsuccessful.

Sullivan: 01:07

Okay, now I wasn't even aware of this search. I know about some others, but he had tried to do this at Bell Labs?

Goldstein: 01:14

Well, for me the OH business started in about October 1st or late in September of '63, and Ed Lilley had a preprint of the MIT effort at Millstone Hill where they made the first detection of OH in absorption and almost coincident with the appearance of the paper, in fact, somewhat before it was published in Nature, there was this paper by Arno Penzias given in a NEREM meeting describing a negative result.

Sullivan: 01:50

I see. Was that ever published anywhere besides the NEREM Record?

Goldstein: 01:52

No, NEREM Record is the only place, but that is [crosstalk].

Sullivan: 01:55

But I didn't know about that reference, though. That's interesting.

Goldstein: 01:58

You could look it up. So, there I was as the session organizer, and I was the chairman. There were four papers. I can't remember the others right now. But I knew that OH had been detected and here was one of my speakers, that I had chosen, who is saying that OH couldn't be detected, at least the way he was doing it. So that was an interesting position to be in. Well, I tried to [inaudible] him a bit, and the MIT fellows were there at the meeting. I think Sandy Weinreb was one, that Lit Meeks was another, and they didn't want to talk at all about their work because it was in publication in Nature and they didn't want to speak until the publication actually appeared which is ridiculous.

Sullivan: 02:44

Interesting the Australians [crosstalk].

Goldstein: 02:49

Bizarre. They started earlier. Really my first connection with this was a few months before I had visited MIT, and they were getting ready to detect OH. And I think they were willing to admit that they were getting ready to work on OH. And they had an older channel spectrum analyzer with a fairly narrow range and about 50 channels, and they tried it out on the hydrogen line at Millstone Hill which is a very intelligent plan of action, to make it work on something that's known before putting it on something that's unknown. And I'm not quite sure how I got out there. I think, if my memory is right, Bob Price was one of the fellows who got me there. And we decided we'd go out to Millstone Hill and observe a hydrogen line, and I was the expert on the hydrogen line. I could tell them where to point the telescope and how to set the frequency and so that was really a lot of fun. It turns out though with the aid of several PhDs. One PhD had to set the oscillator frequency, and another guy could talk to the machine and point the telescope. Those are computer-controlled experiments and the computer program is very inflexible. It had to do a very certain thing. So, it was impossible to tell the machine to point at the source that I had in mind. I knew of a narrow strong emission line that came from a large region in the sky so that he didn't have to point with great precision. And to make a long story short, after spending a whole afternoon at it we were unable to get the thing pointed. They had no sidereal clock. That paralyzed me. But I called up Agassiz Station and got the sidereal time, [laughter] which is very nearly the sidereal time at Millstone Hill, and we still failed. So, if you'd ask me then, because those guys do anything, [inaudible] I would have told you that they couldn't possibly. But that would be quite incorrect because their detection of OH was a very workmanlike job. And the paper they wrote on it was really very conclusive [crosstalk].

Sullivan: 05:14

[crosstalk].

Goldstein: 05:16

You could scarcely doubt that it was a real detection which wasn't always the case [crosstalk].

Sullivan: 05:20

This interested you Harvard obviously?

Goldstein: 05:23

No, not at all, as I was a 21 cm observer.

Sullivan: 05:26

I see. [inaudible].

Goldstein: 05:28

It interested Ed Lilley, that's true. As soon as they had positive results, he was very much anxious to observe OH. And he asked, "Under what circumstances could we do it?" And we had this Cooper-Jelley-Bloomberg maser which has a very narrow tuning range, couldn't possibly be adjusted. The local oscillator wouldn't cover the right frequency range and there were numerous other difficulties. But I thought that there was a good chance to observe OH with a crystal mixer type receiver or a tunnel diode by getting long integration times. And I had improved the Harvard receiver to make it possible to get long integration times. We had digital data processing then at a very early stage. There was analog processing first and then an A/D converter. And it punched cards.

Sullivan: 06:25

This was on the Agassiz 60 Foot?

Goldstein: 06:26

Yeah. And it was, at that time, a five-channel receiver. And so, I had tried to find 21 cm radiation from a globular cluster and used some long integration times just shortly before this project. And I had succeeded in getting theoretical noise for 10,000 seconds integrations, the same length of time which was quite adequate for observing what the MIT group saw with a crystal mixer receiver. So, I advised Lilley to do that, that it was possible to improve things by using a tunnel diode receiver. And Arno Penzias had a tunnel diode that he had tried unsuccessfully to see emission with.

Sullivan: 07:13

Now, where did he try that and what telescope?

Goldstein: 07:15

It was one of the, I think he calls it the 24 Foot. It was one of the Bell Telephone Laboratory’s forward horns. And, of course, with a small horn-- maybe it was a 10 ft. I get those two mixed up. But with a small horn, only emission is possible. And so, he was anxious to check his observations.

Sullivan: 07:43

Why was his so unsuccessful in hindsight? [inaudible] not in the right place, or?

Goldstein: 07:49

The line is very weak in emission and only visible and, so far, is only seen in [anomalous ?] places. Some HII regions. And I guess in dust clouds also [I can see?] emission. The knowledge you have of the sky from the 21 cm hydrogen line just doesn't apply because the excitation temperatures of the two materials are so different. So, in hindsight, he had chosen the wrong experiment.

Sullivan: 08:19

Yes. He just didn't look in the right places for each, [I mean?].

Goldstein: 08:22

Well, his antenna-- even today, knowing everything, it'd be extremely difficult to see it with the small telescope he was using. I think it's probably faintly possible.

Sullivan: 08:32

[inaudible].

Goldstein: 08:33

So, we took out the maser from the focus of the 21 cm receiver-- the focus of the Agassiz 60 ft telescope. And that was quite a brave thing to do because it had been a very powerful receiver and was still capable of good work. It was at that time the only 100-degree system in astronomy. I think there were-- well, the Bell Telephone Laboratories had a 30-degree system, but not on a big telescope, and even not on a sizable telescope, a 60-footer. It was a--

Sullivan: 09:11

So, this was to retune.

Goldstein: 09:13

--good thing. Well, what we eventually did was to make another maser very similar to the first one with a-- but with a different piece of ruby and different cavity lengths. And I'll tell you about that after a while.

Sullivan: 09:26

Okay. Yeah, [inaudible].

Goldstein: 09:28

Let's stick to chronological order. Lilley said what can be done to observe the OH line? And I said, well, with Penzias’ receiver, we should be able to get a good signal-to-device ratio for the features that have already been seen. And I was interested in the emission too since our 60 ft dish was smaller than their 84 ft dish. It looked like the direction to go. So, we started up right away, got another oscillator cavity, and made it work at 18 cm. That was a fairly-- just a couple days' work. And we operated switched load operation. We still use liquid helium, even though our maser had been removed. And--

Sullivan: 10:17

So, you were expecting just a couple of degrees signal. So, I mean, you were--

Goldstein: 10:21

Well, let's see. I think it works out to be between half and one degree in our 25 kHz filters. We're going to use 25 kHz filters that are roughly [inaudible] size. Arnold was there for a couple of weeks. We got the receiver mounted and going. I also started the work. I told the machinist who was there to begin making duplicates of the parts of the maser. It seemed like a good engineering principle not to tear up the very good old system, but really to make duplicates. And he came to Agassiz station every day and didn't have anything else to do, so.

Sullivan: 11:02

Also, Ellen Gundermann was involved in this, was she not?

Goldstein: 11:05

Yes. Not quite yet. But she was a student there and was looking for a thesis. And I think she probably came to the project in December. Let's see. I remember she was certainly with it while I was working on the 18 cm maser. Because she made an important contribution to my understanding of that. There was quite a crew there. There were two observers who pointed the telescope. And there was Ivan Johnson-- two technicians, Ivan Johnson and Dexter Smith, who built the electronic things. The machinist, [Karan/Karin?], used to come out to Agassiz station several times a week partly to-- mainly to maintain the optical things, but he was usually available to do mechanical things for us. And there was another technician [Milwaukee?] who maintained the 60 Foot. He would lubricate the thing and did mechanical work on it, and--

Sullivan: 12:08

So, when did you finally get the thing running?

Goldstein: 12:10

Oh, it was--

Sullivan: 12:11

The first test, or whatever?

Goldstein: 12:13

Let's see. We began observing in mid-November, I think, on a temporary basis, and then the first really convincing observations came on November 23rd or 22nd. It was the day Kennedy was shot. I was sitting--

Sullivan: 12:31

The actual day in '63?

Goldstein: 12:32

Yeah.

Sullivan: 12:32

Oh. [laughter] That's an easy thing to remember.

Goldstein: 12:37

I remember we'd had some preliminary observations that might have shown the effect, but I had three different two-hour-long observations, one of which had the Cas--

Sullivan: 12:49

Which source is this, now?

Goldstein: 12:51

This is the single feature in Cas A. We had the Perseus Arm source, I guess it's called. On one of those records, there was a low point in the middle, by design, and another record there was nothing because we moved off the source, and for a third record, it had been moved over to the fourth channel instead of the third. You can understand that five-point spectra don't give you the most convincing possible presentation of what's happening, but having those three records, I think, would convince anybody, especially-- you have also a theoretical signal-to-noise ratio and a kind of real one from the four points that you can see--

Sullivan: 13:30

[inaudible] message.

Goldstein: 13:31

--to give you an idea of what the signal's noise is. And there was the feature at more or less what you predict it to be from having read the paper by the MIT group. So, we--

Sullivan: 13:43

You mean the absorption feature?

Goldstein: 13:44

Yeah. The single absorption feature. On the day Kennedy was shot. And I was plotting those points when somebody yelled, "Hey, the president's been shot."

Sullivan: 13:51

So, this was that night before that you'd taken them or that morning, or?

Goldstein: 13:54

Yeah. And so, I remember taking these things to see Lilley a couple of days later. I had described them on the phone to him more or less right away. Perhaps the third one that didn't actually exist that day. But I took the stuff to Lilley, and he was at his house and he was watching the funeral of the President, which-- and I told him, there's no doubt in my mind that we'd seen the effect. They were right, though, [which?] does emit in our present equipment--

Sullivan: 14:41

Well, absorbed now.

Goldstein: 14:42

Yeah. Okay. Absorbed. Thank you. Well, this is a really-- well, let's see. How does it go? Well, anyhow, I said we ought to build a maser for 18 cm. Arno's receiver was supposed to be 6 or 700 degrees, but it turned out to be 900 or a 1000. Still a bit better than the crystal mixer we'd had, but we needed considerable improvement. Arno had, incidentally provided another gadget, which was very helpful to the project. He built a current-controlled attenuator, which allowed us to match the temperatures exactly. We put that in the reference arm, and that was a very nice thing. And so as soon as we -- Lilley also thought it was very important to increase the number of channels, and I didn't really object to that at-- five channels is clearly not enough. And we had built a model of cheaper and somewhat simpler phase detector and were ready to start constructing more. But the key thing, it seemed to me, was the maser, and I didn't want to be committed to do those other things on any schedule, but I was willing to. So, let's see, this was late '63. Anyway, we did start to work. We went to see Bloembergen to ask for advice on the maser. I don't really understand the process of maser action in theoretical terms very well, but I understood how to build a cavity maser. And Lilley wanted to see to what extent Bloembergen would help, and he wanted to get Bloembergen's opinion on whether I could do it or not, I think. And so, we had some discussion--

Sullivan: 16:58

So, now it was--

Goldstein: 16:59

--on how to do it.

Sullivan: 17:00

This receiver you're talking about now, this maser, was the one that the emission was detected with--

Goldstein: 17:03

Yes.

Sullivan: 17:03

--eventually?

Goldstein: 17:04

The one that we started with. Yeah.

Sullivan: 17:05

I just wanted to make sure.

Goldstein: 17:06

Yes.

Sullivan: 17:07

So--

Goldstein: 17:08

So anyhow, that was a lot of fun, building a maser, and we-- it's--

Sullivan: 17:12

How long did that take?

Goldstein: 17:14

Oh, it must have taken three or four months. Trying to remember. Some of it [crosstalk]--

Sullivan: 17:20

Was there any time pressure at this time? Did you know that Berkeley was also in this game?

Goldstein: 17:22

Oh, yes. Of course. Lilley decided that it wasn't worthwhile publishing our evidence. Nan Dieter was then at Air Force Cambridge Research and went to Berkeley shortly after. Let's see, it must have been the spring of '64. And, anyhow, she and a group at the Air Force, using a receiver that Harold Ewen-- or, using a preamplifier that Harold Ewen had found, published a paper on OH. Pretty much the observation that [?] at MIT--

Sullivan: 18:14

Oh, I didn't know that. Where is this paper? Do you remember?

Goldstein: 18:16

Oh, it's in Nature.

Sullivan: 18:17

It's in Nature? I don't remember that.

Goldstein: 18:18

Yeah. Ewen and Dieter and probably somebody else.

Sullivan: 18:20

I didn't remember that one.

Goldstein: 18:21

And there's one by Weaver--

Sullivan: 18:25

Yeah, the Weaver one and--

Goldstein: 18:26

--about December in Nature. Late December, early January. Those papers were more or less simultaneous. We could have added to that one more.

Sullivan: 18:35

So, they were working on OH even before Dieter got out there in Berkeley?

Goldstein: 18:39

Oh, yeah. Yeah. But those papers didn't add that much--

Sullivan: 18:43

Yeah. Just confirmation [essentially?].

Goldstein: 18:45

The original paper was so conclusive that it hardly needed a kind of-- oh, let's see, a kind of inaccurate confirmation. So, I don't place the time on this too accurately, but sometime in the spring of '64, we had the maser working, and by summer working pretty well. And it had a performance very much like the original 21 cm maser.

Sullivan: 19:21

You mean in the lab you had it working?

Goldstein: 19:22

No, on the telescope.

Sullivan: 19:23

On the telescope?

Goldstein: 19:24

Yeah. No, let's see. I'm trying to think of how to date this. My wife and I and our kids went to Puerto Rico for a vacation, and it must have been May of '64, and it wasn't really working before that, and it started working pretty soon after we returned.

Sullivan: 19:42

So, once it got working--

Goldstein: 19:44

So, the first thing we did with it was to -- well, we did look at Cas A and got a better signal-to-noise ratio, but, still, our resolution wasn't adequate [crosstalk].

Sullivan: 19:54

Still having 24 kHz?

Goldstein: 19:55

Yeah. That was the narrowest filters we had. We had 80 kHz, and we were then we then had increased to 10 channels. But we were striving very hard to find emission. And so, we picked a long path through the Center. I think our point was l=4, b=0.

Sullivan: 20:14

Also, excuse me, but by this time Ellen Gundermann was--

Goldstein: 20:18

Yes, Ellen joined up-- Ellen had joined probably around Christmas time. And she was doing her thesis as soon as I think [inaudible]. She was scanning the three of the four transitions. And the maser can be fairly easily re-tuned.

Sullivan: 20:41

Right. Which one was missing?

Goldstein: 20:42

Let's see. I think she did 1612 and 1720 was missing. But I may have it backwards. So, I think emission was our goal from an early stage. And one place that we looked at a lot was l=4, b=0, kind of minimum in the Galactic Plane with a long path line through the Center. And eventually we saw not emission there but absorption. And we also looked at the Galactic Center, Sagittarius A.  And we got a publication out of both of those two spectra. And it was [Penzias, Gundermann, and ?].

Sullivan: 21:30

In Nature?

Goldstein: 21:31

In Nature. So that'll give you some date on the thing. I think that Gundermann’s thesis-- probably her thesis observations didn't begin until after that was done.

Sullivan: 21:45

So, you had the thing working. You've seen other absorption. How did you get onto emission source?

Goldstein: 21:51

Well let's see. Realizing the very different conditions that existed between hydrogen and OH, we just started exploring looking at optical sources. And I'm trying to think of the ones we looked at. All the standard optical sources. We looked at Virgo and Orion.

Sullivan: 22:10

Crab.

Goldstein: 22:12

Taurus. I don't remember Orion. What were some of the others? This is probably going to overload your tape.

Sullivan: 22:23

I know. I just make sure-- because when it runs out you don't have any audible sign.

Goldstein: 22:26

I see.

Sullivan: 22:27

Now you turn it over.

Goldstein: 22:29

Well, we saw absorption features in several of those sources and no emission in any of those directions. Trying to think of some of the other-- why don't you stop it for a second, see if I can recall.

I remember feeling that it was surprising that you could see a fairly strong feature in Cas and nothing in Cygnus I thought, "What's different about these sources?" One-- they're both near the Galactic Plane. Cygnus is at roughly 90 degrees longitude. So, there's a long path line through the spiral arm. My gosh there ought to be something there. But we got pretty long integrations without much evidence. The hydrogen line is weaker there too. And I thought, "Well, I wonder if Galactic latitude could be effective?" Cygnus is six degrees out of the plane and Cas is only two-- is it two or two tenths of a degree? I can't recall.

Sullivan: 23:30

It's much closer to [inaudible].

Goldstein: 23:32

One night I looked at several sources that were chosen just to be closer to the galactic plane. By God they all hit.

Sullivan: 23:42

Is it like Westerhout’s Catalog?

Goldstein: 23:43

Yeah, and it came out of Westerhout’s Catalog. W51 was one of them. So, the next thing I do is go through Westerhout’s list and we had I would guess a dozen sources with OH absorption quite a sizeable catalog for those days. And W49 came out after that first dozen. And we were using, first 80 kHz filters in this survey. And Ellen called me up on the phone one day. She was in Cambridge then. Our machinery punched cards and she would submit them to the computer and get the spectrum. She said, "Sam, what's going on in 49?" And I said, "It's the 49th source on--

Sullivan: 24:31

Westerhout’s.

Goldstein: 24:32

-- Westerhout’s List." She says, "Wow. We see strong emission."

Sullivan: 24:37

This is with 80 kHz filters as well.

Goldstein: 24:39

Yeah, we did 80 kHz filters. And she said, "W51 probably has some too. It's not really as convincing." So, we told Ed Lilley and started to work the next day on 25 kc filters. The procedure is that you give the observer a set of coordinates and instructions. How long the integration? What filter width is? What the frequency should be? And he does all the rest. And the cards automatically-- came in the next morning. I don't remember who took them. In piles of three then. The machinery, especially in those days, I didn't know how to punch the cards or how to feed the computer. But we soon had-- we did all the usual things. We changed the settings.

Sullivan: 25:48

Went off the source.

Goldstein: 25:50

Went off the source. There was no question it was real. It was a huge signal to noise ratio. We eventually got to 3 kc filters, which I had bought the parts for but never put together. But we had a big rush project. I think these filters. By then I reached a-- I don't think I gave the instructions anymore after the 3 kHz filters were completed. This must have been in January.

Sullivan: 26:28

Had Berkeley published their thing yet?

Goldstein: 26:32

No, the Berkeley publication was in June or something. I probably got the date wrong. But there was a lot of complex things going on that affected the operation of the station. It was pretty well established that this was my last year there. I thought I was going to be leaving at the end of the year. That would have been January 1, 1965. But in fact, I left in June. I went to Charlottesville. So, there was a lot of dissension between Lilley and myself, especially in those last days preceding them. And I remember once while I was writing all the observing instructions for the crew there. That probably stopped about the end of the year, '64. That Ed apparently felt that he ought to play a more detailed part in running the station, especially since I was going to be going. So, he called me in and said, "Goldstein, let's make an observing schedule." This was in Cambridge. And I said, "Geez Lilley you can't make the observing schedule here. I need all kinds of data. I got to know the coordinates of all these sources and what the sidereal time is. Standard time. All that's out on my desk [back at Agassiz Station?]." And he said, "Well, let's agree what sources are going to be on the list tonight." As usual I would say, "Oh, come on Ed. Geez, I can't tell that until I see what's available." So, he finally gave in. It was impossible to do even general instructions on what's to be observed without all details in hand. I used to do all that. So, I got back out and wrote an observing schedule as usual.  What I was doing very much through my last year there was trying to map data. And we made a survey with a traveling wave maser. We tried to find neutral hydrogen--

Sullivan: 28:52

This, your thesis or--

Goldstein: 28:54

[Virgo?] Cluster. No--

Sullivan: 28:55

Thesis--

Goldstein: 28:56

Long before that.

Sullivan: 28:57

Okay, you were on there as a [inaudible] or anything?

Goldstein: 29:00

Well, I was lecturer on astronomy and the radio engineer in the Harvard College of Observatory and Research Hall. Three different appointments at Harvard. So, I built a horn with the aid of the [carpenters?] out there and Russell Anderson and his nephew. We called it the Virgo Horn because it had a pattern and this guy calculated to be the same as the Virgo Cluster. We set that up and stuck the traveling wave maser that Tracy had built at the labs. That –

End Tape 2A

Begin Tape 2B

Goldstein: 00:00

It's complex, man. I think you can lay it through a conflict between Lilley and Goldstein. By then our relations were really so bad that we weren't even polite toward each other. And I'd say that I told him rather emphatically that this stuff was true without any doubt, not to be published. And anything I would say he would contradict and he decided it needed more checking. So, this was before the use of 3 kc filters on it. We've done this with 25 kc filters.

Sullivan: 00:36

And he must surely have known about Berkeley working on the same stuff, though. Did he feel still obliged to some pressure, or?

Goldstein: 00:42

That's an interesting question. He felt that W49 was an obscure source and it was the only one he knew that we looked at least a couple dozen sources. And I think that he felt if Goldstein was for it, it must be wrong, regardless of the reasons.

Sullivan: 01:03

And I guess, did Ellen Gundermann want to publish it also?

Goldstein: 01:06

Oh, I'm sure she did. Yeah. But as a student, she couldn't really apply any pressure of any sort. She had to do what her advisor wanted.

Sullivan: 01:17

And was it only finally when Berkeley came out with that [crosstalk]?

Goldstein: 01:21

Well, what happened was that Lilley, as soon as-- let's see. I think as the observations with 3 kc filters began, I sort of left that and went to my [Virgo?] horn and that was all I was doing toward [inaudible], that plus some other important centimeter observations done without the 60 Foot. And he wanted to observe the other transitions, all four transitions. And that was done in due course. It took him a few weeks to retune the maser each time. I had to act only in an advisory capacity there because Ivan Johnson understood the function of the maser very well. And he would retune it. And also, let's see, I'm leaving out Hayes Penfield who was there, and who began then constructing what was the Harvard 21-channel receiver. A portable receiver, quite similar in principle, but quite different in design from the one that we had then, the 10-channel. And so, he was nominally in charge of the--

Sullivan: 02:42

three [inaudible].

Goldstein: 02:43

Yeah. And Gundermann was hard to work on her thesis by then. In fact, sometime in mid-January-- I cannot recall these times at all. I was looking for a job. I went to Texas at Christmas time, I remember, and I went to Colorado and was working on the 21 cm stuff, and.

Sullivan: 03:14

Well, Gundermann put the stuff in her thesis.

Goldstein: 03:17

Yeah. And I didn't really mind too much. I thought my permission should be asked, though. I felt I had some rights to this data, but it never was asked. Of course, I served on Gundermann’s committee and that was a kind of embarrassing experience at times. She was plainly one of our good students and I hope that this kind of stuff doesn't really--

Sullivan: 03:44

Oh no. I mean, belong in your book or something.

Goldstein: 03:48

She knew that she mustn't defend me, regretfully. In fact, we were more or less simpatico as I understood her position. She understood law. And I told her I expected her cost of having the stuff in the thesis to say exactly, somewhere in the thesis, under what circumstances it was taken. Well, she did this on a piece of paper stuck into the front for the copy that I had--

Sullivan: 04:13

Only for your copy.

Goldstein: 04:14

--and removed for Lilley's. It really was funny in ways. But of course, lots of times it didn't seem funny. But anyhow, I felt sure that Lilley was going to publish that stuff eventually. And how could he keep from having my name on the publication because I had chosen the source, given the instructions in the observer, built the receiver, and various other valuable services. But he did. He called me up after I left, and I think they tried to measure polarization and I don't really recall the details. He said, "How about my sending you the original stuff on OH emission detection," he said to me over the phone, "you and Gundermann can work on it yourself." And I said, "Well, it's pretty late."

Sullivan: 05:05

By this time Berkeley--

Goldstein: 05:06

Almost two months after Berkeley had published it. "But I'll look at it," I said. And he said, "Well,"-- He didn't say, "Well, then I can remove your name from it and include your work," but that's what happened.

Sullivan: 05:21

So, who was on that paper, finally? Lilley, Penfield?

Goldstein: 05:25

Lilley, Penfield. I don't know who else is on it. But they referred to the detection by Gundermann, Lilley, and Goldstein or something in that order--

Sullivan: 05:38

[inaudible]

Goldstein: 05:38

--of OH, yeah. And then they gave us references in Gundermann's thesis which struck me as really irregular.

Sullivan: 05:45

But she only mentioned you in your copy.

Goldstein: 05:48

Yeah. Well, as far as explicit reference to my name, of course Gundermann's thesis had some kind of acknowledgement.

Sullivan: 05:56

Yeah. Yeah. Now did Berkeley-- as far as actual detection date, who was first?

Goldstein: 06:02

Oh. We were way ahead.

Sullivan: 06:03

And that - When did that happen?

Goldstein: 06:06

I don't recall the details, though.

Sullivan: 06:07

It was a few months, at least.

Goldstein: 06:08

They published very quickly, as soon as they had good evidence of the-- and when was their paper? June?

Sullivan: 06:15

It was the summer sometime, yeah.

Goldstein: 06:18

And our detection must have been in March. It could have been February.

Sullivan: 06:26

And then Ed Lilley had a heart attack in here somewhere, didn't he?

Goldstein: 06:29

Yeah. And that could very well have influenced him somehow.

Sullivan: 06:31

Lit Meeks, he said that he felt that if Lilley hadn't had the heart attack, that the work might have gone faster or something like that.

Goldstein: 06:40

Yeah. Let's see. Well, the work wouldn't have been any different, but--

Sullivan: 06:48

[inaudible]. But he may have been a different man, you're saying.

Goldstein: 06:50

Yeah. I'm sure there are extenuating circumstances from his point of view. I'm sure that I wasn't an easy guy to work with, so, at all times. But I really built in most [sensitive?] radiometry in existence then. You might argue that I didn't really build the Harvard maser which was the top flight piece of equipment. And that's true. I didn't. On the other hand, it was a one-channel device and I made it into a multi-channel device, and I conceived of the digital data processing. And though I didn't make it myself, I wrote the specifications under which it was made and made it work. So perhaps one more thing that did gripe me a lot is that that same receiver was used for quite a bit of other work. And Lilley had to refer to his receiver in each of his papers and he didn't do it on the first-- let's see. Yeah. My name was never mentioned in--

Sullivan: 07:55

Any [other?] [inaudible]?

Goldstein: 07:56

--in any of the stuff that followed. So, I don't feel I was terribly fairly treated there, but--

Sullivan: 08:05

Do you remember when that heart attack was or when that occurred?

Goldstein: 08:08

Oh, golly. Let's see. I remember coming to visit him once. I know it was pretty close to the time that I was leaving. And--

Sullivan: 08:23

Oh, so this was long after you detected the OH line.

Goldstein: 08:34

Yes. Long after. It could have been published before his heart attack, I feel pretty sure.

Sullivan: 08:38

Oh yeah. Okay. Well, is there anything else--

Goldstein: 08:46

Relevant to this.

Sullivan: 08:47

Not necessarily this, but any other project I might not know about that you've been associated with some historical value?

Goldstein: 08:53

Oh, yeah. Gosh. Sure. Lots of.

Sullivan: 08:58

Yeah. Let's try [inaudible].

Goldstein: 09:01

Oh boy.

Sullivan: 09:01

The chapter on radar is [crosstalk].

Goldstein: 09:03

You're really getting it from the horse's mouth or something. So, let's see--

Sullivan: 09:10

When are we here?

Goldstein: 09:12

Well, in 1959, '58. The MIT group wrote a paper on radar echoes from Venus. And in 1959, a year and a half later, they made a subsequent attempt to see radar echoes.

Sullivan: 09:33

No, these are negative results.

Goldstein: 09:35

The first one was a positive result claimed using the Millstone Hill equipment at a frequency around 1100 Mc, I think. And in 1959, several groups looked, if I remember--

Sullivan: 09:52

Where was that published, excuse me?

Goldstein: 09:53

Hudson Science.

Sullivan: 09:55

Hudson Science. And who is this?

Goldstein: 09:57

Well, Price is the first author. It must be, or [inaudible]. And in 1959 at Jodrell Bank, there was an attempt or a much cruder one, much less powerful transmitter. But I think by then the 250 ft telescope was working. And they published an equivocal paper supporting the MIT size of the Solar System. You see the whole problem is making the first guess on the range. They looked at five range gates-- geez, this sounds like my five great-children. Five range gates, the middle one giving an answer, if the MIT value for the Astronomical Unit is right. And--

Sullivan: 10:45

A very marginals confirmation [inaudible].

Goldstein: 10:47

Yeah.

Sullivan: 10:48

And which group was [inaudible]?

Goldstein: 10:49

Evans and Thompson. Evans anyways, mine is one of the observers and he said there's a 5% chance that our results are wrong. But--

Sullivan: 11:00

This is in Nature.

Goldstein: 11:01

This is in Nature. But we think the MIT value for the solar parallax is all right. So, this was my brother's. Well, my brother had gone to work at Jet Propulsion Laboratory in January of '58, and by that fall had become a student at Caltech. And a part-time employee at JPL. And so, in '60, I guess it was, they began to make plans for the March '61 conjunction. And I got hired as a consultant to help design a ranging system. My brother built a spectrum analyzer.

Sullivan: 11:50

What is his name?

Goldstein: 11:51

Richard M. Goldstein. He's a leading radar observer at Caltech now. And he wanted to measure the spectrum of the echo and with a autocorrelation type receiver, and incidentally, what radio astronomers call a Weinreb receiver, although this particular machine was built considerably earlier than Sandy's machine or different astronomical [inaudible]. So, I went out there for a few days to think about how to do this. And brother and I jointly conceived to the idea of doing ranging on the planet in addition to the experiment of measurement spectrum.

Sullivan: 12:42

Now you believed the MIT results?

Goldstein: 12:44

No.

Sullivan: 12:45

You didn't?

Goldstein: 12:46

We didn't.

Sullivan: 12:46

You felt they still needed that--?

Goldstein: 12:48

We felt they needed confirmation.

Sullivan: 12:50

Was it generally believed in the astronomer community?

Goldstein: 12:54

I don't think so. There's considerable skepticism, although they could be right, when the technique was extremely complicated. There are a lot of talented people involved in it, so you wouldn't want to say it's wrong. But considering the belief of it, I do recall that there is a new way of measuring the Solar Parallax in those days. Pioneer 5 spacecraft went [beyond?] the Sun. And I can't remember the name of the guy who produced that data, but he came up with an AU that disagreed with the MIT value. So, I'd say that it wasn't well thought out.

Sullivan: 13:36

Yeah, I sidetracked. You're talking about you and your brother. You say you developed the autocorrelator, or?

Goldstein: 13:40

No, brother invented the autocorrelator independently of Sandy Weinreb, actually.

Sullivan: 13:46

And a couple years earlier?

Goldstein: 13:47

A couple of years earlier. But the parochial nature of we astronomers is--

Sullivan: 13:51

Right, the same nature. Probably someone else did it earlier than that in some other fields.

Goldstein: 13:56

That's probably true.

Sullivan: 13:57

Some electrical engineer.

Goldstein: 13:58

Yeah. Well, in fact, Dan Black is the guy who analyzed the clipping method of autocorrelation. So, he certainly deserves some credit.  Oh, yes, well, I can't avoid sidetracking, but I'm sure the Dicke radiometer is not really first invented by Dicke. The switching principle is such a fundamental thing in physics [inaudible].

Sullivan: 14:21

Right. Well, so just a null experiment, essentially.

Goldstein: 14:24

Yeah. Everybody knows that that's the way to do physical experiments, to compare nearly equal things and look for a small difference. That's a kind of intuitive idea. But anyhow, I'm forgetting that. My brother and I working together, I think I proposed the first thought of just making the radar work like a Dicke radiometer, send out equal signals or something. I kept saying brother made the really vital contribution by saying, 'Well, you don't need a switch for this Dicke radiometer. All you have to do is turn off the transmitter."

Sullivan: 15:01

Right. In time.

Goldstein: 15:01

In time, yes. And so, you have a switch waveform, and you look at it with the--

Sullivan: 15:05

So, this was the first pulse radar. In other words, the MIT was the--

Goldstein: 15:08

Well, the MIT radar was pulse. You might call it CW. It worked in a square-wave fashion. The duty cycles were 50%. Yes, so the Caltech method, they had a CW transmitter. Well, either ways [inaudible]. But at the time of closest approach, when the radar signals are strongest, the Doppler velocity goes to 0. So, one can't do it by measuring velocity. You need to measure range. Well, if you think about a Dicke radiometer that has a switch that comes disconnected from the phase detector, suppose there's an arbitrary phase difference, there's a kind of triangular autocorrelation function that you measure, and it's perfectly adjusted. You're at the peak of one of these things. And if you're 90 degrees out of phase, you pass through 0, and another 90 degrees, you get a negative, a minimum and so, if you switch very slowly, even though you don't know the delay in the source system, you'll still be all right. So, that was the basic principle. You start switching very slowly, well, taking into account what you do know about the distance to the Sun. Then you speed up the switching and measure the phase difference that's required to bring you to the peak of the autocorrelation [inaudible].

Sullivan: 16:31

And this is what you did. What year was this now?

Goldstein: 16:33

This was the conjunction of Venus in '61.

Sullivan: 16:38

All right. And where was--?

Goldstein: 16:39

So, this was about the 1st of March '61.

Sullivan: 16:40

And where was this published?

Goldstein: 16:42

Oh, published. Well, it's in various internal reports at the Jet Propulsion Laboratory.

Sullivan: 16:55

But you got a definite return from Venus.

Goldstein: 16:58

Oh, yeah. You can see it in the data and--

Sullivan: 17:01

But if you felt that confirmation was what was needed for the MIT results, why did you not publish it?

Goldstein: 17:08

Why didn't we publish our own? Wow. Well, from my point of view, I had advised how to design an instrument. I hadn't done it at all. I was there to see it. There's several other complications. Since you don't know the rotation of the planet, you have to have filters of various size to adjust it to the return spectrum [inaudible] have an optimum signal-to-noise ratio. So, we started off with the pessimistic assumption that Venus was rotating like Earth and worked eight octaves downward to narrow our filters because we had an eight-channel recorder and stopped there. So, when I was there, on the narrowest filter, we could plainly see the echo. And we did various things. You turn the transmitter off and six minutes later, the echo disappears.

Sullivan: 18:07

Right, which is pretty much a clincher.

Goldstein: 18:09

Yeah, in that case-- so we did that several times. And that was really all the participation I had in the experiment.

Sullivan: 18:18

So why didn't your brother? Wasn't he interested in the--?

Goldstein: 18:21

Well, he was anxious to finish the thesis. And radar echoes from Venus might be a reasonable thesis, but he wasn't convinced of it. He wanted information on rotation. But anyhow, here, let's replay that a bit. It was the 1st of March, and it was nearly-- must have been 20 or 30 days, 20 days before the closest approach. And that time, he had to redesign the experiment and make a series of narrower filters in his receptor through the signal-to-noise ratio, which he did. And the same thing applied to his correlator, that is the resolution of it was too broad. Now it had to be redesigned. So, he had a very frantic two months getting that stuff changed to be ready to make observations. And his point of view is to get a thesis.

Sullivan: 19:13

I see. Yeah.

Goldstein: 19:14

And we were all JPL employees. I didn't feel [crosstalk]--

Sullivan: 19:17

So, did any group ever-- what was the next group to--? This was using an 85 ft of Goldstone?

Goldstein: 19:24

Yeah, there are a lot of-- two 85 Foots in Goldstein [inaudible].

Sullivan: 19:30

One sending, one receiving.

Goldstein: 19:31

One sending, one receiving. There were a series of observations of Venus, and I think I got paid $80 a day consulting, which seemed to me to be a generous thing. And I didn't complain.

Sullivan: 19:44

Yeah. Well, thank you very much.  This ends the interview with Sam Goldstein at the Green Bank high-velocity clouds conference, 6 September 1971, namely Labor Day.