Summary Session, 19 February 1957
Tracking and Communications Panel
Participants and Their Topics
Clark A. Potter (Panel Leader), Head, Electromagnetic Propagation Branch, Navy Electronics Laboratory, San Diego, California
Max Fishman, Research Scientist, Lockheed Aircraft Corp "Tracking Techniques"
Physicist, Rand Co ., Santa Monica, Calif.
"Tracking and Communications for a Moon Rocket"
James A. Marsh,
Vice President, Systems Laboratories Corp., Sherman Oaks, Calif.
"Survey of Communications Problems Associated with Space Travel"
Head, Electronics Branch, Office of Naval Research, Washington, D. C.
Section Chief, Jet Propulsion Laboratory, California Institute of
Technology, Pasadena, Calif.
H.E. Singleton, Litton Industries, Beverly Hills, Calif.
Dr. Max Fishman of Lockheed presented a comparison of various types of satellite tracking systems. Dr. Fishman first presented a very short discussion of the pulse type of system which he discarded without much further attention. A little bit of numerical work will show that for the ranges in which we are interested for satellite tracking, a pulse system will require somewhat unusual transmitter powers.
The types of system which one might consider, then, in a general way are CW systems. These can be broken down into amplitude-sensitive systems and phase-sensitive systems. A system which uses amplitude comparison, such as a conical scan direction finder, can be used if one is satisfied with comparatively crude measurements. Dr. Fishman quoted an angular accuracy of the order of a mil.
On the other hand, the phase-comparison system using two antennas or sets of two antennas offers some possibility of better accuracy. We might consider three types of the interferometer systems. The first type involves a fixed base-line .and a fixed array in which nothing moves. Examples of this sort are the Minitrack system or the Microlock system, both of which are intended for use with the Vanguard type of satellite. The next type of system that one might use would have a fixed base line, in which the antennas are steerable. One might be interested in using a system of this sort at some frequency which is high enough so that the antenna configuration is reasonable. Steering a 160-foot diameter paraboloid may be difficult. At a frequency which is high enough one can use a smaller steerable paraboloid or some sort of directional antenna. The third type of system which Dr. Fishman discussed involved using a steerable base line. This system involves antennas of fairly large size which are separated by ten to twenty wavelengths. It is the base-line rotation which is used to detect the angle of arrival.
All of these systems are sensitive to angle of arrival. They have the same general sort of limitation in that they require the beacon in the satellite to be stable. None of the systems which were discussed in this paper involved any sort of a dosed loop type of transponder.
In addition, the type of system which uses a fairly extended base-line such as the Minitrack or Microlock system is subject to certain mechanical limitations. If there is a shift in the electrical center of radiation of the antennas so that the baseline changes by a fairly small amount, there is an error which may be unacceptable. If we are satisfied with angular accuracy comparable to that of Minitrack of a tenth of a degree or so, this error may be negligible. However, the type of a system that uses a steerable base line has the advantage that any shift in frequency or any shift in mechanical base line will appear only as a sensitivity change rather than a change from the precision of the null.
The second paper was by Mr. Gabler of Rand Corporation. This paper considered the problem of tracking and communication for a moon rocket. This particular configuration was intended to impact on the moon and thereafter to telemeter data back at a fairly low rate. Presumably some things that one is interested in learning from the moon's surface do not occur at rapid rates. In any event the use of low power does not permit high data transmission rates. In this configuration the rocket is tracked (and navigated) from a set of possibly three or four stations which are separated in longitude. The reason for this is that the earth rotates which makes it desirable to have the rocket in view by at least one station and preferably two.
In this type of configuration a typical trajectory from the immediate vicinity of the earth to the moon takes something like two and one-half days. The tracking parameters one wishes to measure are range, range rate angles, and angular rates with respect to some earth reference. One way to obtain range is to use a transponder beacon in the rocket itself. Rate can be obtained from the Doppler frequency shift, assuming that the beacon is sufficiently stable.
In the moon rocket configuration of Mr. Gabler a comparison of the measured orbit with the precalculated one can be made in the usual way. The tracking accuracy is determined in this type of system by the thermal noise. For the purposes of design, this system involves a very small bandwidth of the order of 1/20 of a cycle per second. We may then improve accuracy by integrating over a fairly long time and programming relative time. In addition, we may program such things as relative earth to moon motion, if we are satisfied with a slow data rate.
Presumably one might locate the point of impact on the moon's surface to within some 25 miles under favorable conditions. Such conditions occur when random errors are made smaller than systematic errors by integration and by use of redundant data. If one chooses the design frequency for this system at 200 megacycles, the Doppler is of the order of 10 cycles per second. We assume that the oscillator is stable to 1 part in 108.
The next paper was somewhat more ambitious in scope. Dr. Marsh of Systems Laboratory Corporation presented a minimum communication system between a ship and a base. It is assumed in this case that we know where the ship is and that one has only the problem of getting information back and forth. This system is restricted to the case where we have only one rocket ship about which to worry. The base may be an extended type as in the previous paper where we had more than one station, but we will consider this to be equivalent to a single station or a single base. At all times it is assumed that the base knows where the ship is and the ship knows where the base is, so that we have solved the location problem. We merely wish to transfer information.
The problem here is simply to get adequate range for the bandwidths we intend to use. One can do this by using antennas which are quite large physically. It is possible to use large fragile arrays which may be assembled or extended only in the low-gravity region of outer space. They can be extensible arrays such as an umbrella type or possibly some sort of metallized film which is quite weak mechanically. Since the antenna array is exposed to little or no drag, such construction methods would be satisfactory.
There is, however, a difficulty involved in orientation. If one builds a large array so that there is sufficient gain, associated with it will also be quite a narrow beamwidth. The ship must be properly oriented as well as the array on the earth. This means that considerable attention must be paid to the programming beforehand. As you know, if you have had any experience with attempting to line up narrow beam antennas, this can get to be somewhat annoying.
One can also take advantage of certain other facilities that are available in outer space. One of them is simply to use the low ambient temperature to keep detector and mixer crystals cold to decrease crystal noise. In this type of system, we can take advantage of a fairly long integration time if we like. How long the integration time will be depends on the bandwidth of the type of information we wish to transmit. We can use redundant transmission provided the required data rate permits. Dr. Marsh indicated that the use of the Maser would provide an improvement of the order of 30 db. This is a cheaper way to get 30 db than building increased transmitter power into the ship. He also quoted some figures on how much weight is actually required for the ship to get a given payload. He said the normal ratio one uses in aircraft is something of the order of seven pounds of airplane to one pound of useful payload. But for a moon rocket, the ratio is more likely to be between one hundred and one thousand to one. Therefore, one ought to pay considerable attention to miniaturization techniques. He also indicated that in this particular case, there is no major breakthrough that is required. It is simply a question of using the known techniques.
The next paper was a discussion by Dr. Shostak of ONR concerning Radio Astronomy. To navigate in outer space and to communicate between earth and moon or between earth and isolated ships, it is desirable to know something about background radiating sources. These sources can then be avoided where they may disrupt communication systems or they can be used as navigation aids.
The most common example of radio astronomy is the well-known 21 centimeter line of hydrogen. If one looks at the transition probability that produces this line, one might feel that there would be no reason to expect it to occur in the universe. However, our galaxy is large enough so that we get a quite respectable signal return in the twenty-one centimeter region. Incidentally, this radiation is not quite black body, but the intensity seems to increase with frequency, rather than decrease.
Most of the radio astronomy that has been done has been a postwar development because of the instrumentation aspect. One requires fairly large antennas and quite good machining tolerances as far as the mechanical components are concerned.
In connection with navigation techniques, one might do some sort of sun tracking. This will provide a line of position. If one has a position line and dock or a stable oscillator, then one can presumably navigate. The sun is a possible way of obtaining a position line but it does have a finite diameter. When solar flares occur, the diameter is not always necessarily as small as we would like to have it, so that people have looked at other possible sources. The next one that comes to mind, of course, is the well-known one in Cygnus A which is somewhat smaller but has the disadvantage of also being less intense. One can also go a little bit further in this direction and use sources which are outside the galaxy, if the system is sufficiently sensitive to pick them up. We might also possibly take advantage of the twenty-meter radiation which comes from Jupiter.
In general, one ought to use fairly short wavelengths when one is attempting to do something over long distances because of the dispersion of the beam. The other general rule is to avoid regions of operation which are quite close to well-known sources such as the hydrogen line.
The next paper was presented by Dr. Singleton of Litton Industries. This paper concerned a system in which an unmanned vehicle is navigated from a circular orbit centered around Mars. The whole process would require approximately one year. The region of interest here is the transition orbit. In this particular system, the navigation is automatic. Thrust is quite low; in this case, something like 10–4 g which gives a transit time of about one hundred days between orbits.
The system is essentially inertial. One has a sun line of position which can be looked at from time to time. This can be programmed to provide correction to the thrust vectors both in magnitude and in direction. Since it is a fairly low-thrust operation, presumably the corrections would not be particularly critical. The navigation system for this type of vehicle would use a set of multiple accelerometers. The drift for such an accelerometer would produce a three-thousand mile error in lateral position per year. Presumably, if one can get within three thousand miles of Mars, one can acquire a spiraling type of orbit which will then become a stable circular orbit. What one does once the vehicle is in a stable orbit around Mars is something else again. Soundings can be taken by auxiliary vehicles (on the surface). The data can then be telemetered back to the main vehicle and in turn to either intermediate stations which are orbiting around earth, or possibly even to an earth-based station.
The final paper was a discussion by Dr. Rechtin of Jet Propulsion Laboratory on the Microlock system. The Microlock system is another way of tracking the Vanguard satellite. The design objectives for this particular system were to develop a fairly light-weight system which would: (1) have a range of between 2500 and 15,000 miles, (2) last for several months as far as the duration of the beacon was concerned, and (3) provide four channels of useful telemetering information having a total bandwidth of the order of 1 cps. The system was intended to operate from horizon to horizon with approximately one-mil angular accuracy. The frequency was chosen as 108 megacycles to be compatible with the Vanguard program.
This system has quite a narrow bandwidth with correspondingly low data-rate transmission. Once again this involves the use of a fairly stable beacon. The transistorized version is quite similar to the beacon which is to be used in the Vanguard vehicle. It is essentially a correlation system in which we maintain a phase lock on the incoming signal. One locks not only onto the frequency but to the phase of what comes in. Since it is a narrow-band system, we simply lock on what is in the band. This may be the true signal or it may be signal plus some unwanted bit of noise.
In the Microlock system considerable care has been expended to make the system linear all the way until, of course, the detection point. This system has been tested in several locations. It has been tested at White Sands and also in Earthquake Valley which is in the eastern region of San Diego County. The system has been found to work quite satisfactorily. One can acquire the signal at quite low power levels. The sensitivity for this system is 155 dbm, which is quite respectable. The bandwidth is approximately one cycle per second. The antenna used was circularly polarized and is a helical array. This type of system uses an extended base line which is similar to the Minitrack system.
The phase-locked receiver is really the part that makes this system work as well as it does. One simple difference between this and Minitrack is that this system involves quite a long listening time in comparison. It is of the order of ten minutes whereas the Minitrack system will listen for something like eight seconds.
In general, the same thread runs through all these papers. The systems that people are thinking about are narrow-band systems. They are CW systems, and all of them contain the concept of a fairly stable oscillator. It seems this is the direction that instrumentation is going as far as tracking and communication for vehicles of this sort are concerned in the near future. Thank you.
Dr. von Kaman:
I thank you very much, Dr. Potter. This is a very dear presentation and I think especially that some of us who were not able to listen to the original papers got a picture of what was discussed. Now I have two suggestions. The one suggestion is that you stay here because I may know something about re-entry, but I certainly know nothing about this subject, so that you can answer directly the questions. The second thing is that I would like to suggest that the speakers come out here and talk in the microphone. I regret very much that we are not so far advanced as to have a slow speed guided missile which could be on the pushbutton here to send to the speaker.
I would like to say a few words. I am not so sure whether they are a question or comment, but would like to get your reaction to them anyway. There is a very large area in this tracking problem for further work and investigation. Reference was made to pulse techniques and various other techniques, presumably with reference both to skin tracking techniques and to transponders or transponding or non-transponding beacons. We are all familiar with the conventional radar used on aircraft, ships and the like in which an effort is made to make the pulse as narrow as possible giving high range discrimination, with all sorts of difficulties. I think that perhaps a few of us (not so many I hope) are familiar with this iniquitous device used by the local constable which radiates nothing but CW then comes back as a Doppler effect. There is a whole range in here from the infinitely narrow pulse to the maximum amplitude pulse where we can explore and find best ways of utilizing this power. I would mention only one with which I happen to be quite familiar. That is the moon radar experiments which were done some years ago. Although a pulse was used for transmission, in many people's eyes it would be called CW, because the pulse was 0.2 second in duration. So there is quite an area here for original work, new ideas and the like.
I can only agree with that statement. Actually the type of CW system that I really meant here was the sort of thing that is characterized by a sine wave rather than by some odd-shaped wave form. Of course one can devise all sorts of intermediate things (pulse Doppler and so forth) and these all have their place. However, if one is interested in making fairly precise measurements he is driven either to a CW system at some form of quite well-known modulation or to a short-pulse system. Personally, from what I have seen of it, I would like to avoid short pulse work. Short pulse circuits are just not well behaved as far as my experience has been, compared to the results that one can get with an equivalent amount of effort in other directions. We may say more about this in the classified session.
Lieutenant Wells (WADC):
Dr. Potter, I noticed that you mentioned that the Microlock system will lock onto whatever signal it receives including some noise. It is my impression that Dr. Rechtin made a point that the system is sensitive to a unique frequency of sine wave and therefore would be less susceptible to noise. Is this correct or not?
Well, I'll stick by my statement, although what you say is also correct. What happens is that the system has quite a narrow bandwidth so that if you want to say what comes through this bandwidth is essentially a sinusoid, well and good. Now if one has a noise source which has an appreciable amount of power within this small but non-zero bandwidth, the system will respond to it. In a practical case, we would expect that the signal-noise ratio would be quite good because the bandwidth is quite small. But if there is energy there and it did not come from the source, this is noise. We are idealizing here by saying that anything that comes through this narrow bandwidth is essentially a sinusoid at the desired frequency.
Dr. Wheelon (Ramo-Wooldridge):
At several points in your talk you emphasized the need for very stable oscillators in doing the phase tracking. I wonder if you could comment on this in the light of the fact that radio-astronomical tracking uses essentially noise sources.
The requirements for the stable oscillator result from using an open loop type of system where one has a beacon that he sends away and over which he has no control. There is a little bit of difference in the constants of the system for noise-source tracking. As far as I know the angular resolution that one can obtain with an optical-type system for incoherent sources is not quite the same as what one could get by using an antenna baseline system for a stable source. I could be wrong about this. Here comes Dr. Rechtin who will resolve this problem.
Dr. Rechtin (Jet Propulsion Laboratory):
Dr. Wheelon was asking a loaded question. I think he knows it. You can do very good angular resolutions of the DF type on a radio source if the source is a small pinpoint. I think what Dr. Potter is concerned with is what we call tracking and communications. The basic problem which we have now is that if we have a limited amount of power the only way we can detect the signal is to have it extremely narrow-band or at least to have its wave-form known within a very narrow band, which is the same thing. If you wanted to determine the position of a satellite, whether you are doing it with a one-way or two-way antenna makes no difference. You are forced to an extremely narrow-band for at least final detection in order to get the end answer. The more stable the oscillator, the further you can track This is essentially what's being stated. For example, if we want to send information from Mars it seems unreasonable to ask that it come in at a 6-megacycle rate. If we are just trying to find out where the beacon is we could probably accept data once every several years and be satisfied if we finally find the answer. The idea here is that things happen very slowly. Astronomers compare over centuries. The bandwidth here is very small. Now in a radio star you have a noise source of very small diameter. You can measure to within a fraction of its diameter and it can be a noise source. You could also transmit a very wide bandwidth from a satellite providing you decode it into an exceptionally narrow band. You can do the same thing using the radio noise source. You can transmit extremely wide band (of the order of megacycles) providing you knew everything that's going on within the megacycle-bandwidth to the equivalent of a very narrow bandwidth.
(Ref. Lieutenant Wells' question).
There is also a question here on Microlock by someone from the Air Force. You were correct.
When I was referring to the use of extremely stable clocks of the cesium type or the ammonia inversion type, then I was talking about using these aboard space vehicles as clocks for time. Bearings on extra-galactic sources or the sun could be used along with the clock much as a navigator aboard ship tells his position at sea. I would like to just mention in connection with the resolving power of the radio apparatus versus the optical equipment that it is well-known that the optical theodolite has much more resolution than the radio type which varies in resolving power as 70 (λ/d). If you make a calculation you will find that it can't compare with the d optical. Finally, I would just like to throw a little more coherence into the question of what is being radiated from stars in our own galaxy and from extra-galactic sources. Namely that there is indeed black body radiation which decreases towards the 20 meters inversely as the square of the wavelength but there are also some places where you are talking about colliding galaxies such as Cygnus A. There is also much hydrogen HI which gives rise to the statistically fluctuating generation of 21 centimeter incoherent radiation. Finally, because of the differential rotation of the spiral arms of these nebulae there is synchroton radiation of some kind which is not decreasing as the wavelength increases but rather is going up as λ2. This is what Dr. Potter meant when he suggested that for intercommunication between stars in our own galaxy, or our planet and other elements of our galaxy that we want to use short wavelengths in the radio region.
Dr. von Karman:
Thank you very much, Dr. Potter and members of your panel.