Amsat-NA Logo

Proceedings of the 11th annual Amsat Space Symposium, Dallas, Texas, USA, 1994. 10 pages.
Satellite Operator No. 37, October 1993. p46-55
Amsat-VK Newsletter No. 104/5/6/7, 1993 Nov - 1994 Feb
Oscar News (UK) 1993 Dec No.104 p16-21 & 1994 Feb No.105 p18-23
The Amsat Journal (USA) Vol 17 No. 1, Jan/Feb 1994.
Amsat-DL Journal (D), Jg. 21, No. 2, Jun/Aug 1994.
Jamsat Newsletter (JA) 1994 Mar - 1995 Apr
BelAmsat Newsletter (ON) 1994 Dec; reprints of Oscar News and A-DL.

Managing Oscar-13

by

James Miller G3RUH


Abstract

1. AO-13 as a Control System

It is instructive to view the management of Oscar-13 in terms of a feedback control system, figure 1. Objectives, which are tempered by known disturbance factors, are used to formulate a control strategy. The strategy is implemented by means of commands transmitted to the satellite.

The spacecraft itself responds in a deterministic way to the commands and environmental disturbances, but in a random way to the much smaller unknown factors, mainly drag effects. This determinism is fortunate, since it means outline plans can be formulated several years in advance, and potential problems anticipated very early.

The satellite tells us about itself via the telemetry beacons, and the received data is then subject to interpretation. This analysis of the actual (as opposed to expected) spacecraft performance can then be compared with the original objectives, and the control cycle repeats.

 Control

Figure 1. Factors on the right hand side are only accessible via radio links. The left hand side is represented by pieces of paper, telemetry decoders, computer analysis programs, telephone, fax, e-mail and of course command stations or managers.

From a practical point of view, having people as an integral part of the control mechanism introduces little degradation because the system dynamic is very slow indeed, measured in days rather than milliseconds. Thus it is sufficient to check out the telemetry no more than once a week or so - although for the author it is, by choice, a daily operation.

So this review paper takes that model as a guide, and the various topics will be discussed in this context:

2. Objectives
3. Disturbances
4. Control Strategy
5. The Oscar-13 Satellite
6. Telemetry
7. Interpretation
8. Telecommand

2. Objectives

The AO-13 control objectives are remarkably concise:
  • Healthy battery charge
  • Optimise orientation
  • Optimise transponder availability
  • Publish information

    These factors are interlinked. The primary task is to keep the battery healthy, otherwise there could be no transponder usage. But this requires the satellite's orientation with respect to the Sun and users on Earth to be controlled in the best way.

    This in turn dictates which transponders should be on during which part of the orbit, because some of the antennas have a more narrow beamwidth than others. With the demise of the 435 MHz transmitter in May 1993 this requirement was somewhat simplified because both remaining transponders, mode-B and mode-S can be operated simultaneously for long periods.

    A healthy battery is one that a) retains a net positive power balance, and b) does not drop below the threshold of 12.6 volts at any time; fully charged is 14.5 volts.

    The preferred satellite orientation is that which points the antennas at the Earth when the satellite is at maximum distance, or apogee. However this ideal can only be achieved for 20 weeks of the year, and at other times up to 60 of off-pointing is required. This is no great disadvantage, since the satellite's range during the best part of these periods is still substantial, giving good Earth coverage.

    3. Disturbances

    In the control sense, there are several factors which change the satellite's circumstances and require action if the objectives are to be realised.

    Known Factors

  • Sun's annual movement
  • Eclipses of the Sun by Earth and Moon
  • Orbital precession

    Unpredictable

  • Random torques
  • Radiation effects
  • Failures

    3.1 Known Factors
    Oscar-13 is spin stabilised at typically 25 r.p.m. In principle it points towards the same spot in the heavens all the time, just like a gyroscope. Therefore from the satellite's point of view, the Sun is continuously moving around at a rate of 1°/day, and consequently sunlight strikes the solar panels at a slowly changing angle.

    At best, the Sun is perpendicular to the spin axis, and gives maximum illumination. At other times it can be up to 45° off axis, resulting in Cos 45° = 0.71 or 71% illumination. This represents the safest minimum with the mode-B transponder ON for half an orbit.

    When the Sun's position lies close to the satellite's orbit plane, Earth casts a shadow across the satellite's path. When the satellite then passes through this point it is cut off from sunlight, a condition known as eclipse and the battery receives no charge. Oscar-13 experiences eclipses for typically 20 minutes per orbit most of the year at perigee. For a few weeks eclipses occur at apogee, and can last nearly three hours. Short eclipses can easily be accommodated by the battery, long ones cannot, necessitating transponder closedown for part of the orbit.

    Occasionally the Moon interposes between Sun and satellite, and this too needs to be anticipated. Periods can be a few minutes and a partial obscuration, to a total eclipse of several minutes and an overall encounter lasting an hour or more.

    While the spin axis orientation is fixed in space, the orbit plane is not. It drifts very slightly. The plane rotates around Earth's equator at -0.17°/day, and the apogee-perigee line moves round the orbit plane at 0.07°/day (in 1993). So the spin axis direction, with which the antennas are coaxial, as seen from the orbit plane appears to drift slowly by a degree or two per week. This calls for occasional correction.

    3.2 Unpredictable factors
    The forgoing known factors can all be explicitly accounted for in the management plan. There are however some less predictable effects. As the satellite swoops past Earth at perigee, even though the altitude is well over 500 km in 1993, it receives a slight buffeting. This shows up as a) a steady fall in spin rate (-0.06 r.p.m./day in 1993) which is consistent over time, and b) a random wander in spin axis direction by as much as 1°/month. This might seem to be a trivial matter, but the whole basis upon which measurement of the spin axis direction is founded is its property of gyroscopic invariance, so considerable judgement is needed when interpreting sensors' data.

    There is also the effect of radiation. This is detectable as "hits" on the flight computer's memory, but this is Hamming-code protected, and so self correcting. Hits on the computer itself appear to have resulted in just two software crashes in the 5 years since launch. In the longer term, solar panel output will degrade, but presently this effect is not really discernible.

    3.3 Failures
    There have been (up to 1993) only two failures aboard Oscar-13. The RUDAK digital module refused to boot up properly immediately after launch, despite a year's exhaustive on-air testing from a water tower in Ismaning, Germany. For a short while it responded to short test programs, but soon that ceased also. No explanation has been found; a cracked PCB trace or IC failure is suspected.

    The 435 MHz transmitter failed abruptly on 1993 May 19, thus eliminating mode-L and mode-J operation. Tests with large EME antennas and DSP techniques have so far not heard the exciter stages, but this work continues. [2]

    4. Control Strategy

    Given the mission objectives and environmental influences, a strategy or plan has to be drawn up to realise these goals. This is by far the most time consuming job in AO-13 management. Factors to be considered are:

  • Solar illumination vs attitude
  • Attitude drift
  • Squint plots
  • Schedule planning
  • Magnetorque simulation

    To examine these factors the command stations use a suite of analysis programs evolved by the author since 1984.

    4.1 Solar Illumination vs Attitude
    The most important influence on Oscar-13 is the Sun angle. This is the angle the Sun makes with the satellite's spin equator. See figure 2. Maximum illumination is received when this angle is zero, while if it were 90° Sun and spin axis would be coaxial, and there would be no illumination at all.

     Sun angle

    Figure 2. The most critical environmental factor is Sun Angle, defined as illustrated.

    The spin axis direction has to be controlled so as to maintain a favourable Sun angle at all times. To investigate the value of Sun angle as a function of spin axis direction, also known as "attitude", program ILLPLAN is used, and figure 3 shows a typical screenshot.

    1994 Apr 18 [Mon]  SEL= -33.8   SAZ= 158.9  WP= 338.2  RAAN= 257.5  IN= 58.0
    
    LA\LO| 260 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100  90
    -------------------------------------------------------------------------------
      15 | -17  -9  -1   7  14  21  27  33  37  40  41  41  38  34  29  23  16   8
      10 | -15  -6   2  10  17  25  31  37  42  45 +++  45  43  38  33  26  19  11
       5 | -12  -4   5  13  21  28  35  41 +++ +++ +++ +++ +++  43  37  30  22  14
       0 |  -9  -1   7  16  24  31  39  45 +++ +++ +++ +++ +++ +++  40  33  25  17
      -5 |  -6   2  10  18  27  35  42 +++ +++ +++         +++ +++  44  36  28  20
     -10 |  -4   5  13  21  29  38 +++ +++ +++                 +++ +++  39  31  23
     -15 |  -1   7  16  24  32  40 +++ +++                     +++ +++  42  34  26
    -------------------------------------------------------------------------------
                                      Figure 3.
    
    Spin axis direction is conventionally defined in terms of an attitude "longitude" around the orbit plane, starting at perigee direction = 0°, apogee direction = 180° and so on, whilst direction out of the plane is called attitude "latitude". Figure 3 tabulates latitude at the left hand side, and longitude across the page, so the presentation is map-like.

    For any combination of attitude longitude/latitude a number is displayed, which is the Sun angle that will be experienced. If that number exceeds +/-45° a +++ or --- is displayed instead; this indicates a no-go zone due to low illumination. If outside +/-60° a blank is shown. At those attitudes the on-board navigation sensors will fail as they cannot see the Sun. The blank area is of course the direction of the Sun; its longitude Saz and latitude Sel is printed at the top.

    By pressing the < or > keys this table can be instantly indexed backwards or forwards by 7 days, and so provides a dynamic cartoon film of Sun angle vs attitude prospects.

    The nominal Earth-pointing-at-apogee condition is longitude 180°, latitude 0°, i.e. the middle of the chart. In the example illustrated this would result in a +++ or no-go condition. So the satellite would have to be oriented either to the left or the right of centre, for example at longitude 210° or longitude 120°. The choice is made from considerations of squint angles for northern and southern hemisphere users (see section 4.3), and favours 210°. In this example the no-go zone moves vertically up the diagram some 1° per day, and allows a nominal 180°/0° spin axis direction to be restored on July 11th.

    By liberally exercising this program a sequence of orientations can be chosen as much as two years in advance which will guarantee favourable attitude and Sun angle combinations.

    4.2 Attitude Drift
    A program used in conjunction with the forgoing is called ATTHIST. Once an orientation of the satellite has been achieved it is left untouched for as long as possible. Reorientations are time consuming, and confusing for users. A typical analysis is shown in figure 4.

    DATE               ALON   ALAT     SA  ILL %    SEL    SAZ  Arg P   RAAN
    --------------------------------------------------------------------------
    1994 Jul 11 [Mon] 180.0    0.0   36.4   80.5   50.9  199.9  343.8  242.9
    1994 Jul 18 [Mon] 180.2    1.0   29.8   86.8   57.4  206.7  344.3  241.7
    1994 Jul 25 [Mon] 180.3    2.0   23.1   92.0   63.4  216.3  344.7  240.4
    1994 Aug  1 [Mon] 180.5    3.0   16.5   95.9   68.5  230.6  345.2  239.2
    1994 Aug  8 [Mon] 180.7    4.0    9.8   98.5   71.7  251.2  345.6  238.0
    1994 Aug 15 [Mon] 180.8    5.0    3.1   99.9   72.0  276.1  346.1  236.8
    1994 Aug 22 [Mon] 181.0    6.0   -3.6   99.8   69.3  298.1  346.6  235.6
    1994 Aug 29 [Mon] 181.2    7.0  -10.3   98.4   64.5  313.7  347.0  234.4
    1994 Sep  5 [Mon] 181.3    8.0  -17.0   95.6   58.5  324.1  347.5  233.1
    1994 Sep 12 [Mon] 181.5    9.0  -23.8   91.5   51.9  331.3  348.0  231.9
    1994 Sep 19 [Mon] 181.6   10.0  -30.6   86.1   44.9  336.5  348.4  230.7
    1994 Sep 26 [Mon] 181.8   11.1  -37.4   79.5   37.6  340.4  348.9  229.5
    1994 Oct  3 [Mon] 181.9   12.1  -44.2   71.7   30.2  343.6  349.4  228.3
    --------------------------------------------------------------------------
                                 Figure 4
    
    Given an initial orientation this gives a week by week breakdown of key data, in this case the familiar Earth-pointing-from-apogee condition longitude 180° latitude 0°. The Sun angle is of greatest importance, and is seen to start at +36°, fall through zero, reaching the nominal limit of -35° in late September, dictating a re-orientation.

    Other parameters of interest are the percentage solar illumination (100 Cos SA), the Sun's position in longitude (SAZ) and latitude (SEL), and the familiar keplerian elements argument of perigee and RAAN. An understanding of the Sun's position is essential in management as it influences the ease or difficulty of attitude measurement via the on-board sensors, though it is of no importance from the user perspective.

    4.3 Squint Plots
    Once an attitude schedule has been fixed, driven only by Sun position constraints, the next task is to devise a transponder mode-schedule. [14]

    By far the most important quantity in Oscar-13 communications is known variously as the pointing or squint angle. This is the angle between the spacecraft's antennas and the user. See figure 5. If this angle is large, communications are poor quite regardless of anything else. So a study of squint angle forms the next stage of the planning process.

     Squint angle

    Figure 5. 'Squint' is the angle between the spacecraft's antennas and the user.

    Some of the spacecraft antennas have a narrow beamwidth, notably the 1269 MHz receive which is a 5 turn helix, and the 2400 MHz transmit, a 4 turn helix. Their -3dB beamwidths are of the order of +/-20°, so there is little point in seriously exercising them when signals arrive or depart at angles worse than this.

    On the other hand, there are also omni-directional whips for 145 and 435 MHz, and these must be used sideways on and at short range.

    With this in mind program SQPLOT is used to determine the pointing angles as a function of time through each orbit for a representative 10 day period. This can be done for a typical user at 45° N or 35° S, and for any specified satellite orientation.

    A typical plot is shown in figure 6. Squint is the vertical axis, while the horizontal is mean anomaly in 1/256th orbit period. Superimposed on the plot are candidate periods for mode-L and mode-S operation. (Since the demise of mode-L, that time is now allocated to mode-S).

     Squint plot

    Figure 6. Plots of squint angle as a function of MA are used to devise the transponder mode schedule.

    In this way a mode switching schedule can be devised, and is eventually published in the familiar format:

        L QST *** AO-13 TRANSPONDER SCHEDULE *** 1993 May 10 - May 31
        Mode-B  : MA   0 to MA 130 !     Omnis MA 250 - MA 60
        Mode-BS : MA 130 to MA 180 !<- S transponder; B trsp. is ON
        Mode-S  : MA 180 to MA 190 !<- S transponder; B trsp. is OFF
        Mode-LS : MA 190 to MA 195 !<- S beacon + L transponder
        Mode-JL : MA 195 to MA 210 !        Alon/Alat 210/0
        Mode-B  : MA 210 to MA 256 !  Move to attitude 120/0, May 31
        Please don't uplink to B, MA 180-190. Interferes with mode S.
    
    4.4 Magnetorque Simulation
    Re-orientation of the satellite from one spin direction to another is achieved by an on-board system called a magnetorquer. Wound around each of the three spacecraft arms is a large solenoid. These three coils form the rotor of an electric motor; the "stator" is the Earth's magnetic field. Commutation is provided by the on-board computer which switches current through the coils sequentially in synchronism with the spin.

    When the Earth's magnetic field lies on the satellite's spin equator we have a pure motor, in that spin rate can be increased or decreased or simply maintained. If it lies off-equator then the resulting non-coaxial torque allows the spin direction to be changed as well.

    Magnetorquing is strongest when the Earth's magnetic field is strongest, and so only takes place at perigee encounter, essentially within MA = +/-10. During fly-by the field also changes greatly in direction, and this ensures a rich profile of torques can be generated. Typically a movement of 5° can be achieved per perigee pass. However during eclipses, the spacecraft is deprived of its reference direction the Sun, so magnetorquing is impossible at that time.

    Program MAGSIM output is typified in figure 7. Once again the orientation coordinate system of orbit plane longitude and latitude is used. As the simulation progresses the interim orientation is plotted. Each perigee instant is marked. This example has been exaggerated for clarity.

     Magnetorque Simulation

    Figure 7. Prior to actually commanding a magnetorque, a ground based simulation is essential to find out how many perigee's worth of torquing is required, the impact of eclipses and to investigate trade offs between spin rate and re-orientation rate. Also, the spacecraft's Earth modelling is a simplification, and simulation can account for this.

    Across the top of the screen a running display of the most important numbers is maintained. The first line gives time in utc, keplerian elements RAAN and argument of perigee (AP), Sun angle (SA), orbit number and MA. The quantities below the line are columns containing certain vectors expressed in orbit plane (German = Bahn) coordinates. SUN is the Sun's direction, SSV and SV are target and present spin vector as from the spacecraft's on-board magnetorquing algorithm, whereas the column SVr is the real spin vector determined by simulation. Hr, Mr and TQr are the Earth's magnetic field magnitude and direction, spacecraft magnet direction and resultant torque ( TQr = Mr x Hr ). M-SOLL is the angle on the spin equator between the magnet direction and the Sun. The flight computer uses this to commutate each of the three coils at the right moment.

    Over the years this simulation has been refined and calibrated to the point where complex reorientations can be planned and executed with considerable confidence. Errors amount to only 0.2 r.p.m. and a degree or two after a dozen perigee passes and a net movement of 60 deg. These errors are mainly due to short term variations in the Earth's magnetic field. Finally the program automatically generates the 7 number command sequence that is to be sent to the Oscar-13 on-board computer.

    4.5 Other planning programs
    ECLIPSE calculates the time and duration of eclipses by the Sun, and is used to determine when temporary changes to the transponder schedule have to be made. MOONECL does the same thing for the Moon. SMOOTH13 takes many sets of Norad keplerian element sets and smooths them, resulting in a secondary set of elements that is free of the short term luni-solar variations and measurement noise that is so readily apparent [5]. About 6 month's past data, typically 20 sets, is processed at once. The smoothed set is used by all the management programs, rendering them consistent and accurate. The set is also converted into Oscar-13's on-board computer format for direct uploading.

    5. The OSCAR-13 Satellite

    The spacecraft comprises a number of subsystems [1,3] including:
  • On board computer or IHU
  • Transponders B, S, (J and L failed)
  • Solar/battery power system
  • Sun and Earth sensors and SEU
  • RUDAK-1 (failed)
  • Propulsion system (spent)
  • Telemetry system and beacons
  • Telecommand system

    The on-board computer or IHU (integrated housekeeping unit) orchestrates spacecraft activities. Based on a Cosmac 1802 processor it has 32k of radiation hardened RAM, and input/output interfaces for analogue and digital services. The flight software turns the transponders on and off as directed, switches antennas, monitors the battery systems and navigation sensors. Finally it generates telemetry and accepts telecommands from the ground.

    6. Telemetry

    This is transmitted continuously; there are text bulletins in CW and 50 baud RTTY, but the most important data is sent as PSK at 400 bps or 50 byte/s. This rate was selected (in 1978) because the system fundamental timing is a 20 ms clock, and it is also properly matched to the limited beacon power using popular ground station receive antennas.

    Data is sent in 512 byte blocks, preceded by a 4 byte sync code, followed with a two byte CRCC or checksum and lasts 10.24 seconds. Blocks, interspersed with a few seconds of idle code hex50, consist of engineering data and plaintext messages alternately. A complete description of the telemetry can be found in [3].

    The information available is:

  • General status
  • Navigation
  • Power
  • Temperature
  • Messages

    Decoders have been available since at least 1984 [6,7]. Audio enters one end, RS232 data comes out the other and the data can be displayed on virtually any PC. This is a routine no-skill operation enjoyed by perhaps as many as 500 operators. Taking telemetry is not the exclusive domain of command stations, and never has been.

     Telemetry display

    Above: 'Q' block Status and 'M' Message. Below: 'Q' block Navigation and 'N' Message  Telemetry display

    Figure 8. Typical screens from the author's display program. The message blocks are displayed in sequence as they arrive, and can be saved to disc or printer. The lower window can be cycled between Status, Navigation, Power and Temperature data. Windows showing a small square in the top right corner arrived uncorrupted; those with a cross failed their CRCC checksum test and therefore contain errors. One is apparent in the message that begins "N de VK5AGR ".

    7. Interpretation

    Once telemetry has been received it can be analysed in detail if required. This is particularly important during a re-orientation, to ensure that nothing is going wrong, and goals are being achieved. Effort goes mainly into studying:

  • Whole orbit data WOD
  • Battery state
  • Attitude determination
  • General monitoring

    7.1 Whole orbit data WOD
    There is a facility to dwell on one telemetry channel and sample it at any interval from 1 to 256 MA counts until 384 points have been collected. This means that more than a whole orbit's data is acquired each time, so "WOD" is a slight misnomer. The data is telemetered in one of the message blocks.

     WOD display

    Figure 9. A typical dwell would be on the battery voltage. Here the 384 points of numerical data have been plotted using program WODDISP and give a profile of 1.5 orbits activity. This plot has been annotated to show which transponders were in use, and to highlight perigee eclipses. The heavy mode-L loading is apparent. The danger threshold is 12.6 volts, and though not reached, shows why mode-L time was relatively restricted.

    Frequent plots such as this have enabled the command stations to predict the exact battery condition under all conditions of transponder loading and solar illumination, and so maximise the facilities on offer at all times.

    7.2 Attitude Determination
    Oscar-13 carries two optical sensors which are used to determine the satellite's attitude. They are mounted on the end of one arm, and as the satellite spins they scan the spin equator for their respective targets. The Sun sensor measures the Sun angle, (see figure 2), while the Earth sensor responds when the Earth comes into view, generally within the hour or two around perigee.

    The art of position finding is centuries old. The principles are much the same whether you are surveying the backyard or aiming a satellite. All you need is two or three objects whose position IS known. You make an observation about each of them. Then you plot your position on a map; it's the only place that could give rise to those observations. A simple two-dimensional analogy will make this clear.

     Trivial attitude fix

    Figure 10. You have a map of a room, you are somewhere in the room and want to plot your exact position on the map. The obvious way to do this is to measure the distance to a known point, say a corner, and then the distance to a second known point. Then draw arcs of these distances about each point, and your position must lie at their intersection.

    Of course two circles intersect in two places, and you need to resolve this. Frequently the alternative is unreasonable, but you could easily measure the distance to a third point.

    It's just the same with measuring a satellite's position. You have a map of the sky. You are on-board the satellite and want to plot the spin axis direction on the map. So you measure the angle between the spin axis and two known points in the sky. These could be a star, the Sun or the Earth.

    On the map one draws the positions of the known objects, and draws the two arcs. The spin axis direction must lie at their intersection. The ambiguity can be resolved with the help of a body in a third position, or by some reasonableness criterion. This is illustrated in figure 11.

     Trivial attitude fix

    Figure 11. By measuring the angles between a spacecraft axis and two objects, intersection of two arcs as shown locates the direction of that axis.

    As indicated earlier, Sun and Earth sensors measure the angles for this purpose. Figure 12 shows the resulting plot.  Attitude fix (spherical)

    Figure 12. The "map" is in orbit plane coordinates, that is longitude 0 to 360, and latitude -90 to 90. A Mercator style square map (e.g. like figure 7) is very clumsy to work with; a spherical presentation is more appropriate since we are entirely concerned with angles rather than distances.

    The measured data is superimposed, and although two intersecting arcs would suffice to determine the spin axis direction or attitude, in practice one takes a number of measurements over several days. To make this clearer, figure 13 shows the centre magnified x4. At this scale it can be seen that the data is indeed slightly noisy, hence the need for confirmatory measurements.  Attitude fix (spherical) x4

    Figure 13. The central portion of fig.12 magnified x4. Data from Sun and Earth sensors is plotted.

    The program which generates these plots is called ATTPLOT, and takes as its input date/time and value of Sun and Earth sensor measurements. It takes 4 seconds to compute and draw. Recently the author discovered a means of measuring squint or pointing angle from observations of the S-band beacon doppler shift [8] (the antenna is mounted eccentrically), and this data can also be input and plotted.

    Another program called ATTFIX takes the same data set as the plotting program, and computes the least squares solution numerically that is depicted in figure 12 graphically. This is useful for analysing the fine structure and jitter properties of the Sun and Earth sensors, and indeed has been used to calibrate the mounting axis of the Earth sensor itself. The telescope axis has been found to be misaligned by about 2.5° upwards, i.e. toward the antennas, and its two beams are separated by 10.14° rather than the specified 10°. These may seem like insignificant amounts, but small uncertainties in measuring the spacecraft attitude take a disproportionate amount of time (i.e. days) to resolve.

    Sun sensor data is quantised to 1 part in 256. This equates to about 1° resolution. Thus a spot measurement is likely to have up to +/-0.5° error in it, as well as added random noise. The whole orbit data collection facility helps resolve this.

     Smoothing Sun data

    Figure 14. Shows the Sun sensor data gathered over a 12 day period. The quantisation and random noise is clearly apparent. A least squares fit is drawn through the data, and allows extraction of extremely accurate values of Sun angle.

    Using this technique, it has been found that the spin axis direction, which by gyroscopic principles ought to be a rigidly fixed direction, does have a tendency to wander over a period of a month. This is attributed to the present (1993) low perigee height.

    The many techniques developed for Oscar-10 and Oscar-13 attitude determination have been published, and the interested reader is directed to the references [11,12,13]; [15] is outstanding.

    8. Telecommand

    Oscar-13 is a computer in space; its keyboard and screen merely happen to be on Earth. Thus one composes instructions on the ground, transmits them to the spacecraft, and watches telemetry for the response. This round-trip takes about 30 seconds.

    Just as a home computer will be running a program that creates an environment (perhaps BASIC), so Oscar-13 is running a language/environment called IPS, Interpreter Structure for Processes. [4]

    8.1 IPS Interpreter
    The IPS interpreter offers a plain language environment not unlike FORTH. Instructions are accepted, parsed, compiled and then executed.

    Essential Primitives are provided. e.g.

    Arithmetic   0...9  #  +  -  *  /
         Relational  =  <  >  <>   etc
         Logical  NOT AND OR EXOR BIT
         Peek & Poke bytes, words, fields
         Manipulate stack (Push Pull Swap Dup)
         FOR...NEXT   DO...WHILE   REPEAT...UNTIL
         An 1802 assembler
    

    All other objects defined by user

    Variables & Constants
         Functions/Procedures/Routines
    

    Examples of commands; entering

    2 3 +
    would result in the response 5. The command sequence
    : SQUARE DUP * ;
    defines a new function which will be compiled and ready for global usage. So a subsequent command
    9 SQUARE
    would result in the response 81.

    8.2 Software Installation

  • First the flight computer is reset.
  • Next a short self testing loader is uplinked
  • Then the IPS interpreter is uploaded (only 7200 bytes)
  • Finally the flight routines are uplinked (36 kbytes).

    When the flight routines are running they serve simultaneously:

  • 20 ms loop analogue multiplexer
  • 20 ms clocks
  • Telemetry buffering
  • Spacecraft control. Interpreted and then compiled code.

    8.3 Flight Routines
    This is approximately 10 pages of plain language, and is compiled as it is received by IPS. It then totals about 16 kbytes including buffers etc.

    At the highest level it is organised around a chain. Calling this is the very last command sent:

     0 EINH SERVICE   ( Battery service, watchdogs )
     1 EINH BAKEN-ST  ( Beacon sequencer  )
     2 EINH NAV-ST    ( Navigation )
     3 EINH MEL-ST    ( Memory errors )
     4 EINH UL-ST     ( Transponder scheduler  )
                      ( 5 - spare )
                      ( 6 - spare )
     7 EINH MODE-S-ST ( Mode-S control )
    
    The program is "merely" cycling around these eight tasks. Each is defined elsewhere, and of course invokes other definitions. For example, item 2 on the chain, "Navigation start" is defined below, and calls in turn NAVIGATION, SENSOREN and TRQ-ST:
    : NAV-ST  NAVIGATION SENSOREN TRQ-ST ;
    
    : NAVIGATION   Z-ALT @B Z @B <> DUP Z-MARKE !
      JA? Z @B DUP Z-ALT !B =0
          JA? RECTAS @ KO + RECTAS !
              PERIGA @ KW + PERIGA !
          DANN ORTSDATEN MAGNET @B
          JA? A-CONTROL
          DANN
      DANN ;
    
    : SENSOREN Z-MARKE @
         JA? #447 @ SS>BETA BETA !
             #44D #44F JE I 6 + @B
             JA? 0 I 6 + !B Z @ I !
             DANN 2 +NUN
         DANN ;
    
    : TRQ-ST Z-MARKE @
        JA? E-FLAGS @ #14 UND >0  Z @
            MZEITGRENZE @ = ODER
            JA? 0 M-EIN !
            DANN Z @B 32 + #FF UND 64 < M-EIN @
                 UND 1 UND MAGNET !B
        DANN ;
    
    8.4. Commanding
    Day to day commands are required to update Message blocks (which can be read on the beacons in PSK, RTTY and CW), invoke magnetorques, check sensors, invoke WOD data collection, change transponder mode schedules and so on. Occasionally 0 TRANSPONDER (transponders off) is needed ...

    Commanding is a very small part of Oscar-13 management. It is not always straightforward. The 435 MHz command uplink is interfered with by radar, and man-made noise invariably ruins 145 MHz downlink telemetry, especially at poor pointing angles and at maximum range. Mode-B commanding is tiresome at best.

    Fortunately the 1269 MHz uplink is still available and needs only a few watts e.i.r.p., and the 2400 MHz S-band downlink is very strong on compact antennas [9,10] and error free. So provided the mode-S facility is exercised daily, which it is, an excellent command link is assured.

    8.5 Reliability
    Crashes are very infrequent. One initial load, and five reloads have taken place as follows:

                                     Total   Life   Rate     Crash
     Load    From           To       Commands Days  Comms/day Notes
     --------------------------------------------------------------
      1  1988 Jun    -  1989 Oct 09   1173     480   2.44      *
      2  1989 Oct 09 -  1989 Oct 28    390      19  20.53      rt
      3  1989 Oct 28 -  1989 Dec 10    224      43   5.21      *
      4  1989 Dec 10 -  1991 May 13    759     519   1.46      pt
      5  1991 May 13 -  1992 Jan 29    248     261   0.95      rs
      6  1992 Jan 29 -                 934    1082   0.86      (*)
                                      ----    ----  -----
                               Total  3728    2404   1.55
    
    
     Notes
      *   reason unknown - probably radiation hit on CPU
      rt  after intensive RUDAK testing; probably an induced mistake
      pt  investigation of navigation coding error induced a "poke"
           with its address and data transposed.
      rs  reset command inadvertently sent.
      (*) as of 1995 Jan15
      A reload uses 70 commands.
    
    8.6 Coding
    The COSMAC 1802 IPS interpreter was designed and implemented by Dr Karl Meinzer DJ4ZC [4] in the late 1970s. The flight software was also written by Dr Meinzer, and has been edited and added to by Peter Guelzow DB2OS and the author.

    9. The Future

    Oscar-13 management has essentially stabilised, and a regular diurnal rhythm is well established. The spacecraft is "calibrated" to a high degree, and is extremely reliable.

     Keplerian evolution

    Figure 15. It is now well known that the orbit is unstable in that the eccentricity slowly oscillates. Unfortunately in 1996 perigee passes (Hp) will strike the atmosphere causing the orbit to decay rapidly in December. This has been documented by several authors; [5] contains a full bibliography. It will be an interesting period.

    10. CONCLUSION

    Many command stations have contributed to Oscars 10 and 13 management at one time or other. These include DB2OS, DJ4ZC, DK1YQ, G3RUH, VE1SAT/VE6, VK5AGR, W0PN, W3GEY, ZL1AOX. Many more potential command stations work or have worked behind the scenes. Some have pulled out to work on other satellites or in other fields. For some their work is over (e.g. RUDAK specialists), or they simply retire. A thousand or more hours a year is a lot to give.

    The present day command team consists of DB2OS, VK5AGR and the author G3RUH. Having a transmitting station in each hemisphere is important, but geographic location is otherwise irrelevant. In fact it is quite possible to have non- transmitting command stations, since others can do the actual uploading.

    The core team is in daily and immediate contact via telephone, fax and e-mail. Contact with the user community is via similar routes, the post and by listening to the transponders. All significant decisions are made jointly, and by common consent can be vetoed by the others.

    Anyone can do this job if they show the level of commitment typified by the people whose callsigns are listed above. These folks are self-motivated, self- trained, self-starting and self-sustaining.

    There is absolutely no "how-to" manual, and no training course. They have a strong inquisitive streak, and an ability to work things out from scratch unaided. There is no glamour, and no "power". Far from it. They do it because it is interesting.

    References

    1. Davidoff, M; The Satellite Experimenter's Handbook, ARRL 1990. ISBN 0-87259-004-6. Appendix B-3.

    2. Emerson D.; Digital Processing of Weak Signals Buried in Noise, Proceedings of the 11th Annual Amsat Space Symposium, Arlington TX, USA, 1993.

    3. Limebear R. (Ed); The Oscar-13 Operations and Technical Handbook, AMSAT-UK 1989.

    4. Meinzer K.; IPS, An Unorthodox High Level Language, BYTE, January 1979, pps 146-159. Book from author.

    5. Miller J.R.; The Re-entry of Oscar-13,Proceedings of the 12th annual Amsat Space Symposium, Orlando, Florida, USA, 1994.

    6. Miller J.R.; Phase 3 400 bps PSK Data Demodulator MK II, Oscar News (UK) No. 98, December 1992. PCB available from author.

    7. Miller J.R.; Telemetry Decoder for Oscar-10, Electronics and Wireless World, (UK), 1984 October, pps. 37-41, 59-60. Part 2, 1984 Nov, pps.37-38. Also: Ham Radio Magazine, (US) 1985 Apr, pps. 50-62. Now superceded by Mk II version.

    8. Miller J.R.; Measure AO-13 Squint Directly!, Oscar News (UK) No. 99, February 1993. Also: The Amsat Journal (USA) Vol 16 No. 1, January 1993. Also: Amsat-DL Journal (D), Jg. 20, No.. 1, March 1993.

    9. Miller J.R.; A 60 cm S-Band Dish Antenna, Oscar News (UK) No. 100, April 1993. Also: The Amsat Journal (USA) Vol 16 No. 2, March/April 1993. Also: Amsat-DL Journal (D), Jg. 20, No.. 2, Jun/Aug 1993.

    10. Miller J.R.; Small iS beSteSt, Satellite Operator No. 33, June 1993. (S-band 16 turn S-band helix antenna).

    11. Miller J.R.; Oscar-10 Attitude Determination, Proceedings of the 4th annual Amsat Space Symposium, Dallas Texas, USA, 1986. pps 20-34

    12. Miller J.R.; Sun's Up Part 1, Oscar News (UK) No. 50, 1984 Dec; Part 2, No. 51, 1985 Feb; Part 3, No. 52 1985 Apr.

    13. Miller J.R.; Sensorship - A Question of Attitude, Part 1, Oscar News (UK) No. 54, 1985 Aug; Part 2, No. 55, 1985 Oct.

    14. Miller J.R.; Planning AO-13 Mode Schedules, Oscar News (UK) No. 77, 1989 June.

    15. Wertz J.R. (Ed); Spacecraft Attitude Determination and Control, D. Reidel, 1984. ISBN 90-277-1204-2.


    CREDO
    Discovery is easy; the difficulty is to acquire what we discover.
    Roberto Gerhard, composer 1896-1970.

    Feedback on these pages to KB5MU. Feedback on the article should be sent to James Miller

    Created: 1995 Jan 15 -- Last modified: 2005 Oct 29