Interview with Gart Westerhout on 15 July 1993

Gart Westerhout, interviewed by Steven J. Dick at the U.S. Naval Observatory, Washington D.C. on 15 July 1993.

Papers of Gart Westerhout

Contact Archivist for rights information.

Oral History

Gart Westerhout

Audio cassette tape

1993-07-15

Additional Materials About Westerhout Series

Begin Tape 3 Side A

Dick

Our last session was on March 5 [1993], and quite a bit has happened since that time. Curiously enough, the last thing we were talking about was Senator [Robert C.] Byrd and the clocks in Green Bank. Just a few weeks later, we had a major episode. Would you like to run down what happened with the Green Bank situation?

Westerhout

Yes. I've forgotten what we talked about by way of introduction.

Dick

We were talking about how we originally got into Green Bank and the VLBI and that there was a new push to get the clock system.

Westerhout

To get the clock and the new VLBI antenna there. That was a push that was initiated by Senator Byrd. We always felt that we did not want to operate in Green Bank because of the overhead in Green Bank. The overhead cost and so on was enormous. So already in 1987, we had attempted to remove ourselves from Green Bank, and the senator at that point intervened and said to the Secretary of the Navy basically, "No way."

So the oceanographer, Admiral Seesholtz, was called on the mat, basically, not by the senator, but by the CNO [Chief of Naval Operations], saying, "You can't do that without asking permission." So he had to cough up, out of his own pocket, the conversion of Green Bank to a VLBI station, because we had made it clear that we would no longer get compatible data, accurate enough data, out of the interferometer we were operating at Green Bank. So it had to become a VLBI station if we were going to stay there. So that started that whole VLBI effort in Green Bank, thanks to the interference of Senator Byrd.

Then around about two years later, the great big 300-foot telescope fell down, and the senator imposed upon the NSF [National Science Foundation] to replace that telescope with a new one, a bigger and better one. The NSF said, "That's not in our plans, sir. That was a telescope that was on its last legs, and we have absolutely no plans, nor does the astronomical community, to build a telescope in Green Bank. How about the LIGO, the big gravity wave machine that they were planning to build in Stanford?" They presented him, indeed, with some designs saying, "This would be a marvelous use for Green Bank, a nice place to put it." The people in Stanford didn't think so.

"Naah," said Senator Byrd. "I want a telescope, a big one. So as a result of that, the NSF said, "Okay." "And I want to know how much it will cost." So the NSF provided a number, and the senator then increased the NSF budget. He didn't simply order the NSF to take it out of somebody else's pocket; he increased the NSF budget to pay for that telescope. So it came out of no other scientist's pocket in the NSF. At that point, of course, everybody was enthusiastic.

Very shortly thereafter, the senator's staff contacted the oceanographer of the Navy and said, "You guys are doing things in Green Bank, and we're planning to build this large telescope in Green Bank, but we'd like to increase the Navy presence in Green Bank also. So can you do something about that?" We said, "No. We don't have any particular uses of Green Bank."He said, "Well, let me know what you can do."

So what do you do then? You informally provide--that was the oceanographer of the Navy, not us--some information as to what you could do. What could you do with 1 million? Well, for 1 million we could do a lot of work in the shops there--there's a good electronics staff in Green Bank--related to the Time Service. We could use their shops. That's basically what we came up with. This was all in conversation. I was never involved in that conversation. That conversation took place between one of the officers on the oceanography staff.

"Naah," said the senator's staff. "How about $10 million?" Gulp! Far too little, a million. So at that point we then came up with the two projects: one, replace the VLBI antenna that we are now using, because that's more than twenty years old. It'll probably have another five, or maybe even ten years' lifetime, but it would be nice to get a real modern one that can really see the whole sky rather than the limited motion of the old antennas and so on and so forth. And the other thing is to build a ultra high precision precise time reference station--UHPTRS. Similar to the sort of things we have been building in Hawaii and other places, but with greater capability. At some point the name "ultimate clock" was added to that, not by us at all. Somewhere people became to perceive that as an ultimate clock, which an ultra high precision time reference station is, of course. It has very high precision and will be capable of standing in for the master clock if that were so.

So in the course, then, of that development, the senator plussed-up the Navy by $10 million to build the clock and the VLBI station. Now, I have never quite figured out whether that was Navy money which was simply redirected, or where it came from.

Dick

In any case, it was in March that the press got hold of this thing.

Westerhout

No, no, wait a second. This is an important story, because I think the chronology is important there. So that was FY '92 money that we got. In the course of time, the guys in Green Bank had been talking to the senator, and they had been telling him, "With this great big telescope, we really need a new control center, an operations building."

The senator's staff asked how much that cost. Some estimates were made, and this Joint Operation Center was going to cost about $4 million in addition to the clock vaults which we expected to cost about a million and a half. In fact, we were not going to build any clock vaults. We were going to put the clock in the basement of the interferometer building, which is big enough. We had budgeted to put environmental chambers in there. But the senator said, "No, I want a building."

So he appropriated, or told the Navy to appropriate, $5.4 million MILCON, construction money, to build a Joint Operations Center in which the Navy had no role, we had no rooms in there or anything, except for operations of the VLBI, and a clock vault. We insisted the clock vault should be separated somewhere in the hillside.

All that was beautifully designed by the NAVFAC people in Norfolk, and the Joint Operations Center was entirely designed with the help of the Green Bank staff. We were not involved. The clock vault was the absolute tops of a thing that you can build that is vibration-free, that is air-conditioned to the hilt, a tenth of a degree, that is fantastic humidification and so on, three separate vaults, lots of backup power, this, that, and the other. One typical example is that everything was strictly redundant, except that in the final review, Paul Wheeler found that pipes that pipe the coolant to the air-conditioners were single. At the end of the one pipe, there were three units. At the other end of the pipe there were three units, but the pipe itself was single. So that was a single point of failure. So that was quickly changed.

So this was a beautiful, beautiful thing, and it was almost ready. It went out for bids, and the bid packages were almost ready to be opened. In the meantime, of course, we had that $10 million, so we started designing a new telescope and we started procuring clocks, some of them through NRL [Naval Research Laboratory], the cesiums, some of them through NRAO. However, NRAO somehow or another felt that they needed a new memorandum of understanding (MOU) to govern this whole clock business, and they sat on that for a long time. They also figured that with this so-called joint operations center, they needed the memorandum of understanding.

Both NRAO and we submitted drafts of a new memorandum of understanding to NSF in July of 1992. In the end of November, the NSF finally came back with their answer. In the meantime, they said, "We're not going to spend any money on clocks until you've signed the memorandum of understanding." Well, the memorandum of understanding said, among other things, that the operations building was, of course, a Navy building. Therefore, Navy was responsible for furnishing it, heating it, cooling it, maintenance, and the whole thing, which, of course, since we weren't using the building, was outrageous.

So I countered with a slight revision of the MOU, which added one sentence: "And the NSF shall pay normal cost per square foot for occupying that building." So they called and said, "What is the normal cost?" I said, "Ten dollars per square foot." Whatever the number was, maybe it was $100. Anyway, it came up to $200,000 a year they had to ante up to, and they said, "No way!" I said, "Hey, as far as I'm concerned, also no way." I mean, how to get a person to change his mind quickly is by charging him money.

So that negotiation went on for a while, and they dropped all of their requirements. They were going to furnish it, they were going to maintain it, as long as we provided the building. Then it went through Navy lawyers. We were just ready to sign that contract and thereby then the NSF would have released money to pay for the clocks, the masers and stuff like that, that they had already sent them.

They sat on it, when some journalist, after pork, called. He had been reading the congressional testimony over the years, and there are several pieces of congressional testimony, including testimony from the oceanographer, that said, "We do not need this." There was a letter from the secretary of the Navy on record, said, "We do not need it." The letter of the secretary of the Navy said, "However, if the Congress feels it is in the interest of national security to build a clock and a VLBI thing there, we will interpose no objection." So the Navy had interposed no objection, but the Navy had not said, "We badly need this stuff."

Well, somehow or other, this journalist came upon that. The call came to me eventually, and he asked me--he read from the testimony--"Is it true what I read here, that the Navy doesn't need that?"I said, "That's true. The Navy doesn't need it. The Navy didn't ask for it." And a few more of those leading questions, and before you know it, the next day it was in the paper, "Dr. Westerhout states that the Navy doesn't need it. This is pork." This was only the clock. He never discovered anything about the VLBI. So that became Senator Byrd's clock.

Dick

And it was repeated in the papers around the country.

Westerhout

Around the country. People started writing to their congressmen and so on and so forth. Well, it then took quite a while for the Navy to make up its mind, because there was one other peculiar thing in this whole story. This whole business of removing ourselves from Green Bank came up again in FY '92. In FY '92, we basically did not have money to operate Green Bank, so Senator Byrd, in his largesse, provided $900,000 to operate in Green Bank.

For FY '93, we had no money to operate in Green Bank, so we were ready to send a letter to NSF saying, "In spite of all these efforts, we're going to remove ourselves from Green Bank, because the Navy has not provided operating money," in the light of the letter of the secretary of the Navy, which said, "If Congress provides the funding, we will interpose no objections," since Congress hadn't provided the funding. So Congress provided the funding to operate in FY '93.

Now we're at the end of FY '93, and you may ask what is happening in FY '94. Well, sir, it is currently submitted as an unfunded project, operations of the Green Bank station in FY '94. It's still not funded. However, they have told the senator that they will fund it.

Dick

This is for VLBI.

Westerhout

This is for VLBI now. The disposition of the clock was that the senator decided to take the ball in hand, because the Navy was sort of sitting there wondering what to do, and simply rescinded the money. He says, "I have determined that this clock, with the changing world situation, this clock is no longer in the interest of national security, and I am the first, of course, to save the budget of the United States." And a big press release.

Dick

But he was really bowing to public pressure there.

Westerhout

He was bowing to public pressure, but he turned it neatly around. All his supporters said, "Oh, that Senator Byrd. He's such a good guy." Boy, oh, boy. So what that finally resulted in is that there is about $5 million for VLBI to build a new antenna and all the accoutrements. We did, in the meantime, already receive twenty cesium clocks, so they are now being integrated in Washington and in Richmond, Florida, and so on--twenty brand-new cesium clocks and some other auxiliary equipment. We were able to give back about $3 million of the 5 million for the clock.

The clock vault was completely canceled. The operations building--the Navy is going to talk with the senator to see whether they can transfer that money to the National Science Foundation. The National Science Foundation said, "Yeah, sure, we'd like to get that." That's probably still going to be a process that will lag for a while, but I'm sure that's eventually happening.

We are washing our hands, in other words, of this building, because that was like a big building, a $3.5 million, $4 million building. That's a big building. There's an auditorium, absolutely everything in it. It was a stone around our neck, because it would be on our property records. What the hell do you do with an operations building in Green Bank? The only thing you do is operate a VLBI antenna for which you need one operator.

Dick

But your position is that the observatory doesn't need to be there for VLBI either, isn't that right?

Westerhout

That's right. We could do it much cheaper by doing it at Westford, for example, which is a Haystack antenna that is being used by NASA and by NOAA, but NASA and NOAA are dropping out there, and therefore that antenna is in difficulties. But we're basically forced to stay in Green Bank.

We have always made the very strong point that we need two antennas on the West Coast and two on the East Coast. The ones on the West Coast are, of course, rather far removed from each other, one in Hawaii and one in Alaska, but we need the long baseline. Therefore, if one thing drops out of the East Coast, we always have the other one to go. It's all part of this whole business of having two stations on each side, to have this long baseline to measure the rotation of the Earth, because the rotation of the Earth is the quantity that is the hardest to predict. That you have to measure, therefore, with the greatest accuracy. So, the security of always having antennae available.

So, you will ask what then happened when the antenna in Richmond, Florida, fell over through Hurricane Andrew. You certainly didn't have an antenna. Well, we engaged the people at Haystack, and, for free, people in Italy and Germany, who gave us generously of their time, but that is not something you can continue forever. At some point that is very nice to do, but -

Dick

That antenna is being rebuilt, right?

Westerhout

That antenna is now being rebuilt. Well, the bids are due as we speak.

Dick

But the money is there to do it.

Westerhout

The money is there to do it. So that antenna should be back in operation. There should be a new antenna, a brand-new antenna, in about a year from now. At that point, NAVNET will have three antennas that they own themselves--the one in Hawaii, the one in Green Bank, and the one in Florida. Three of the four antennas ain't bad in a period of only three years, to build up a thing with three $2.5 million antennas, to get that going.

Dick

What, by the way, was NRAO's position in all of this flap at Green Bank?

Westerhout

NRAO, of course, would like to have the clock. In principle, of course, I have to say that we would have liked the clock. It was the best clock system that we'd ever built--ten hydrogen masers, twenty cesiums, God knows what fancy equipment, several satellite relay antennas on the roof, so on and so forth. A really beautiful station. But it was supposed to be sitting there more or less with two technicians, one technician and an engineer, just keeping time. It was completely steered from Washington or Florida, depending on which one was dropped out. In other words, it was in addition to the master clock here or to the alternate master clock in Florida. It was not its own master clock, because if you have your own master clock, you need a whole staff of mathematicians and people who now are capable of distributing their time and measuring, which is a staff of six we have in Richmond, Florida.

Then there were arguments, of course. Well, if we must stay in Green Bank, let's close Richmond, Florida. We finally got that point across to the oceanography staff, and in particular to the oceanographer, that would hardly save any money, because that staff would then have to be moved to a place outside of Washington that is not Richmond--Green Bank, Flagstaff, whatever. Moving costs would be hundreds of thousands of dollars, and the salaries. What you would save would be the $60,000 or so that it cost to keep the physical plant operating. That, of course, was no argument, because it's a very important VLBI station. So that idea was canceled, thanks to the understanding by the oceanographer and the staff.

Dick

Let's branch out some from radio astronomy, then. Certainly one of the innovations during the time that you were here was radio astronomy. What other innovations stand out during your time here? The optical interferometer would certainly be one. How did that get started?

Westerhout

The optical interferometer business started around 1978, a year after I got here, basically.

Dick

That early?

Westerhout

Yes. In the sense that Shao, under the tutelage of -

Dick

This is Mike Shao at MIT [Massachusetts Institute of Technology]?

Westerhout

Mike Shao at MIT. But he was working for his Ph.D. under Staelin. Staelin was his thesis advisor at MIT, and they built a little prototype thing that could only look at the pole star and surroundings. They had a little bit of NSF support. At the same time, Townes was trying, using some of the mirrors in the top of the McMath solar telescope, to see what he could do in infrared interferometry. Both of those came to ONR [Office of Naval Research] to see whether there was any money there, and somehow or another we got ONR to put in about $100,000, $50,000 to each, for proof of concept principle, with the Naval Observatory the main organization that would eventually benefit from that, and therefore being asked to look after that. That went well, in the sense that Shao proved what he was doing.

Dick

That was in the late seventies?

Westerhout

'79, I think that was. They both got relatively good data. We then talked extensively, and NRL was very heavily in that, because Shao at one point moved to NRL after he got his Ph.D. That led to ONR providing a research option for interferometry, and the beneficiary of that research option was the Naval Observatory. The research option was to research the possibility of doing very high precision astrometry in the infrared and/or the optical. The money was divided equally, more or less. There was a third party involved--Curry at the University of Maryland. But after a year, he dropped out because he didn't produce. The other two went full speed ahead. The infrared went to the University of California at Berkeley. Townes also produced another 2.5 million for DARPA to build the big mirrors, because he had 1.5-meter mirrors. You needed that much, because he was using heterodyning, which means a very narrow bandwidth. So he needed a lot more signal.

Shao, Johnston, Simon, Mozurkewitz, all that crowd at NRL, together with Hershey and Kaplan at one point, started building an interferometer on Mount Wilson. The result of that is the Mount Wilson interferometer, of that $5 million over four years, basically. That went into that. We (USNO) started getting involved in that around 1982 or '83. Before that, I was the only person in the observatory who kept pushing this, because the Astrometry Department, Dr. Hughes, did not believe that this could provide anything better than what he already had.

Dick

With transit circles.

Westerhout

With transit circles. Because the transit circles, you know by now, are bothered by refraction, and all the irregularities in the refraction you make, you may be able to get a factor of 3 or 2 or 50 percent better if you really do a lot of finagling and so on, but not any more.

Dick

Hughes, of all people, knew the limits of the transit circle because of refraction.

Westerhout

Exactly.

Dick

What were his objections?

Westerhout

Because when you're working with an interferometer, you still look through the same atmosphere, and therefore there's no question that you get the same effects. I had worked in the radio field with interferometers, and I knew that you are looking at delays, rather than directions, and the delays are equally affected by the refraction in the two arms of the interferometer. For a year and a half, on and off, I tried to explain that to Dr. Hughes, but I am not a mathematician. Somehow or other, he didn't buy it. Then I finally got Mike Shao in Hughes' office. We closed the door, and three hours later, the whole blackboard was full of tensors, because Shao and Hughes both spoke mathematics. Shao completely convinced him. From there on, it was Hughes' project, which it was always supposed to be, but it then became the interferometer project.

Dick

You're saying that was in '82.

Westerhout

That was about '82 or '83. '83, probably. Then John Pohlman got involved in building the huts for the siderostats and other parts for the siderostats and parts for the delay line. It was really getting more and more a joint project. MIT was still involved. Then Mike Shao moved from NRL to SAO, so SAO got involved. They got a little bit of money out of it. That telescope is still working, that interferometer on Mount Wilson.

Dick

The sign up there says NRL-MIT interferometer. It doesn't say USNO, I noticed a few weeks ago when I was up there.

Westerhout

You were up there? I see. NRL-MIT. Well, that is right, because it was not a penny of USNO money in it, except for the salaries occasionally. Pohlman worked on that. He was paid by NRL. There was indeed no USNO money, that's quite correct.

Dick

But some of the observers, at least, were USNO staff.

Westerhout

Yes.

Dick

That now is not operating, is that correct?

Westerhout

It is still operating, but we expect that this is pretty well the last year. We did five years' worth of astrometry and, of course, a very large amount of stellar diameters and double stars and so on. The thing has been very productive, with lots of papers on 1 milliarcsecond double star orbits, where, really, 1 milliarc second separation of a double star, that's a double star that goes around in a year. Therefore, you can get one or more orbits of the damn thing, and they look fantastic. They also did some wide pairs that they could compare with speckle things, and always the interferometer was a hell of a lot better. Now, that's relative astrometry, as is the diameter measurements of stars. What we're after is absolute astrometry.

So what we did was measure about 12 or 20 FK5 stars over and over and over again, five years running. Combining those data then, and see how they fluctuate from one year to the next, and you come to the conclusion that the accuracy is of the order of 100ths of an arcsecond--the absolute accuracy. One of the things we did right in the beginning, and we have kept doing that, is plotting the difference of the FK5 position against the interferometer position. You put the theoretical FK5 error box on that diagram, you find all these stars sort of spread to two or three sigma around the error box, which you expect, but, of course, that immediately tells you that when you take an FK5 star and you say, "That star is good to .05 of an arcsecond," that ain't so. That .05 arcsecond RMS and the outliers run into the 2, 3, 4/10s of an arcsecond. But the error bars themselves and then putting them on top of each other from year to year really gave a terrifically good thing.

That project was over in '88--four years. In fact, we got some more money in '88. It was really five years. Originally it was four years.

At that point, the next question is, what are you going to do now? We felt that it had proven completely that you could do astrometry with an optical interferometer. Townes was not doing astrometry; he was doing imaging. He never came up with good astrometric results. So he's sort of now off on the side and doing his own things. He probably still gets a little bit of Navy money, but it doesn't go through NRL anymore. And I presume he gets NSF money.

We then embarked on two projects--the USNO optical interferometer, astrometric interferometer, and the NRL big optical array. The NRL big optical array was going to have arms about a kilometer long, with many stations that you could move your siderostats around and synthesize the aperture. Ours was always 20-meter baselines, four siderostats so that you could do phase closure, you can do two baselines at the same time, and all of that. That was designed, worked out.

We looked for a site. It finally became clear that the real topnotch atmospheric circumstances that we wanted for the astrometric interferometer were only to be had on tops of mountains. We can't put a 1 kilometer array on the top of a mountain. So for a while, the project split. It then became clear that the amount of money that the optical interferometer was going to cost was far more than we had available. The MILCON money that we thought we had kept for FY '93 to build on that mountaintop in California was canceled. So we suddenly lost $3.5 million there.

At that point, we completely revamped the project, and NRL and USNO went jointly and used the next best site that we had been able to find, which is one on Anderson Mesa in Flagstaff, belonging to the Lowell Observatory. So we got the Lowell Observatory involved. Everybody is traveling there. We just sent two flatbed trucks and a trailer with all the equipment over there. The buildings are now finished. Lowell built all the buildings for us. Installation is now taking place as we speak. So that's sort of the history of that part. But now it's a joint project. Our four astrometric telescopes are, for the time being, part of the big optical array as well.

Dick

And NRL is still more interested in imaging than astrometry?

Westerhout

That's where NRL gets their money. They get their money out of the development of sensors. That's where that comes from--proving that you can do imaging with an interferometer. Part of the defense for that is, of course, based on the fact that in principle, you can use such a thing also to look down, but you have to first understand the whole technology of astrometry. I mean, you can do 1 cm on the surface of the Earth if you look down from 200 kilometers, and see the insignia on the Russian general's breast pocket. But, of course, the Russians have dropped out, so you never know what is going to happen. All right, let's say the insignia on Saddam Hussein's breast pocket.

So that's very much imaging, and that array will consist of six moveable telescopes plus the four Naval Observatory telescopes, but has been built such that eventually we can split off completely if there's more money available, and NRL gets money to build their own delay lines and so on. The whole project is aimed at astrometry first and imaging second. Therefore, the delay lines that have been built are all the astrometry delay lines. The arms of the imaging array are only 100 meters long or so, so they can also use those short delay lines. The longer the arms, the longer the delay lines.

Dick

What will the observing program be for the Naval Observatory? What do we hope to accomplish?

Westerhout

We aim is that by mid-1996, we will have a catalog of 1,000 stars--bright stars.

Dick

To what accuracy of position?

Westerhout

To a precision of 5 milliarc seconds or better. In other words, within two years, that will be better than HIPPARCOS, but HIPPARCOS is, of course, a project that does 100,000 stars in one fell swoop. We can't do that with that speed. So the aim is an astrometric catalog to begin with, and then expanding that. The other aim is to convert it to infrared relatively early in the game.

Dick

How does this affect the transit circle? Will the interferometer supersede the transit circle, do you think?

Westerhout

The interferometer will supersede the transit circle in terms of providing absolute positions of stars. The transit circle at the moment looks like it is still needed for the planetary system. People are worried to throw out the transit circle when there is not yet a good new system available. You need overlap. Eventually, of course, you can do the planets with respect to a star background with an astrograph if you know the star background well. Do it relative and assume that the star background is tied to the extragalactic sources.

Now comes the whole story about tying the VLBI reference frame based on 400 quasars, which is another joint NRL-USNO project. Determining that reference frame is a joint USNO-NRL project.

Dick

Why do you need a dynamical reference frame based on the solar system if you have an extragalactic reference frame?

Westerhout

Because if you don't know how to tie those two together, how do you tie the proper motion system? You have to somehow or another convert the whole system into an extragalactic system, get the rotation out, because obviously the dynamical system is rotating also. Everybody always says it isn't, but at the milliarc second level, it is. So that is where the observation of planets still comes in.

The question, however, is, how much longer? Is it one or two years? Is it five years? Looks like what we will be doing here is we will continue work with the 6-inch. But Steve Gauss is already coming back from the idea that we should now start equipping the 6-inch with a CCD camera, for the simple reason that then it's no longer the same instrument.

Dick

That we should not do that?

Westerhout

Not do that. Because the main reason why we continue with the 6-inch is to continue precisely the same series and eventually end up into a combination of the two, and at that point terminate it. Whether that eventually is what will happen is, of course, another question.

Dick

By the way, you mentioned phase closure, which I keep hearing about, but I don't understand what it is. Can you explain that?

Westerhout

Phase closure is when you observe on one baseline, you determine a certain phase of the fringes. There are fringes sitting in certain location. How accurate you know that phase, of course, determines how accurately you know the position of the object. Now you do it on the second baseline. You get another phase. Now you do it on the third baseline. One, two; two, three; three, one. Now the three phases of those signals measured at the same time have to add up to 360 degrees. If they don't, then something is wrong, and you have to find out what's wrong and finally get to that point.

That, however, does not make it absolute yet. That now is relative. Phase closure is absolutely necessary if you use the precise phases to do imaging. You're not interested where the object is, because you can make a zero-point error in all three of those phases. For the astrometrist, that would be disastrous, but for the imager, hey, that doesn't even matter. So that's where in our case, then, all the metrology comes in, tying the whole system to bedrock. So phase closure is basically making sure that you have at least three stations measuring simultaneously so that the phases of the fringes add up to zero or to 360 or whatever.

Dick

That's a difficult thing to achieve, I take it.

Westerhout

That is a difficult thing, but it is now used regularly in VLBI. They have managed to do that in VLBI, and that is why VLBI now produces maps of extragalactic radio sources with -

End Tape 3 Side A

Begin Tape 3 Side B

Dick

We were talking about astrometry. As long ago at least as 1980, I remember we were talking about an astrometric satellite, which never came about for the Naval Observatory. The Europeans took off in that area. What's the story there? Why did the Europeans get ahead of the Americans in that effort?

Westerhout

Basically because they were ahead in the effort from the start. The French took that up initially in 1969, and then again, when that first attempt completely failed, in the late seventies--'77 or so. Then they turned the whole thing into an instrument that didn't particularly do absolute positions, but parallaxes. That got the entire astrometric community riled up about what fantastic parallax you can get, what you can do for all these stars that we have always wanted to have parallaxes for.

In fact, I organized, on behalf of HIPPARCOS, a one-day session at the 1979 Montreal meeting, where I invited a whole lot of people from different walks of astrophysics to give talks about what HIPPARCOS will do for them. It was a very successful thing to make propaganda for the project, basically. So they were ahead in that sense that they had a big group of people in France, Sweden, and England, basically, working on that telescope for a long time. This was a major effort. Don't forget it was a major effort.

Dick

But we had a plan then to get into astrometric satellite business. How far did that plan go, and where did it go?

Westerhout

The idea of us getting into the astrometric satellite business didn't start until the early eighties, I would think.

Dick

1980, for our sesquicentennial publication, there's a drawing of an astrometric satellite and even a time line.

Westerhout

Very good! Very good. Okay. Must have been '79, then.

Dick

Was that put in for funding?

Westerhout

That was put in for funding. That was defended everywhere, but everybody said that's far in the future. That's typically 6.1 money, so we talked to the 6.1 community, the ONR folks, and they said that's an NRL project. So Ken Johnston and I got together and we talked about its possibilities, and NRL is now picking it up and have, in fact, put plans in for funding of some sort of a satellite.

Dick

They call it the Newcomb astrometric satellite. At least that's the title that was announced for the talk that Simon is giving next week.

Westerhout

Oh, really? Hey, I hadn't heard that one yet. That's good. That's good. And the NRL side of the project has been sort of under way for the last two years. Our plan was more to try and get money for, let's say, a Phase A study, and we never even got that. We simply did not have the manpower to put into doing any intensive study of an astrometric satellite, what it could do.

So when we made plans for an astrometric satellite, we basically wrote down in a number of pages, and George Kaplan was very instrumental in that--what such a satellite would do, what it would be good for, the reasons why the DoD [Department of Defense] needed an astrometric satellite, and all those good words, and how it would overlap with HIPPARCOS. That was an important aspect, because by the time we started talking about that, HIPPARCOS was scheduled for launch two years later, and we would come in long after HIPPARCOS was down again. Of course, it's now still the case, because we are nowhere and HIPPARCOS was delayed by several years. So in the end, that isn't all that bad, if we can indeed get a satellite up by the year 2000. And that was our plan a number of years ago when we started reviving that again, we said the year 2000, and I'm still saying that in every funding discussion, that somehow or another you must work towards that. However, I can say NRL is now heavily involved in that, to do the initial 6.1 work.

Dick

So they have funding for the study.

Westerhout

They simply took that out of their own pocket. So it's still really in the concept stage, but there are some good ideas there. One of the things I don't like is that they are immediately saying a hundredth of a milliarc second or something like that, and that is not what the Naval Observatory needs. I keep asking the NRL people why would the Navy need that.

Dick

A hundredth of a milliarcsecond?

Westerhout

Yes. In principle, when you have a good interferometer in space with no other things, you might be able to do that. I don't quite see it. But it would, of course, allow you to do a lot of things on gravitational waves, on determination of gravitational deflection of light by massive bodies, second-order effect, the Sunyaer effect, all sorts of effects that could be measured when you have that precision. Of course, in addition, people are talking about parallaxes. You could do a parallax. You could get a very good handle on the parallax of the Magellanic clouds, provided that the thing could look faint enough when we observe stars of that kind. So there's a lot of astronomical value in going to these very great accuracies, leave alone finding of planets around stars. You can start seeing wiggles in stars of tens or hundreds of a milliarc second. You suddenly have a much bigger chance of finding something.

Dick

So I don't see what your objection to it is. You said you didn't like it.

Westerhout

I don't like it because I don't see how I would defend it to the DoD. I cannot personally find specific DoD reasons to build that. The 6.1 guys are not all that worried about that, interestingly enough. But I at the Naval Observatory should be worried about that, because we're an operational organization, and we have to therefore make it clear why we need those kind of accuracies. That I still find a hard one. A satellite that would do a milliarc second on large numbers of objects, that's a different story. That one I could defend instantly, there are so many things there we can do, but tenths or even hundredths for a few stars, I have difficulty with that.

So some of the things that we were thinking of when we originally talked about an astrometric satellite was a satellite with a wide field camera on board, either working in the HIPPARCOS-type mode or simply in the stare mode. But we did not think in terms of the way HIPPARCOS does it by looking at stars 56 degrees apart simultaneously, so that you always have the angle reference. We were thinking--and that's what the NRL was doing, too--of an interferometer in addition to a camera, with the interferometer determining precisely where you are for one or two stars in the field, and then you take a CCD picture of the field, and you have the field completely oriented by means of the interferometer. But, of course, that's two instruments in one, rather than one. The question is, how big is it going to get if you have to do that sort of thing?

So that now is basically in the talking stages, and that's why next week's talk will be very interesting to hear. We already had one by Simon about a month and a half ago for a small group of people, but now it's for a bigger group.

Dick

How about HIPPARCOS II? Is that going to be funded, in your opinion?

Westerhout

I doubt it. I doubt it, because they have to first compete for design money among twelve groups, and then out of those twelve, four will be chosen for design. Which of those four will then finally fly? So the chance of that going off in the light of what other priorities there are, I mean, HIPPARCOS was fighting against the X-ray satellites and so on. There was a lot of infighting going on there among the European people in finally getting HIPPARCOS going. Parallaxes was one of the main things that won them out.

Dick

So the argument now will probably be that you had your satellite; now let's do something else.

Westerhout

I suspect that that might well be the case. But you cannot, with the philosophy, stop designing. Somebody may be very enlightened to say, "Hey, maybe we should do that."

Dick

If the NRL astrometric satellite goes forth, what role would the Naval Observatory play?

Westerhout

I would hope that it's going to be a joint project. So far it isn't yet, because again we simply don't have the personnel. I have not been able to assign personnel to such a project and say, "From now on, you're going to design satellites," because everybody is needed somewhere else.

Dick

And we don't have much expertise in designing satellites anyway.

Westerhout

That's all right. That's all right. That means that you simply send the person to NRL or APL or some place like that for half a year or a year, let them work with the guys there to get that expertise. We haven't done that.

Dick

Let's move on to another area, talking about innovations. You mentioned the CCD effort in Flagstaff. I guess this has been one of your pushes, was to use Flagstaff as an R&D place for light detectors, is that a fair statement?

Westerhout

That's a fair statement.

Dick

How did all that develop?

Westerhout

When I came to the observatory, CCDs were already in use. Among other things, there was the ground-based CCD that was to be used for the space telescope and which Seidelmann was involved, and Kitt Peak was using some CCDs. People were experimenting here and there with CCDs. It was clear that somehow or another, the electronic camera was around the door. I arrived in Flagstaff in late '77, and found out that there was a photographic parallax program. We had been doing photometry and were doing some spectroscopy and doing double stars and so on. There was quite a bit of photometry knowledge. But photographic astronomy?

So after I looked it over for a while, I sat them all down and, among other things, I declared that in ten years, there shall be not a single photograph made in Flagstaff anymore. That shook the hell out of Ables and Conard Dahn and these people who were all completely wedded to the photographic plate, with membership in Kodak committee, so on and so forth. Well, it wasn't quite ten years. We are still using some photographic plates, but it looks like within a year the photographic plate business will finally be ended. Seidelmann then imported the CCD chip that was going around as part of the Space Telescope preliminary work, and so he imported that to Flagstaff. He used it, and the Flagstaff guys sort of looked on a little bit.

Dick

For his planetary satellite work.

Westerhout

For his planetary satellite work, yes, and partly also these were test observations to find out how well these things work and so on. That's why it was provided free, with all the kit and caboodle attached to it. Then the Flagstaff guys started trying it out a little bit. In the meantime, Monet was working with a CCD at Kitt Peak, and invited Dahn over. So Dahn and Monet did some CCD observations of parallaxes and so on in Kitt Peak, and Dahn came back and said, "Hey, that looks like it works."

Then about a year and a half later, we hired Monet. Of course, then it really took off, although when Monet came, we were already into using one particular chip, which by then the observatory had inherited. It didn't travel around anymore; it stayed at the observatory. In fact, it's still there--the banana chip. It's called the banana chip because the imperfection it has on it looks like a banana. But it's already referred to as the banana chip. It's still there and it's still being used.

The aim of all of that was, of course, to (A) do things faster, because there's a lot more quantum efficiency; (B) get to fainter objects; and (C) avoid the use of a plate measuring machine. And then, of course, the fourth thing is, what are you going to do? That's, of course, always the subject of discussion as to what should be done with a camera--any camera. What do you need in the way of star catalogs? How deep do you need to go? You need the tenth magnitude catalog? Twelfth? Fifteen, sixteen, seventeen, eighteen? A lot of people keep saying seventeen. Now, that comes, of course, partly from the fact that seventeen can be reached with reasonably modest telescopes, although Christy was in the process of designing a telescope with 1.5-meter aperture to get to seventeenth magnitude photographically.

Photography is out, basically. Kodak is beginning to give up. They're more or less the last ones in this country that are still doing emulsions. Making photographic emulsions is an art that goes from one person to another. When these persons finally die off or go elsewhere, it simply disappears. I mean, some of these emulsions contain the innards of peculiar animals, like loggerhead turtles or stuff like that. I heard the other day, but I forgot what animal it was whose livers were used to make the emulsion for the 103AO plates or something like that. Absolutely amazing. Apparently this art is beginning to disappear. Moreover, it doesn't pay anymore. It's really a service of Kodak to the scientific community, and in this day and age, you can't afford it. So everybody is looking towards the demise of photographic plate and switching rapidly as they can do using CCDs and other devices.

Dick

In the case of the 61-inch for the parallax, how much increase in accuracy have you got with the CCD versus the photographic plate?

Westerhout

By a factor of three. We could go to fainter objects, too. The typical result of the last twenty years of USNO parallax program with photographic plates, which was analyzed by Harrington and company a number of years ago in the Astronomical Journal, comes out to about 3.5 milliarc seconds typical accuracy for a parallax. A few parallaxes like Van Biesbroeck 8 and so on, this one that was observed for a long time because it was assumed to have a planet, with many plates, you simply beat down the statistics, and they come close to 1 milliarc second, but that's hundreds of plates. Typically a parallax program is 20 or 30 plates. With the CCD, you get 3 milliarc seconds in a single exposure of two or three minutes, and therefore, since so far it still looks like it goes with the square root of N, and the exposures are short so you can take as many as you like, 1 milliarc second parallaxes are completely routine at this point. In effect, you see that in the latest publication.

Dick

You said that you couldn't justify to the Navy a hundredth of a milliarc second accuracy. How do you sell parallaxes to the Navy?

Westerhout

Well, in the first place, I sell parallaxes by making the statement that if you have a requirement of better than a tenth of an arcsecond, you'd better know the parallaxes, because the nearby stars go around with that sort of thing. So that's immediately accepted. I forget to mention that we are usually determining 15th magnitude star parallaxes.

However, now comes the crown on that work. We have, in the course of time, over the last eight years, seven years, developed filters with an Inconel spot depositor. Inconel is a highly absorbent material, on the filter, so that you can put a bright star shining through that spot, and the faint stars are the reference stars, because bright stars, of course, it's very hard to observe parallaxes on, because there's no comparison star around.

That is harder than you think, because you're talking astrometry, so you're talking not wanting the image center of an object to move much more than a tenth of a micron. What does that spot do if it is not completely 100 percent homogeneous? Suppose there's one tiny little hole in it. Now put a fifth magnitude star on it. That tiny little hole lets perhaps 1,000th of the light of the star through, but that moves the image of the stars. But we succeeded. It works fine. We mastered that. We had industry master that technology.

Dick

But that's a technique with the photographic plate.

Westerhout

That's a technique with a photographic plate. So that has resulted in the last catalog having quite a number of bright stars in it, which everybody is cheering about, and HIPPARCOS people are already using to test their preliminary parallaxes. They now show graphs of Naval Observatory parallax versus HIPPARCOS parallax, with a straight line at the 45-degree angle all the way through it. The spread is much smaller with respect to the parallax when you get to large parallaxes, of course. They're both approximately the same magnitudes, so that's interesting. A lot of bright star parallaxes will be solved by HIPPARCOS.

Dick

We were talking about how you sell the parallax to the Navy.

Westerhout

So I was going already to the bright stars, and I was going to say that obviously we will have to look at the brighter stars, but in order to understand the technology, we have to concentrate on the faint stars for a long time, develop all the software, the hardware, and so on. We started doing that with the photographic plates.

Then we started developing a mosaic of CCDs, which is a 6.2 project that failed, which the 6.2 people like. It is high risk; 6.2 is high risk. It failed because the CCDs that we chose were no good. This was a project which was contracted to Auburn University, and Auburn University is not to be blamed. We selected the CCDs. These were five CCDs that were the long integrating CCDs where you can look at a lot of stars, and one that could be read out fifty times per second. There, of course, you have your problem solved of having a bright star that would completely obliterate everything else. That star doesn't look bright on that CCD.

The trick is, however, you want to do that to a tenth of a micron, and therefore that whole system of CCDs, that mosaic, has to remain stable over a long period of time. So what was done was bonding those CCDs to a substrate made of the same silicon material, so it's basically one wafer with CCDs on it, which will act completely uniform.

Now that we know what went wrong there, we are now developing that technology ourselves. We got the right kind of guy to do that. Fred Harris has been with us for two years, three years, and we're working on that ourselves. The next step will probably not be small, but rather some bigger CCDs together in a mosaic form.

Now you understand how we slowly but surely have begun to understand CCDs, right? We're developing light detectors. At some point those light detectors have to come into use. Well, they came into use in the 8-inch, in an experimental program where we are comparing FK5 stars and radio sources with the same instrument, instead of making three steps, as has been done up until now. Secondly, it will come to fruition in the electronic astrograph, which we now have been talking about for the last four or five years, six years, maybe, but is now beginning to get somewhere. There is a team working on the design of that. Flagstaff has provided a big report of what back end you would put in there, and the typical back end will be eight CCDs, 2048 by 2048 each. That would then cover about one and a half by three-quarters of a degree. Size of the telescope would be about 75 cm diameter.

There are still a lot of things to be discussed, like what do you do with the bright stars? Because, after all, you want to attach this faint star catalog to bright stars, of course, because that's where your reference frame is. Can you observe them simultaneously? Would you put an objective grating in there so that the bright stars have images in them that...all sorts of possibilities. Would you put a beam splitter in? Take two pictures simultaneously, two different CCD arrays. And so on. So there is a lot of discussion, but I would hope that discussion to sort of start gelling in the course of the next half year, because in FY '94 there is money available to start building such a telescope.

Dick

And the goal is what, a sky survey?

Westerhout

Sky survey, talking about two, two and a half years in the North, and two, two and a half years in the South.

Dick

What site do we have in the South?

Westerhout

We don't have one, but we'll make a portable telescope. We'll take it to Cerro Tololo. Fred Vrba already several years ago, at my behest, on his various visits to the South, went and spied around and talked to people, and there are several empty domes available. That should not be a particular problem.

Dick

And in the Northern Hemisphere, would the site be here or Flagstaff?

Westerhout

Flagstaff. That's sort of self-evident. You wouldn't want to put it anywhere else. If we had an interferometer site in California, we would probably put it there, but we haven't got that.

Dick

Shifting gears here now, you've mentioned 6.1, 6.2. I want to talk about the budget. First of all, what the budget was when you came versus what it is right now. How much has the budget increased?

Westerhout

By a factor of four.

Dick

So our current budget is?

Westerhout

About $20 million.

Dick

So it increased from 5 to 20.

Westerhout

Yes, approximately.

Dick

How about the sources of the funding? Are they basically the same now as they were then?

Westerhout

They're basically the same now as they were then, except that we've got a little bit more what they call reimbursable funding, funding from other organizations, that helps us. At the time we started, we had the regular operating budget and we had a small--about $500,000 a year--R&D budget in the 6.2 arena, that basically kept Flagstaff going. That R&D budget slowly went up to 1,500, 000 what it is now, and still basically supports Flagstaff.

Dick

Can you explain the 6.1, 6.2, and 6.3?

Westerhout

Yes. 6.1 is called basic research. That's funded by ONR. 6.2 is called exploratory development--high risk development of things that might or might not pan out. Design a new paint for the hull of a submarine. Well, will it work or will it not work? You don't know that until you've tested it, and maybe it's the wrong kind of paint so you don't do it. To give you a ludicrous example, but all that sort of stuff is paid from 6.2 money. Then 6.3(A) is called advanced development. 6.3(B) is called prototyping. We have 6.3(B) money from the oceanographer. At 6.3(A) level and below, it all comes from the R&D community. In other words, it comes from the under secretary for R&D, which is mainly ONR and some of the labs. The labs are all funded from those things.

Dick

Can you give me specific examples of programs here that are funded by each of those, just to put it in concrete terms?

Westerhout

The interferometer is 6.3 and OPN, because we felt that...well, in the first place, there was not nearly enough money in 6.3 to build a whole interferometer, and we felt that in the 6.1 program that built the interferometer at Mount Wilson, we had learned a lot and we could specify certain things. There was no question we could specify, without any trouble, the delay lines. So we simply contracted the delay lines out to JPL. We bought delay lines off the shelf, if it were. That's OPN money. That's no risk.

Dick

OPN meaning -

Westerhout

Other Procurement, Navy. On the other hand, a lot of the work that is done by designing new things and experimenting with new things is really still high risk, so that was paid by 6.3. Ken Johnston still puts a lot of 6.2 money in as well. We didn't have 6.2 money to put into it. So that is a project that is already on the way of becoming a finished thing, whereas Flagstaff, we feel, is never at the point where they produce finished things. At that point, it transfers from 6.2 to 6.3, and when it's 6.3, it sort of dies on the vine or gets used for certain experimental work to show that you can do this or that. We feel that the double star work, for example, is really R&D. We feel that the double star work is not operations.

Dick

Because of -

Westerhout

Because of the fact that it's in support of some of the other things. That's historical. You could just as well call the double star work completely operational, but, therefore, when you buy a speckle camera and so on... Of course, I'm contradicting myself. The interferometer, I said you build the delay lines because you know how to build them. You build a speckle camera because other people have built them, so you buy it off the shelf. But we've defended that as being real R&D money.

Dick

So this 6.1, 6.2, 6.3 system was here when you came, and still here.

Westerhout

Right. But we get no 6.1 money. 6.1 money is basically reserved for the pure researches. A lot of that goes into universities. If we really have a research project, we can apply for money to ONR just as though we were a university, and we may or may not get a grant.

With 6.2, we submit a proposal every year telling them what we do now, what we're going to do in the next years and so on, and on the basis of that, they determine how much you're going to get the years thereafter. That's gone up for a while. Now it's on the way down.

In the 6.3(B) arena, you again submit a proposal, but much less detailed, to the oceanographer of the Navy, and he decides how to distribute his money among R&D work and operational work. He can do that as part of the POM process. The POM process basically is a process whereby the oceanographer, the user of Navy money, decides how we will distribute his money.

Dick

POM stands for -

Westerhout

Program Objectives Memorandum. The POM exercise is an exercise to write a program objectives memorandum. That memorandum contains the objectives that you want to do, why you want to do them, what impact it will have on the Navy, on the war-fighting capability, and what it costs.

Dick

How often is this done?

Westerhout

Every two years. So we are starting off next month and the months thereafter in POM '96. In POM '94, which Captain Donat ran, he was extremely successful. That's why our budget in FY '94 stands to go up by $2 million, which is absolutely unheard of in everybody else's. Keep your fingers crossed. We may not hang on to it. But the chance that we go down the same as everybody else is nil.

Dick

POM '96 may be another story, though.

Westerhout

POM '96 may be another story, but the important thing is that you have the oceanographer and his staff understanding what you do. Therefore, I was one of the few people who cheered when the oceanographer decided to come and move into the Naval Observatory. This must have been in 1980 or '81. They all said, "Westerhout must have flipped." Everybody hated the idea that they had to move out of their offices when all these people came.

Dick

I was one of them. [Laughter]

Westerhout

I'm sure you were. But it has done us nothing but good, because that is the person who goes up to the Hill and up to the secretary of defense and defends your program. Instead of having him sitting in the basement of the Pentagon where you never see him, he sits here right next to us.

Dick

Was there a change, then, when they came here? Were they still our resource sponsor before they came here?

Westerhout

They were, sort of, but nobody really knew how that worked, which was another funny thing. Because Captain Smith, for example, pushed very hard--he came here a year before I did--to get the budget up, because we were in a real slump. But Captain Smith hated the oceanographers, and I have the feeling that Captain Smith didn't really know to whom his justifications went, because he sent them directly to OP-09B, which is the vice chief of naval operations. The vice chief presumably then sent it on to the oceanographer, but the oceanographer in those days was not really in the picture, although he was quite clearly the fellow who eventually had to put it up. It was a peculiar situation.

Dick

There is a difference between line officers and restricted line. Smith, for example, would have been a line officer.

Westerhout

He was a line officer.

Dick

And the oceanographers were restricted.

Westerhout

Right.

Dick

So that may have something to do with the antagonism.

Westerhout

Absolutely. Then when Captain Smith found out who his relief was going to be, which was Captain Vohden, he comes into my office, he closes the door, he says, "I have bad news. I just found out who my relief is. It's a flyer." No matter who he was, he was a flyer, and that meant the end of the observatory, as far as he was concerned, because he was a submariner. He always carefully avoided to mention that he was a diesel submariner. That's another detail. He would always talk about "my friend [Hyman] Rickover." He had taken a course at one point from his friend Rickover. He wasn't a friend at all, of course, but that then made it clear that of course he knew everything about nuclear subs, but he never was on a nuclear sub in his life. He was one of the last diesel guys, basically. But we were talking about the budget.

Dick

Which you say has increased four times over the time that you've been here.

Westerhout

Basically, yes.

Dick

How about the relative proportions of the 6.2, 6.3? Has that changed?

Westerhout

There was practically no 6.3 at all. There was $100,000 to $200,000 to support the master clock a little bit.

Dick

And now?

Westerhout

There's now $1.5 million, but that's going on its way down. In fact, at one point we were at $2.5 million, but then somebody got wise and said, "They have too much money." Somehow or other, we couldn't get it. We are getting better at that, and I think we're more and more getting our foot on the ground, as far as making it clear to people what it is that we're doing and why it's so important to provide that continuity.

Dick

How about the size of the staff when you came versus now? Has it gone up or down?

Westerhout

Up, up. There were 158 people when I came.

Dick

And we now have?

Westerhout

We now have 180.

Dick

And that's gone down because of the public works.

Westerhout

Exactly. At our peak we had 212 or 213 on board, and it's gone down some. It's gone down by 34 when public works left.

Dick

We're at the end of this again. There are lots of other things I want to talk about.

Westerhout

There's a lot of stuff we have to discuss. Reserve that for some time next week.

End Tape 3 Side B