Weber, Fritz

Weber, Fritz

University of Alabama in Huntsville Archives and Special Collections, Huntsville, Alabama

This material may be protected under U. S. Copyright Law (Title 17, U.S. Code) which governs the making of photocopies or reproductions of copyrighted materials. You may use the digitized material for private study, scholarship, or research. Though the University of Alabama in Huntsville Archives and Special Collections has physical ownership of the material in its collections, in some cases we may not own the copyright to the material. It is the patron's obligation to determine and satisfy copyright restrictions when publishing or otherwise distributing materials found in our collections.

.MP4

en

Interviews

Audio

ohc_stnv_000050_A

Bilstein, Roger

Weber, Fritz

[00:00:00] Roger Bilstein: Interview with Mr. Weber. I've been working on the chapter on the instrument unit, and I think that's what we'd like to concentrate our discussions about. I'll just kind of work through the structure of the instrument unit to start with or maybe we should start with some basic design concepts? Can you tell me how the instrument unit came about as a separate stage from the canisters in the S-I in the spider beam area and the interstage area until it was made into a separate stage for the S-IV and then the S-IV-B common IU?

[00:00:43] Fritz Weber: This is primarily a management decision, however, it has also considerable technical points involved. As you know, we had the three different Saturns: Saturn I, Saturn IB, and Saturn V. Of course, this is many years ago, and my memory may not get everything in the proper sequence. We started out with the canister type as you know, and I don't think we need to go into this one. One of the main reasons for this was that we could use designs of hardware and component which did not need to work in vacuum or in strange environments. We protected those by putting them into canisters and plug the canisters into the instrument unit. However, when we did find that we needed to design and develop most of the components directly for Saturn use, we applied the environmental requirements right to the component design, and therefore, did not need the canister anymore, which would make assembly and test of the instrument unit somewhat easier.

[00:01:59] FW: Did I give you a good enough point there? The separation of instrument unit from the S-IV-B stage was primarily a management decision. Each stage was only working for a part of the flight, yeah? However, the instrument unit is required for the whole flight. If we would have given the instrument unit to the S-IV-B stage contractor, we would have assigned with that IU assignment also the coordination and integration of the stages, which according to headquarters decision should be a center function and not a contractor function at that time, yeah?

[00:02:43] RB: So did the center coordinate these and then just let IBM do the fabrication? [inaudible…interfaces?]

[00:02:49] FW: Yeah. Okay, right, I see your question. You wanted to know how Marshall and IBM shared responsibility. It is kind of a development. The development of the instrument unit was responsibility of Marshall, the production was responsibility of IBM. So with this we integrated the design during design period, yeah? We built the first instrument units right here in our own shop as you know. We developed at the same time the IBM capability here in Huntsville to take over this assignment and start producing and assembling instrument units.

[00:03:32] RB: How did you go about developing this capability with IBM? Did you give them personnel to help them?

[00:03:37] No, we did not give them personnel. We gave them a contract to develop such a capability. This was a specific contract.

[00:03:47] RB: This is before the NASA 14,000? Before that contract?

[00:03:52] FW: Yes, yes, yes. If you want to, let me pull out a little folder. This folder from which I take the information here is a personal folder which has to do with the first IBM contract to develop this capability. This specific item is the evaluation of the performance of IBM. It was the first contract at all which had an incentive award involved. Now if you have specific questions I think I can, in reference to that folder, find out what you want to know.

[00:04:32] RB: Well, I am interested in the reason that IBM was selected as the contractor and why they were moved here to Huntsville to undertake this operation. I know they started doing the computer, the launch vehicle data adapter and the launch vehicle computer. That was their first contract.

[00:04:51] FW: Yes, in fact they did. That first contract, the development of the southern computer, southern flight computer, we should say. By this effort, IBM automatically got involved very strong in—how shall I say?—in the overall vehicle aspects and gave them a very good advantage over any other contractor to get involved in the instrument unit, which again involves all aspects of the vehicle and its operation. Now, the actual reasoning I could not tell you, I do not know how the selection and when it was made, really. This was done pretty much on a higher management level. The one primarily involved was Mr. Weidner's, or Dr. Weidner's deputy, Jerry McCall. He had spearheaded that. He's now with IBM and I…maybe you can give…he's still in Huntsville, maybe you can give him a call and find out. He may be able to talk with you I do not know.

[00:05:58] RB: Wasn't he with IBM before? Before the contract and then came with Marshall?

[00:05:59] He was…Jerry McCall was at that time with...I do not know where he came from. I must admit, no.

[00:06:13.] RB: And he was Weidner's deputy?

[00:06:16] FW: Weidner's deputy, yes. And on a special assignment, handled these items, yeah? I was more involved to direct the first contract. Now, we had made a decision then, here within Marshall, that the development of the instrument unit would be an in-house project. The manufacturing, as I said before, IBM. It would be a production effort, would be managed by what we call now P.M., program management. At that time, we called it I.O., industrial operations.

[00:06:56] FW: So, the first contract, which I referred to, to build up the capability during the development of the instrument unit in-house, was my responsibility. I had the assignment to coordinate in this all Marshall inputs. When the industrial contract started, P.M. or I.O., at that time, took over and managed the production contract. It would be good if you... I try to answer your questions, but I wait then and see which direction you want to follow. So, maybe that's it.

[00:07:34] RB: Well, did Von Braun play any part that you know of with top management of IBM, trying to get IBM into Huntsville, or to support the space program in general?

[00:07:46] FW: You know, I really do not know. I must admit, I was not involved in these type of operations at all, on these type of discussions. My role, as I say, was restricted to coordinate the development of the instrument unit, and see that IBM, during this period of time, supported us in that effort and same time build up their capability. So, I can tell you quite a bit of that period of time and the aspects there, how IBM started, and how they built it up, but I cannot tell you very much about the high-level decisions, management decisions, and why IBM was selected, and so, yeah…I do not know. I would tell you, but I don't know.

[00:08:29] RB: Oh, yeah, yeah. Maybe we can switch over and talk about some of the technical aspects of the IU for a while. We have first a structure which is built in three parts of this bonded honeycomb. Why was bonded honeycomb selected rather than just skin and stringer construction? That was a decision that was made here at Marshall.

[00:08:53] FW: Yes, decision was made primarily on what we call a trade-off study. We calculated, and we made preliminary design studies for either approach, and we found that the honeycomb was lower in weight. Weight is our problem because the IU weight is directly one-to-one to payload weight for most missions of the Saturn. It is based strictly on technical features, weight, strength, because it's a carrying member, the shell of the IU. Do you have one of these little brochures on the IU, which I..the yellow, the yellow-red one, the old one, we…it has a little description of the subsystems.

[00:09:40] RB: I've seen some things. I'll talk about that later. I might, well, maybe ask you to see the booklet when we're through here. [I’d like to see that?] Then the system was built into three sections with those splices, and this was for transportability, but it was only to transport the sections, wasn't it? There was no thought of ever disassembling the IU as a get around?

[00:10:03] FW: Yes, it was. It was. This was a very—how shall I say?—very unwanted situation. At that time, we did not have a good transportation means to the Cape. We could, we know we could use a barge, yeah? But the environment on the barge is not very favorable. We would have to build big protection containers or so, which is very bulky and expensive and time-consuming. We did not know then that the so-called guppy, or pregnant guppy, as it was called, this huge airplane would be available. So the sections of the IU, in the beginning, were thought we would disassemble the complete IU, disassemble it into three sections and transport those with different means. But later on, when the guppy became available, we kind of lost interest on this kind of a disassembly.

[00:10:59.580] RB: Is the basic design for all the cable trays and all the environmental control system, everything, was that made to take apart? But then it was, then it changed to be a complete system?

[00:11:06] FW: It was, it was in the beginning. Yeah…then it changed to be…Yeah, so these, these many, many interconnections or disconnections would be avoided, yeah? Which is a good feature, yeah?

[00:11:21] RB: Those splices need not have been there originally. It could have been made some other way.

[00:11:24] FW: No. Well, the main thing, one of the trade-off points was, again, transportation, yeah?

[00:11:32] RB: Now, they weren't made here, those things, the ones that you fabric…the ones that you made here in-house. You got those honeycombs from Convair, Texas?

[00:11:40] FW: That is right. Convair. However, the next low bidder, do I remember?

[00:11:44] RB: North America. In Tulsa.

[00:11:48] FW: North America, yeah. And I know a very strong contender for this contract was AVCO. They did not get the job. So there was a good capability available in this country already of honeycomb, very good honeycomb. This was one of the trade-off. We have the capability. We have the state-of-the-art developed to a point in this country that we could use that technology.

[00:12:07] RB: Now, this technology was developed primarily for aircraft control surfaces prior to that?

[00:12:13] FW: I wouldn't know. I doubt it for control surface. For structural members, yes. Control surface, I do not know. Construction…

[00:12:20] RB: Like for a rudder where you need some strength, but you need also…

[00:12:25] FW: But how about a wing? How about the basic wing structure or so, yeah? It's found quite, quite a few instances, yeah.

[00:12:37] RB: Well, can we talk a little about the environmental control system?

[00:12:39] FW: Yes.

[00:12:40] RB: When the system was first designed, you estimated on each one of these cold plates you're gonna have so many watts, so much heat to dissipate, and you designed your systems for that. Then you also had an agreement with the McDonnell people that you would cool off their telemetry in the top of their stage with the same cooling system? But this is just the telemetry for their stage. It has nothing to do, it's not integral with the IU. Their own things that you're cooling, running off your cooling system.

[00:13:04] FW: Yes. Mm-hmm. That is right. Let me make a few comments on this design of the cooling or environmental control system, temperature control system, whatever we want to call it. The design constraints were rather bad. They're rather hard to meet for this reason: in the very beginning of the development, we had considerably more electronics in the instrument unit and the S-IV-B stage. As we carried on later and we developed Saturn, where we dropped very many measurements, telemetry, and other gadgets we needed for either safety or for just to measure different phenomena in order to allow troubleshooting in case we had failures.

[00:13:56] FW: Now, therefore, the environmental control system was probably over-designed in the beginning. We needed it. Now, we did not at this time anticipate how a huge, very complex guided rocket system would go on in development. We had to play it by ear and play it safe, kind of. So we possibly put too much in. In fact, later on, the S-IV-B stage engineering, they commented that they do not need any cooling of their hardware at all. Which again left us with a considerable cooling capacity without the use, yeah? Of course, we had to drop quite a few of our own.

[00:14:39] RB: Well, didn't you have to put heaters in then to balance this cooling capability because it got so cold?

[00:14:43] FW: In some cases. In some cases, heaters had to be…which is very undesirable. But I think by slight design changes, it is not necessary anymore if I…I'm not very familiar with the latest few years right now. I've been drifting out of it.

[00:14:58] RB: Didn't also, in order for this cooling problem, didn't they paint the first couple of IUs white and then to warm it up a little bit to paint the last number of IUs black or [inaudible]?

[00:15:10] This… this I couldn't…I…I do not…I doubt it. I do not know, but I doubt it, yeah?

[00:15:15] RB: And then they covered it with a cork substance after that even. Cork and what's that?

[00:15:19] FW: [inaudible]

[00:15:20] RB: For even more heat absorption and damping for the guidance system.

[00:15:21] FW: Radiation.

[00:15:26] RB: And also they have this kind of epoxy and lead chip paste that they put over the outside around for damping for the guidance system, I believe.

[00:15:37] FW: I, as I say, I couldn't comment. This is later changes, yeah?

[00:15:40] RB: Yeah, they just did this a couple of months ago.

[00:15:48] FW: Are you going through the different subsystems of the IUs?

[00:15:52] RB: Yeah, would you like to do that?

[00:15:53] FW: Could I get a copy of this little booklet…. was started...

[00:15:55] RB: Yeah, sure. You put it out in Astrionics, I believe?

[00:16:00] FW: Yeah, we put it out in Astrionics. [Flips pages of booklet] In fact, I triggered this thing and we had a good sale there. I shouldn't say sales. A demand.

[00:16:10] RB: Do you have an extra one of these so I can demand one? [laughter]

[00:16:12] FW: Yeah, this is one of the very few leftovers. I think I can give you this copy.

[00:16:17] RB: Okay.

[00:16:18] FW: The point at this time was that very few people understood why the instrument unit, what it is for, and what it does. For education of very many people, we had to deal with on specific issues, we provided this little booklet here, which goes through the major subsystems.

[00:16:35] RB: Mm-hmm.

[00:16:37] FW: [Flips pages of booklet] Yeah, so we just talked about the structural system as you see. Very…pictures, a few remarks, the environmental control system, how it works.

[00:16:52] RB: Now, did AVCO have any trouble in making these plates that you know of? Any particular problem there?

[00:16:58] FW: No, no basic problems. Problems first were encountered when these plates were mounted to the instrument unit structure. We needed to change the support or the bolts or inserts for bolts in the instrument unit.

[00:17:19] RB: Mm-hmm. In the core density unit?

[00:17:21] FW: Yeah. Right. And we had another problem, the mounting of the platform, which...

[00:17:28] RB: That's heavy.

[00:17:29] FW:...was quite sensitive because the mounting such plates [flips pages of booklet] or the platform was on a structure which is exposed to considerable loads, we would put stress into the structural members of either the cold plates or the platform, which is very sensitive to this amount, as you know, the IMU, as let’s call it. The design, therefore, of the structural connections between the instruments [flips pages of booklet] and the structure needed special attention and redesign. Of course, the guidance system, by the way, for your information, if you use this, there is a diagram of the instrument unit as it relates to the three stages and to the spacecraft. It shows the basic subsystems in different colors, so it's easy to detect. If we talk about the guidance system, [flips pages of booklet] it is red and black, stabilized platform and computer data, etc..

[00:18:39] RB: Mm-hmm.

[00:18:40] FW: Platform. Shall I make a few comments to this?

[00:18:43] RB: Yeah, would you kind of go into the history of the platform, back to the ST-80 and 90?

[00:18:54] FW: The platform is based on two types of inertial components: gyro and accelerometer. The special features of those two components is that their bearings are practically frictionless, air bearings or gas bearing as we call them. This is very important because any friction or torque in such a bearing would be a direct failure or error in our guidance in our measurements. The air bearing has provided us very good service. Rocket can be easy manufactured. However, a great care must be put in making the bearing parts geometrically exact, very exact in that. However, since these are cylindrical parts, it's easy to make an accurate cylinders. It's hard to make a spherical member.

[00:19:48] RB: Was this development particular here at Marshall, air bearings?

[00:19:52] Air bearings was a Marshall development which started already in the early ‘50s. In fact, at that time I worked in laboratories where we developed the air bearings used for accelerometers. So air bearings, it started here very heavy early in the ‘50s.

[00:20:10] RB: And they started with accelerometers and then moved into gyros?

[00:20:13] FW: No, the air bearing was started for the gyro application. However, it showed so good features in accelerometers that it was adopted there too. It may have been misleading, what I just said before…I said I used to work with air bearings, especially developed for accelerometers because I was in a section which was responsive for accelerometers. The gyro development had been going on already for some time when I got involved in the air bearings, but I utilized their experience then, too. The air bearings were also used for the inertial components for the various army missiles before—that is the Redstone, the Jupiter, and later on, of course, the Pershing.

[00:21:02] FW: The Pershing stabilized platform used practically the same design of gyros and accelerometers as the Saturn basic design. However, the mounting of the gimbals was in such a way that the gimbal angles were restricted to a few degrees because a ballistic missile doesn't need to have all the different features as the need for space launches. In the beginning, we wanted to have an option that allowed us 360-degree freedom around all three axes, which required four gimbals in order to avoid gimbal lock. If two of the gimbals would be parallel in certain cases, it would be gimbal lock. You needed four gimbals actually to overcome that possibility. So we had started the development of the ST-124 with a four gimbal layout. However, when the requirements of the spacecraft, and Apollo especially, were known better, we eliminated the fourth gimbal and used this as an additional option to ask for a fourth gimbal only when needed. We never needed one.

[00:22:15] RB: Was this the so-called redundant gimbal system that they thought of that was redundant in the pitch axis?

[00:22:22] FW: The redundant pitch axis, yes. That's it. What else can I say about the platform?

[00:22:30] RB: Can you say a little bit more about air bearings? Were you ever in any arguments with Draper up at MIT and liquid bearings? Was that ever a thought?

[00:22:39] FW: Yes, yes, there were very many arguments. However, the arguments were usually developed by the users and not by the developers. The developers saw the different advantages and disadvantages of different systems, and they used them accordingly. However, there was also a little bit preference given on personality basis, yeah? It is not a real competition anymore between these two types of bearings. One is that we reached the end of the rope for either one, and we are looking for different, new methods to measure in angles or in space or acceleration in space.

[00:23:17] FW: Secondly, the developments have come to a point where each of the different methods has the features of the other. Let me very shortly describe a few of the major differences. In a liquid bearing—a fluid-supported, a floating bearing—you need to have a very precise gravity in a liquid, and you need to float in which has exactly, overall, this specific gravity of the liquid, which is very, very hard to do.

[00:24:02] FW: You have two problems, the dynamic balance and the static balance, both. One is that the float is just floating and doesn't float up or down, but stays in the liquid where you put it. And the other one is that dynamically, that it would not turn or twist because one side of it is as heavy as the other. This is very difficult to do. Also difficult to do is to fill this liquid, which by, we need a very heavy liquid because the heavy liquid has a damping effect, and it is very difficult to fill a bearing with that type of liquid, bubbles or gases.

[00:24:37] FW: And as I say, it's a very difficult procedure to do these things. The gas bearing does not have that disadvantage. It needs only a very good geometrical size, very good size and very close tolerances. But it does not need to be especially balanced at all. It can have any weight, yeah? Because if you need to carry a little bit more, you increase the pressure slightly. The errors are about the same for each one. If you do drift in both, it's about the same.

[00:25:11] FW: If you put them on a comparative basis, you could compare two different ones and say they are very different. But when you look into it, you find that either the flywheel is heavier or the bearing is not comparable or the environment is different. So if you compare it on a comparative basis, they're almost the same.

[00:25:27] FW: One is a little more difficult to make as the other. The air bearing has one disadvantage. You have to supply gas because the gas is used up. It flows out. It was at least the case in the earlier designs—yeah?—so you need it for long flights to provide a lot of gas. Now, that's the reason for Apollo, we did not use the gas bearing. We used a fluid bearing because the fluid is not used up. It stays there—yeah?—and the Apollo is used for many hours-–the Apollo gyros or IMU. Where the Saturn is only used for a few minutes. I think we upped it now, lately, to seven and a half hours, yeah? We have a seven and a half hours capacity. But the Apollo gyros have a day's capacity—yeah?—many days. You wanted to?

[00:26:14.420] RB: Is the air or the nitrogen that comes in here, isn't it recirculated?

[00:26:20] FW: That is the very new design [flips pages of booklet]. This is the point I want to mention: later on, liquid bearings were improved to make them a little easier to manufacture, test, and so on. The gas bearings were developed to use a recirculating liquid, or gas, actually. So you do not need to carry tanks anymore. You see air bearing supply. It's a huge tank. You had to carry a lot.

[00:26:50] RB: About 50 cubic inches.

[00:26:51] FW: Yeah, and I think if you want to carry more for a longer operating period, you need several of those things. At this time, it was not laid out for seven and a half hours. I know this. I do not know how long this air supply would last. This is the platform. Well, this much is good. The platform has shown a very good reliability, yeah?

[00:27:21] RB: I know you've had one on a breadboard. The gyros have run about 6,000 hours.

[00:27:26] FW: Of course, there is one more disadvantage of the air bearing I should mention, at least the open loop bearing. The closed loop is different. But the open loop bearing, where you use up your air supply or gas supply, every dirt or dust or foreign particles in the supply may get stuck in the bearing and restrict the bearing itself and give you inaccuracy. So there are pros and cons to both methods. It is also much heavier, this platform, as the Apollo IMU, yeah? But it is accurate, and for the Saturn, it was the right decision, very right.

[00:27:59] RB: About 120 pounds, something like that?

[00:28:02] FW: Yeah, I think it was about that much. [Flips pages of booklet] I do not know if we have weights in here.

[00:28:077] RB: That's all right.

[tape cuts out]

[00:28:18] RB: Was this designed for any particular serviceability too, so you can swap gyros out? All the gyros were the same. I mean, you could swap out X, Y, or Z gyro for any other,
couldn't you?

[00:28:30] FW: Yes, you could. You could. It is not very desirable to do it because of the precision. The removal or replacement of a gyro or accelerometer is a major effort. It requires that you adjust the angles very, very accurately to the angle measurement. It's a precision mechanical unit, and so you need to have a good laboratory to do that.

[00:28:55] RB: Can you explain to me, and I'm not sure I'll understand it when you finish explaining it to me, how this is optically aligned to accurately pinpoint the vehicle's position on earth? Then what happens when it is released at, I think, T minus 17 seconds? And how are the gyros spun up in the first place? When are they spun up?

[00:29:21] FW: The gyros need to be spun up before you start the alignment. The inner gimbal has a window, has an optical surface.

[00:29:33] RB: Two mirrors.

[00:29:33] FW: Or two mirrors in order to get this optical surface visible from the ground, yeah? Oh, my…

[00:29:40] RB: That's what I said when I read about it, “Oh, my.”

[00:29:42] FW: This is a, this is a, yeah…[chair moving] The reason it's hard to explain because it is built with a closed loop with torques within the platform, so that you…it is a self-alignment. Once you get the, by eye, the mirror up there.

[00:30:00] RB: Then you can run the other one around with the servos.

[00:30:01] FW: Then you, yeah. And you keep it, you keep it locked in. You keep the optical locked in, so every time the platform wants to turn away. To this scheme for aligning the platform. This is a fairly complicated scheme, which is done in a closed loop after a while. That means the platform is locked in. If it wants to drift away, it will develop some torque in one of the gimbal rings and turn the gimbal back to the position it should have. So it's a fairly accurate and good alignment scheme, automatic. But I think to explain it in all details, I would need a block diagram to go through it, which I do not have available right here. [gets up from chair, moves around] Of course, [inaudible] doesn't have much here. I don't know. [sits back down at table] Of course, there's a model of the ST-124. And you notice here the frame. Which I, which I've all… It was actually, why, why doesn't it move?

[00:31:12] RB: You have a tape here.

[00:31:15] FW: This would be the fourth gimbal, yeah. Yeah, the blue one, as you can see. No, it's locked in somewhere.

[00:31:19] RB: Gimbal, gimbal lock.

[00:31:20] FW: Gimbal lock, that’s right. [Both laugh] Yeah. So, but eventually you end up with all the gimbals. Now, I do not know where the, where the surface is, which we see then from the ground. I believe, if you want to go in some more detail on any of the components, it may be good that you discuss it with the designers who were involved with the design.

[00:31:44]: RB: I want to talk with Dr. Seltzer.

[00:31:48] FW: Seltzer was involved in the operational aspects. That means the flight operational aspects, the flight dynamics, and these things.

[00:31:57] RB: With Mendel's group?

[00:31:58] FW: Mendel is in charge of the platform, and they have a considerable amount of information. Now, the Saturn is one of those projects which is documented in all aspects, management and technical, much better and much heavier and deeper as any other project I know of. You can follow practically any thought or any question down to the very, very detail. What we can discuss here and where I can help you is only in generalities and the overall general approach.

[00:32:31] RB: Yeah, we want to get a feel for, you know, how the equipment works.

[00:32:35] FW: This is, of course, the basic material for the structure, yeah?

[00:32:43] RB: Is that honeycomb?

[00:32:44] FW: This is honeycomb, yeah.

[00:32:45] RB: And these are two different densities that you use?

[00:32:47] FW: No, you just pull them apart.

[00:32:49] RB: I wouldn’t do that.

[00:32:50] FW: See, this material comes in a block which looks almost solid—yeah?—and it is cut then to pieces. This was actually a much heavier block, about two inches, the thickness of the instrument unit. It is cut by a saw, it's aluminum. Then, of course, it's by machines pulled apart and you get this honeycomb.

[00:33:13] RB: And then for different densities you just run it back together?

[00:33:15] FW: No, you don't run it back. Machines do it, pull it just in a way that it is exactly alike, that the density of these different holes or the size of the different holes you pull open is exactly the same, yeah? And then, of course, it is filled with certain plastic foam and a heavier layer of plastic outside foam.

[00:33:34] RB: And that gives it the rigidity?

[00:33:35] FW: It gives it rigidity and damping features too yeah?

[00:33:37] RB: Damping, okay. Oh, yeah, damping would be a good reason for...

[00:33:40] FW: Now, this is here a small cut of the instrument unit, of an actual instrument unit. This was done for testing. We had some problems indicated earlier in providing inserts. So these are inserts here, yeah? As you can see, they are cast into plastic, yeah?

[00:34:05] RB: Now, these inserts go all the way through the wall?

[00:34:07] FW: This is the wall, this instrument unit, they go all the way through, yeah.

[00:34:12] RB: And this is the entire thickness of that wall?

[00:34:13] FW: Yeah, it's the entire thickness of the wall, yeah.

[00:34:15] RB: I didn't know these inserts...And then these are bolted on the outside and, of course, shaved off clean.

[00:34:20] FW: Yeah, right.

[00:34:21] RB: But they're bolted right through. And these hanging structures that we saw in here that the cooling racks are on, those things go all the way through and are bolted on the outside through these special aluminum fasteners.

[00:34:33] FW: These are these fasteners here. What else we...Oh, yeah. If we go to the computer, we can discuss that a little later. Did we cover the platform?

[00:34:46] RB: Yeah. Temporarily.

[00:34:47] FW: Or shall we say a few more words to it? A few of the design features which were new at this time to put as much of the electronics on the gimbals as possible in order to have the minimum amount of wires going through the gimbal. The platform does not have full 360 degrees freedom, but it has this only with a fourth gimbal, which was never required. If you want to know some about the management and the costs, I have made a study some time ago, compared the Apollo with the Saturn. But I would have to dig it out. You could read it if you want to.

[00:35:28] RB: Okay. I might drop by next week and answer for that.

[00:35:31] FW: Okay. Of course, I can get you a copy. You read it some at your leisure and bring it back here. The reason for that, by the way, was primarily...Oh, there are two reports, I should say. One is based on the other. I was called in once to look into some problems developing in the production and assembly of the Apollo IMUs at Milwaukee AC Spark Plugs, or AC Electronics, as it's called now. I did this and prepared a report. Later on, the cost was questioned very seriously by our own management—the cost of a specific platform ST-124M as compared with an Apollo platform. And I was asked to look into this, so I compared the costs, and I have a report on that, yeah?

[00:36:27] RB: You mean the AC unit was more or less than the...

[00:36:31] FW: This I would have to answer a little different. If you buy a gyro or an IMU, you say, “I want one more,” yeah? You pay less for an Apollo as you pay for a Saturn. But if you look into the cost and say, “How much money do I invest total in the IMU for the Apollo?” and “How much do I apply, do I need total for the ST-124M?” and divide it by the number of actual flights to support, the Apollo platform is much more expensive. Did I make it clear?

[00:37:13] FW: The reason for it is that—there are many reasons, yeah?—one is that you need to manufacture about 100 gyros before you get 20 good ones. See, the Apollo—and it's kind of unfair to compare them on that basis—the Apollo needs a longer lifetime. There is one life restricting item in each platform, in each gyro: that is the ball bearings of the flywheel. The flywheel is very heavy, and it has to work very, very high speed—yeah?—I do not remember,
I think in the order of 20,000 RPM, which is extremely high. Heavy wheel, and you cannot afford much tolerance in there because you cannot afford to have this flywheel move at all. It has to stay in relation to the outer bearing where it is.

[00:38:06] FW: Now, you have to put a pretty good preload on these bearings and, of course, little lubrication, so they do not last long. In Apollo, they have found that they make many bearings and then they run them for a long time and test them again and again and again. Those which make it through the first, let’s say, few hundred or a thousand hours—yeah?—they are very likely to last much longer, and the other ones drop out.

[00:38:34] FW: Now, these type of things increase your cost. It is not the only effect, but I give this as an example to show what type of effects you have there. If you want to go in more detail than this, we can do it any time. I think on this I am pretty much at home because I made this study. Do we have enough on the platform or shall we…yeah…okay…

[00:38:56] RB: We better do the computer.

[00:38:58] FW: The computer. The actual computer was developed by IBM and it was a kind of a breakthrough or a change in technology as compared, for instance, with the Titan computer, which was produced by IBM at the time at the same plant when we started the Apollo computer.
This little gadget here is not an Apollo printed circuit board, but it is one which they use in their commercial computer.

[00:39:30] FW: The Apollo is, however, the same type of philosophy. A little chip, as they call it, a small printed circuit board, only the connections in our case were mounted differently. The connections, because we needed printed boards on both sides, not only on one like on this one—yeah?—we needed to have clips which reached around the corners for mounting and connecting the electrical connections. This gave us quite a few problems. These little springs were very, very hard. I am sorry I do not have an example here—yeah?—these little springs gave us a considerable problem. IBM had really to work hard and spend a considerable amount of money to make them perform properly. The point was to...

[00:40:13] RB: They are like pages put in there, aren't they, these printed circuits?

[00:40:16.] FW: Yeah, they are, they are. [Moves away from table, begins writing on chalkboard] But unlike this connection, we had printed circuits on both sides of the board. We needed to have clips which went around, let's say we solder them here to the board, and there's a clip. The clip in this case was the mounting facility. At the same time, it was the electrical connection. Of course, they had the whole board on two sides here. These clips [taps chalk emphatically on chalk board], they were a real headache, and the technology to develop these tiny little clips and work them was very difficult. Every time they soldered here [draws on chalkboard], the heat came in and gave them some problems here. They even developed different types of soldering points [comes back to table] an optical soldering by focusing a very high lamp through a…is it a...sold[er], a solder[er] optical system? And focus it to one point and solder this way not to get a mechanical contact to the printed circuit board. It's very tiny as you can see, and it's hard to work these. This was one of the bigger problems. I think the systematic problems were minor compared with this little clip problem we had for a long time.

[00:41:35] FW: The other novelty together in technology is the small transistors. These little chips you see on each side are transistors. If you look close, you'll see that there three connections on these three little points. These three little points are connected in the board if you can see on the board this little square mirror type looking item—this is a transistor. The way it is put in, it’s just pressed down, and by the pressure, the electrical connection is good enough that it stays on. I believe there is one point where the chip fell off, and where, if I'm not mistaken on this one…No! It's still there. I thought there was one point where the chip fell off where you can see the three connections before here which are open.

[00:42:26] RB: The pieces of the mounting piece there is what? About three quarters of an inch square?

[00:42:31] FW: Approximately, yeah.

[00:42:33] RB: And those transistors are about the size of a pencil point. They're square, yeah.

[00:42:40] FW: [inaudible] They have three electrical connections each. You notice another thing, the black blocks, these are resistors. Now the resistors are painted on with a paint which contains small metal particles or carbon, yeah? Now the resistors are painted on, and after they are painted on, they are adjusted to the proper value by sandblasting part of it off, yeah? That was also a rather new manufacturing method. You see that parts of it is missing on either one of the resistors. They are sandblasted automatically. There is a resistor measuring bridge which in itself has an electrical contact to trigger and to start the sandblast, yeah? The same
sandblast is of course moving along the transistor and when the proper amount is reached it's cut off by the resistor measuring bridge.

[00:43:39] FW: The memory is a magnetic memory, and it is based on tiny little—how shall I say?—magnetic rings, yeah? It's then…wires are put through…you saw these memories, I think, yeah? Now, these are so small that they would not easily work by hand, yeah? It would take years for anybody to align them and put these wires through. The second one from the top—this must be Saturn sized, yeah? This is, I understand, for the Titan computer, but there's a different system. These here, the 1321, is the Saturn size, which is rather small now. The next smaller is the next generation. I do not know where they are used here.

[00:44:33] RB: This is not talking about a pinhead; you're talking about the size of a pinpoint there.

[00:44:36] FW: These are actually little rings of magnetic material. Now [inaudible]

[00:44:41] RB: What kind of material is that? Just steel?

[00:44:44] FW: No, it is not steel. It is a synthet [sic] material. It is very fine powder synthet [sic] together. Now, they are different binders, and I do not know which they are. The way these memories are done is quite interesting. There is a metal plate in which the worker—mostly they are women because obviously they can work much better with these—it's put on this plate, and the plate has a carved in very tiny little half round and grooves. Groves? Grooves?

[00:45:26] RB: Grooves.

[00:45:28] FW: Grooves. Then the vibrator starts to vibrate this metal board and these little memory rings, the magnetic rings they just fall in place. After a while, they are lined like soldiers on the board. Through a microscope the workers see if all holes are filled, and then they are ready to put by a machine very thin wires which are very straight right through the whole line of things, yeah? It's amazing. It works very good, yeah? The vibration makes it. I think it's a very high frequency vibration.

[00:46:07] RB: Was this technique perfected at IBM? Over here?

[00:46:10] FW: It was perfected at IBM. Not here, no.

[00:46:12] RB: In Owego.

[00:46:14] FW: In Owego. The computer was built in Owego, but this technique was not developed…oh yeah, this is Owego. These printed circuit chips or boards, they were developed in another plant in New York, I forgot the name. I was there. I saw it.

[00:46:33] RB: Binghamton?

[00:46:34] FW: Near Binghamton. Not directly Binghamton. Binghamton has a larger plant
and has lots of [inaudible]. But there was still another plant about…oh…eight miles from Binghamton where these things are made.

[00:46:47] FW: I mentioned before the computer itself did not provide any greater problem than its systematic development. But the hardware development—as I mentioned before—the computer has the option that you can add memories at will, yeah? You can have a larger or smaller capacity. The real issue, however, was the software. Now, I must say here that it was the first time that Marshall went into real digital, large digital system. The Pershing, which we worked on before, as you know, was an analog system, completely analog. It was built up on mechanical and electrical analog systems, yeah? I believe...

[00:47:39] RB: Was that still using the ball-and-disc integrator?

[00:47:43] FW: Yeah, that is one of the gadgets used, yes, but it [mustered?] a number of gear trains and potentiometers and various mechanical gadgets—yeah?—in the Pershing. The decision was hard to make at the Pershing time because at that time, it so happened that the digital technology just came up, and the risks to come up with a low-cost development was, with digital means, was very great. We would have to pay considerable effort to really develop the technology for a digital Pershing computer.

[00:48:23] FW: For when we had to make decision on Saturn, there were already several digital systems available, not available, but tried, yeah, the Titan and I think the Atlas II. So we had already experienced available in industry. For Marshall, it was the first experience with digital systems, with a larger digital system. [Flips pages in booklet] So the software was more of a—how shall I say?—I shouldn't say a problem—was new to us even more than the mechanical or the design of the hardware. Design of the hardware, I think there was not a real difference in between analog and digital, small electronics in either case. But the software, that is, the digital programming, that was new, and we had to learn that. It worked quite satisfactorily, as you know. It helped a lot along that line—that is, in the development of the digital programming, as well as the systems—that we started rather early with a so-called breadboard, simulating a Saturn vehicle, including the ground hardware, and exercising it through the breadboard. I don't know if you saw it. It's in one of those field buildings. I think it's still there.

[00:49:37] RB: I would like to go and see [inaudible].

[00:49:38] FW: Yes, I think you should. Of course, you should realize that what you see now is the end of the program—yeah?—in the beginning…It’s one of those steel buildings on the right, when, from the tower to the right, yeah? Realize that we started out with a tiny little setup,
and we enlarged it as we went along, you know. This is Saturn 1B breadboard. The Saturn V breadboard is back in the quality laboratory building.
.
[00:50:12] FW: Platform server amplifier, I think there's nothing. These are straightforward electronic design servos, servo design. Power supply is new. This is based on a digital method, yeah? This was a kind of a breakthrough, too. Before, as you know, the power supply, that means conversion DC to AC, to 400 cycle AC, was done mechanically by motor generators. They were perfected to quite a good reliability. However, for new space systems, they were not sufficient anymore. So this is a digital system which converts from DC to AC. It's quite novel. If you are interested in details, you may want to talk with some people, yeah? It's a quite interesting, tricky system, yeah?

[00:51:05] RB: Can you say a little more about it?

[00:51:08] FW: The point here is that you have a very good power—how shall I say?—efficiency, input output. The digital system is a real interesting thing. I could, to say here on this item again, without a diagram here, I couldn't give you very much detail. The man who knows it very well is Kreider, Bill Kreider. He works over there in, in, uh…

[00:51:38] RB: Can you spell that last name?

[00:51:38] FW: K-r-e-i-d-e-r. E-i, E-i.

[00:51:46] RB: Right, okay.

[00:51:46] FW: Kredier, Bill Kreider, and he works in research laboratory, research branch of Astrionics. He's good at it.

[00:52:00] FW: Data adapter. Oh. I should have mentioned this before. The data adapter is nothing but input-output facilities for the computer, yeah? So it is conversion when, whatever we needed, conversion analog to digital or digital to analog, whatever the inputs, outputs required, just to adapt the computer to the overall system. You would have to go through each of the channels and see what the input is and how you have to convert it to make it digestible by the computer. [Flips pages in booklet]

[00:52:33] FW: Control system. Oh, this is the wrong picture, by the way. We never built this control computer this way. Oh, this, I'm sorry, this is a [rate?] gyro, yes. [Rate?] gyro. This is the first time we really adopted the redundant approach. We use three each gyros. Nine gyros in this one. Whereas in contrast to the main platform, there are only three gyros and three accelerometers. No redundancy. We considered the redundancy to this one to be available from...

[00:53:17] RB: Skylab?

[00:53:19] FW: No, not from Skylab, from the Apollo. The Apollo had two platforms.

[00:53:22] RB: Oh, yeah, that's right. But this wasn't always hooked up with the Apollo platform, was it?

[00:53:27] No, it was not. Let me see…

[00:53:28] RB: When did that happen? I notice you have it here on this schematic that it's hooked up to the platform.

[00:53:34] [Inaudible] command. So this must have been a very old decision—yeah?—if it is here. That was very early during the development of the IU. The decision must have been early.
There were in between, at different times, various discussions on whether it is feasible to do it one way or another way, and the quests for changing the old concept came up again and again, but I do not believe that it ever changed.

[00:54:01] FW: Not on here is the control computer. [Flips pages in booklet] The control computer which is developed by a firm in Tampa, Florida. ECI, I think. It was not a breakthrough at all. We needed quite good amplifiers, substantial amplifiers to run inputs to the various controls for the engine. But I believe there was nothing really breakthrough. The only point was design, features and design problems, mechanical and electrical design. The printed circuit boards were a little bit harder to handle. They were rather big, rather large for their—how shall I say?—for their normal electronic uses. If you have large printed circuit boards and environmental changes, let's say temperature or so, we didn't want to have too close tolerances put on, yeah? Then you have a few problems that the contacts would break or soldering spots or so.

[00:55:19] FW: The control computer has one problem involved in...I have to go a little bit into the system here to make it clear. Practically each of our Saturn flights has a different payload or has a different flight program for it, yeah? If the bending moments, the payload and many other variables affect the control system, it may make it unstable if we have the bending mode filters in. We would break up the payload, we may have an unstable flight, yeah? So the control computer is one of the items which has to be changed for each configuration or each flight, which is not so for the other components, yeah? It applies to the control computer. So filter networks are put in rather late in these control computers.

[tape ends]

0:56:19