Amsat-NA Logo

Amsat-UK's Oscar News, 1997 Feb No.123 p 24-29
The Amsat Journal (USA) Vol 20 No.2, Mar/Apr 1997
Amsat-DL Journal (D), Jg. 24, No. 1, Mar/May 1997
Jamsat Newsletter (JA) - In preparation

Oscar-13's Life and Death

by

James Miller G3RUH


Abstract

In the Beginning

Oscar-13 was launched from Kourou on 1988 Jun 15th on the first test flight (V-22) of the Ariane 4 rocket, along with Panamsat and Meteosat P2. Eight and a half years later AO-13 re-entered and burned up, on 1996 Dec 5th.

Lift-off was 1119:01 utc; at 1239:04 P3C became Oscar-13 when it was ejected from the carrying structure, and at 1403:38 utc sprang to life with the activation of its 145 MHz telemetry beacon. ZL1AOX reported the satellite to be in perfect health.

Its first message blocks contained a welcome in German, English, Spanish, Portuguese and French.

AO-13's first message

Figure 1. Oscar-13's first beacon message: "HI, this is AO13, the international spacecraft to support science/education, amateur space communication and above all international goodwill and cooperation".

Oscar-13 was launched into a highly elliptical equatorial orbit, a "Geostationary Transfer Orbit", with a perigee height of 240 km spinning at 5 rpm. Over the next 7 days it was spun up to 26 rpm, and re-oriented to point the motor along-track such that apogee velocity could be increased by 150 m/s. This increased perigee height to 1100 km, and freed the satellite from buffeting and atmospheric drag.

During the next 2 weeks it was spun up to 60 rpm, and oriented "upwards" so that a burn at apogee would increase inclination to 58°, and perigee height to 2500 km. The delta-V was 1314 m/s.

Two weeks later, 1988 July 22 1500 utc the mode-B transponder was switched on for general use, and the rest is history. (It was in fact discretely tested for a few minutes the previous day).

AO-13 Performance

With its apogee distance of order 40,000 km Oscar-13 gave global VHF/UHF/SHF radio coverage to thousands of operators. Apogee was initially over the equator, drifting maximally Northwards for the start of 1992, back to the equator again by 1995, and ending around 45° South two years later. For the most part, its performance was outstanding, and gave great pleasure to many people. Indeed, for some it became a way of life.

It is instructive to review the factors which made Oscar-13 so remarkable. We should look at both successes and failures. A spacecraft consists of a Bus which carries Payloads.

Payloads

  Payload       Status   Comment
  ----------------------------------------------                    
   2.4 GHz TX    OK
   435 MHz TX    OK      Failed after 5 years
   145 MHz TX    OK
     
   1.3 GHz RX    OK      Low sensitivity
   435 MHz RX    OK
     
    RUDAK        --      Never worked        
  ----------------------------------------------                    
Let's take a detailed look at the payload successes and failures.

Payload Failures

Oscar-13 carried a packet radio digital regenerator experiment called RUDAK. The engineering model was tested to great satisfaction from a water tower near Munich, and pre-launch in AO-13. But in space, to universal disappointment, it never worked, nor could it be persuaded, via diagnostic probing, to reveal any reason for not working. Intensive testing was tried twice, soon after launch, and again a year later, then abandoned.

The mode L 1269 MHz receiver system sensitivity was 10x too low. A requirement for at least 5 kw e.i.r.p. deterred most operators. Had the sensitivity been correct, 10 watts to a long yagi would have been ample, and mode-L might have been as popular as mode-B, maybe more so, because the 70 cm downlink is such a quiet band. No explanation was found.

The 435 MHz transmitter failed abruptly on 1993 May 19. So that was the end of mode-L anyway, and the best telemetry beacon. The transmitter used to run at typically 41C, which may point to the cause of its failure. However the L-band 1269 MHz RX remained to the end, used for command purposes in conjunction with the S-band downlink.

Payload Successes

The most popular mode for over a decade, the Mode-B 145 MHz transmitter was on virtually continuously, some 60,000 hours, interrupted only by typically a couple of hours at apogee for Mode L and/or Mode-S with their more directional antennas. This was a great achievement for the electronics, particularly the transmitter. The design is based on the HELAPS principle where the output signal is decomposed into its amplitude and phase components. The phase information is processed by a class D power amplifier, and the amplitude is modulated back onto the result. This technique results in a very power efficient, and therefore cool-running wideband amplifier. The HELAPS concept was pioneered by Amsat.

The S-band transponder also turned out to be a great success. It was built essentially as a technology demonstration, and as such had known limitations, low power and hard limiting. Initially perceived as a techie's toy, a bit of eloquent persuasion [1] showed this not to be the case. Large numbers of users had a go, and are now ready for what promises to be the workhorse mode for P3D. Bill McCaa K0RZ certainly deserves thanks for his farsightedness over 12 years ago, and for nursing the S-band transponder project to fruition.

The Bus

Turning now to the Bus. That's the metalwork, motors, flight computer, batteries and so on that are needed to support the payloads.
        Successes                   Failures
     --------------------------------------------         
     Thermal Design
     Motor System
     Battery System
     Magnetorquing System            NONE!
     Navigation System
     Computer System & IPS
     Command & Telemetry                         
     --------------------------------------------         
Bus Successes

The AO-13 bus has been extraordinarily successful. Virtually nothing has broken, and what has was trivial. It's an outstanding achievement, even more so when you realise that AO-13's pedigree dates from 1979 and Phase IIIA. Let me touch briefly on a few of the systems we took for granted.

Thermal design

Worked perfectly. The satellite enjoyed a benign shirt-sleeve internal environment whatever the Sun's position. There were no structural hot spots. During eclipses the solar panels reached ­40C, but the internals rarely fell below zero. It's no mean feat to achieve good thermal balance, which is based solely on juggling the radiation characteristics of the various exposed parts of the structure, by choosing reflective or absorbing coatings.

Motor system

The bi-propellant hydrazine and nitrous oxide 400 Newton motor system behaved flawlessly during the two post launch burns. Another outstanding success, and a fine tribute to its highly skilled specialist engineers.

Battery System

AO-13 carried two batteries; the main and the auxiliary or back-up. The main battery was still going very strong, and could support the spacecraft through a 2-hour eclipse. There was a little concern in 1993 when it was thought to be ailing; it just wouldn't hold charge. Then it was realised that it hadn't actually got any charge to hold, because the regulator control set points had not been altered since launch to reflect ageing of the solar panels. Ageing alters a panel's characteristic voltage-current curve quite a lot. Once these control points were altered, a matter of a fraction of a volt, the main battery came back as strong as ever.

The auxiliary battery, which was normally fully discharged, was accidentally charged up in 1991. On 1996 May 26 the auxiliary was again deliberately recharged for 24 hours to see what would happen. It stayed charged for only 7 days, and the delivered capacity appeared to be down to 0.1 A-H. So on Jun 19 it was recharged again, this time for week. Its voltage held up until the end of September. A final recharge was given on Nov 16, "just in case" it was needed during the final weeks. However it was never actually used to power the spacecraft because once invoked, switching back to the main battery would have caused a power surge which might have proved fatal.

There were two Battery Charge Regulators (BCR) on board, both of which were exercised regularly, one at a time. These are essentially 50 watt switched mode power supplies taking some 30-35 volts from the panels and delivering 14.5 volts to the battery with a very high conversion efficiency. Both performed perfectly. Another outstanding design.

Magnetorquing System

Oscar-13 needed to be re-oriented from time to time to ensure sufficient illumination of the solar panels. This is accomplished by the magnetorque coils. These electromagnets form the rotor of an electric motor. The stator is formed by the Earth's magnetic field, and commutation of the coil current is performed by the on-board computer. In this way the satellite's spin rate and spin axis direction can be changed as required. It's a delightfully simple concept, and another outstanding success.

Navigation System

Attitude determination was performed by on board measurement sensors, and by on-the-ground data processing software. Measurements were made by a Sun sensor and an Earth sensor. Both instruments worked fairly well. In recent years we had to use their data with great care in that they were prone to give unexpected results which needed to be explained, and that wasted time. There was some evidence of clouding of their cover glass, and certainly of changed photodiode sensitivities. In fact the navigation sensors were the most delicate part of the bus design, and uncertainties in their operation and the need to interpret their data with considerable skill and judgement resulted in perhaps the largest expenditure of ground station time.

Computer System, IPS, Command, Telemetry

Finally, the spacecraft's many functions were coordinated by the on-board computer system called IHU. This comprised the CPU, memory, data multiplexer, command decoder, telemetry encoder and of course the flight software itself. These things were the "heart" of the spacecraft.

The flight computer proved fantastically reliable. Based on the 1802 Cosmac processor, executing just 100,000 instructions/sec it didn't really miss a beat since launch. There have been so-called crashes, where the program stopped. Two of these were definitely known to be operator error, such as sending the Reset command by mistake, and the others occurred at times of great testing activity and are virtually certain also to have been operator error. No crash can definitely be linked with radiation hits. And more remarkably, no crash can be attributable to faulty software. I think we had here an example of that very rare object, a bug-free Operating System.

In any event, ignoring two stoppages due to simple operator error, the flight software ran faultlessly for at least 7 years. Over its whole life the spacecraft was sent 4628 commands, an average of 1.5 per day.

                                Total   Life   Rate
   Load    From         To     Commands Days  Comms/day Notes
   ----------------------------------------------------------
    1  1988 Jun 15  1989 Oct 09  1173     481   2.44      st
    2  1989 Oct 09  1989 Oct 28   390      19  20.53      rt
    3  1989 Oct 28  1989 Dec 10   224      43   5.21      rt
    4  1989 Dec 10  1991 May 13   759     519   1.46      po
    5  1991 May 13  1992 Jan 29   248     261   0.95      rs
    6  1992 Jan 29  1996 Nov 24  1834    1761   1.04      --
                                 ----    ----  ----- 
                        Total    4628    3084   1.50
                                 ----    ----  ----- 
   Notes
    st   soon after mode-S control routine testing.
    rt   just after intensive RUDAK testing.
    po   investigation of navigation problem induced a
          "poke" with its address and data transposed.
    rs   reset command inadvertently sent. 
    --   s/c failed from overheating at perigee.

Figure 2. Oscar-13 Command history from launch to failure. Over its life the spacecraft was sent 4628 commands, an average of 1.5 per day.

On May 1994 the computer memory error detection and correction circuits started flagging errors at a rapid rate, initially 3, later about 12 per orbit. Command stations could not discern whether these represented real memory corruptions, or a minor malfunction of the EDAC circuitry itself. Either way, the flight software didn't crash, so it was a benign fault, and probably due to a partial logic failure.

Command and telemetry systems proved robust and worked extremely well. The formats carry over unchanged to the P3D satellite.

The Oscar-13 Flight Computer system is an engineering tour de force which deserves unqualified praise. And at its heart is the operating system, written in a Forth-like language called IPS, which was devised by Karl Meinzer DJ4ZC in 1977-78. Its exquisite design is in fact one of Karl's greatest achievements, something that has certainly never been properly recognised.

That the same flight computer, IPS language, command and telemetry systems are built into the Phase 3D satellite presently under construction, is greatly reassuring. [2]

The End Nears

Due to the differential pull of the Sun and of the Moon, Oscar-13's orbit eccentricity began to increase in early 1994, so reducing perigee height. In late 1996 atmospheric drag started to rob the satellite of kinetic energy, and at the beginning of December the satellite re-entered, having failed a couple of weeks before due to excessive heating.

The first evidence of drag is to be found in the record of Mean Motion, which began to increase steadily from about March 01 when perigee height was about 300 km. By September, the rate of increase was large enough to cause users difficulty in keeping tracking programs in sync.

Mean Motion increases ...

Figure 3. Mean Motion - predicted and actual values.

Atmospheric heating began to be clearly discernible by October 07, when live telemetry at perigee showed a sudden rise of 5C over a 2 minute period on a solar panel. This was the first time this phenomenon, perigee heating had been recorded for an amateur satellite.

From this time onwards, the Whole Orbit Data collection was invoked at 1 MA intervals to record solar panel temperature, right up to re-entry. A typical plot taken on Nov 21 is shown. By this time the temperature sensor's limit of 79C was being exceeded almost every perigee.

Solar panel heating

Figure 4. Oscar-13 solar panel #1 temperature. Shows perigee eclipses and heating 1996 Nov 21.

The final re-orientation proposed in [3] to combat anticipated attitude changes due to perigee drag was not implemented. The observed effect was less pronounced than predicted. And because of perigee eclipses it would have taken so long to re-orient that the spacecraft would have failed mid- move. As it happened, excellent mode-B transponder performance at ALON/ALAT = 160/0 was maintained right to the end despite a generally poor Sun angle. In fact the perigee drag torques actually improved the Sun angle, holding it to around 45° (when it would otherwise have increased to a dangerous level), and finally reducing it to 37°.

Final Orbits

On perigee 6478/79, at around 1996 Nov 23 [Sat] 1140 utc Oscar-13's solar panel #3 ceased to function due to excessive heating.

On perigee 6479/80 at around 1996 Nov 23 [Sat] 2009:30 utc panels #1,2,4,6 also stopped working, leaving only panel #5. The transponders were switched off by Graham VK5AGR shortly after his AOS at 1996 Nov 23 [Sat] 2314:15 utc, orbit 6480 MA 99. Orbit 6480 continued on telemetry beacon only.

AO-13 survived perigee 6480/6481, 1996 Nov 24 [Sun] 0432 utc, perigee altitude 107 km, monitored by insomniacs in Europe. The battery voltage was stable at 11.8 volts, much lower than the normal 14.5v. The on-board computer and related systems are regulated to 10.0v.

VK5AGR monitored the remainder of orbit 6481, but one solar panel proved insufficient to sustain a basic system, and the beacon stopped transmitting at 1996 Nov 24 [Sun] 0538:16 telemetry block time, orbit 6481, MA 34. Reset commands, which would have left an unmodulated carrier, had no effect.

On orbit 6482 in Europe, 1996 Nov 24 [Sun] 1300 utc onwards, neither of the beacons was detected, and the Reset command was again ineffective. The command stations then concluded that Oscar-13 was defunct.

Inspection of the very last telemetry block shows that failure occurs just as the 133rd byte was being transmitted. The spacecraft clock was reading correctly at the end, so the exact time of death, taken to be when the flight computer failed, can be recorded as 1996 Nov 24 [Sun] 0538:08 utc.

Keplerian Elements

Keplerian element history Keplerian element history

Figure 5. Evolution of Oscar-13's Keplerian Elements from 1988 July until burn-up. The culprit is eccentricity Ec which directly affects perigee height Hp. The rate of change of eccentricity and also of inclination (In) depends largely on the values of RAAN and Argument of Perigee Wp. We were unfortunate in that their initial values combined unfavourably, with the result as shown. This subject has now received a great deal of attention in order that a similar fate shall not befall the Phase IIID satellite, due for launch in 1997 [5]. Parameters computed as described in [4].

Re-Entry

Re-entry was predicted for 1996 Dec 05/06 approximately, though experience with final orbits of decaying satellites indicates an enormous degree of uncertainty in these matters. NORAD keplerian elements were eagerly awaited daily, while mean motion rose inexorably towards 16 rev/day.

The last element set, #387, released by NORAD was date stamped 1996 Dec 05 [Thu] 0842 utc and numerical integration of the equations of motion indicated that AO-13 would re-enter at about 0900 utc, perigee 6541/2. When pressed, a Space Command Public affairs spokesman said "the predicted decay time for object #19216 / 88051B / AO-13 is 1996 Dec 06 [Fri] 01:57 utc" (perigee 6552/3), and added that tracking was proving "very difficult", a curious admission. The comment was received on Dec 06 at 2142 utc, some twenty hours after Space Command's predicted decay time. At the time of writing (1997 Jan 15), no official confirmation of re-entry time has been received.

Given that a) during the final 3 months, NORAD keplerian elements were exceptionally poor when compared with expectation [4] and observation, and b) on Nov 03 one Space Command analyst asserted our re-entry would be January 1997, it is highly probable that AO-13 actually burned up when we predicted! This was 1996 Dec 05 [Thu] 0854-0900 utc, streaking at a temperature of 3000-5000C across the USA from Denver, Colorado to the Great Lakes.

Archives

The final month was the subject of much analysis and reported day by day on Amsat's Internet bulletin board, amsat-bb. Data, plots, telemetry and so on are archived in the directory ftp://ftp.amsat.org/amsat/satinfo/ao13/

AO-13's last message

Figure 6. This final telemetry message block appeared on Oscar-13 a couple of days before the spacecraft failed.


References


1. J.R.Miller, Mode-S - Tomorrow's Downlink, Oscar News (UK) 1992 Oct No.97 p20-22.
Also: Amsat-DL Journal Dec 1992 Nr4. Jg.19.
Also: Amsat Journal (USA) Vol15 No. 5, 1992 Sep/Oct, p.14-15.
Also: Amsat-VK Newsletter (VK), No. 90, September 1992.
Also: CQ-DL 9/93 1993 September, p.614-617.

2. J.R.Miller, Managing Oscar 13, Oscar News (UK) 1993 Dec No.104 p16-21 & 1994 Feb No.105 p18-23
Also: Proceedings of the 11th annual Amsat Space Symposium, Dallas Texas, USA, 1993. 10 pages.
Also: Satellite Operator No. 37, October 1993. p46-55
Also: Amsat-VK Newsletter No. 104/5/6/7, 1993 Nov - 1994 Feb
Also: The Amsat Journal (USA) Vol 17 No. 1, Jan/Feb 1994.
Also: Amsat-DL Journal (D), Jg. 21, No. 2, Jun/Aug 1994.
Also: Jamsat Newsletter (JA) 1994 Mar - 1995 May.

3. J.R.Miller, From Hardware to Vapourware; AO-13 Re-entry Plans, Oscar News (UK) 1996 Jun No. 119 p 36-41.
Also: The Amsat Journal (USA) Vol 19 No.4, Jul/Aug 1996.
Also: Amsat-VK Newsletter No. 135/136, 1996 Jul/Aug.
Also: Amsat-DL Journal (D), Jg. 23, No. 3, Sep 1996.
Also: Jamsat Newsletter (JA) No. 176, 1996 Aug 30.

4. J.R.Miller, The Re-Entry of Oscar-13, Proceedings of the 12th annual Amsat Space Symposium, Orlando, Florida, USA, 1994.
Also: Oscar News (UK) 1994 Oct No. 109 p 16-20.
Also: Jamsat Newsletter (JA) No. 166, 1995 March 25. p1-4.
Also: Amsat-DL Journal (D), Jg. 22, No. 1, Mar/May 1995.
Also: Amsat OZ Journal (OZ) No. 37, 1995 May.
Also: The Amsat Journal (USA) Vol 18 No.3, May/June 1995.

5. V. Kudielka & W. Drahanowsky, Phase 3D - Feasibility Study of Launch Sequences and Orbits, OVSV, May 1994.

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

Created: 1999 Feb 03 -- Last modified: 2006 Feb 23