Nobody addresses communication satellites

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

Does anyone know if our communication satellites are subject to Y2K failure? Embedded chips?

-- William D. King (derryking@uswest.net), June 14, 1999

Answers

I don't, but I'm pretty sure they are a lot less important now than they were a decade ago, because of the advance of land-based (and sub- oceanic) fibre optic cables, which are cheaper, have more bandwidth, and don't have large speed-of-light delays.

These days, satellites are used mainly either for broadcast (TV, pagers) or for communications to out-of-the-way places with nonexistent or poorly developed communications infrastructures.

Its the ground-based infrastructures that are critical.

-- Nigel Arnot (nra@maxwell.ph.kcl.ac.uk), June 14, 1999.


Where's Arthur C. Clarke when we need him :)

-- Andy (2000EOD@prodigy.net), June 14, 1999.

Here's some previous GPS discussions: http://www.greenspun.com/bboard/q-and-a-fetch-msg.tcl?msg_id=000IYo http://www.greenspun.com/bboard/q-and-a-fetch-msg.tcl?msg_id=000qso

A good link for general explanation: http://pollux.com/defenseweb/1996/oct96/rollovr!.htm

-- Cheryl (Transplant@Oregon.com), June 14, 1999.


Comm satellites remain highly used, intricately iinvolved in billions of transactions daily. Fiber cables (long distance) help, bu don't carry all the load.

Troll Maria has given good info in the past about milspec satellites and their ground control station (most notably in CA (in the building called the "blue box"), but has not gone into equal details about commercial comm satellites. they remain threatened by ground-based program errors, ground-based service interruptions, and intermediate service interuptions between the ground control station, the ground relay stations, and the ground network and both ends of the comm link.

Once the signal has gotten to, been amplified by and tranmitted by the ground stations, then left the earth-based antenna, it will probably be okay until it next is received by an earth-based antenna. Biggest problem in space will be most likely the control and alignment command errors coming FROM the earth-stations. If no "bad" guidance signals go up, the newer satellites themselves will probably be okay. NASA has not said anything about their older weather and survey satellites, nothing about older comm satellites.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), June 14, 1999.


The satellites themselves are not sensitive to Y2K errors, any more than my ground-based transmission/receive equipment is. A communications satellite is essentially just a multi-channel relay (or "repeater," for you Ham-types): it receives ground-based signals and retransmits them.

The satellites do have to be precisely aimed, so there are tiny attitude jets (and an on-board fuel supply) for that purpose. But attitude drift isn't date-sensitive, and the control circuit is surprisingly simple.

Yes, some commands could be sent by the ground-based control station, but I am not aware of ANY company that would trust a millions of dollars' worth of satellite to totally automated aiming with no human supervision! Before ANY re-aim command is sent, human engineers carefully check (and REcheck) the instructions before they're EVER transmitted.

(And then they check AGAIN, after the correction, to make sure it's exactly right.[g] You're talking about MILLIONS of dollars riding on that "aim;" they make SURE it's right.)

(And if you'll please forgive me -- this is one of the disconnects that I harp about; I see things like this all the time in stuff being written about Y2K. People assume that computers are permitted to run billions of dollars worth of equipment without human supervision, which is worse than ludicrous.)

The ground based signal routing equipment could have possible Y2K problems, but I'm not aware of any that would jeapordize the bird itself. To give you an example, we use computers to feed the program audio to our bird; the worst-case scenario is that our program isn't transmitted at the correct time. The receiving stations then politely scream at Corporate Engineering in Dallas, and the problem is soon corrected. :)

Incidentally, we're getting ready to change our most important network over to an all-digital format (MCPC MPEG with a VERY sophisticated CFEC system; it's slick[g]). The receiver is the latest generation state-of-the-art, and can be controlled by us OR the transmitting station with embedded signals.

And yet, this brand-new receiver is totally date-insensitive. It doesn't even know what a "date" is, nor is there any way to enter a date into it.

You will find this to be true of most communications equipment. The layman is confused because some of the program origination and routing equipment COULD be date-sensitive; but that does NOT affect the essential chain of transmission: from earth-based send station, through the bird, and back to earth-based receive station.

-- Stephen M. Poole, CET (smpoole7@bellsouth.net), June 14, 1999.



Automated could be good, could be bad.

Don't have the issue of Physics Journal with me (it's at home) but the Oct 1998 issue had a good bit of detail about how a single erroneous control signal from the ground station "twisted" an earth survey satellite out of control by applying a continous spin to it. The signal was "automatically sent" as part of a software package that was improperly programmed to maintain alignment while the gyros were shutdown and restarted - the erroneous signal itself was not sent manually, as I recall, but the error was the result of a series of commands sent to as part a of a software update.

Nothing in the story mentioned Y2K, just regular maintenance.

Once the thrust error was applied, the antenna was stable in orbit but was spinning about the wrong axis; so its antennas were not aligned back to earth; and so the controls couldn't be "fired" again to stop the spin since no signal could get through.

I'll retype it tommorrow.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), June 14, 1999.


Robert,

Are you talking about GOES-8? Yes, a ground-based controller did send it a bogus signal in October, 1998, but the loss was temporary. The engineers were able to get it back on line.

(Had to look that one up in the Failed Sat List[g]; that's not a commbird.)

In fact, you can view images from the satellite at the NOAA/GOES Website.

-- Stephen M. Poole, CET (smpoole7@bellsouth.net), June 14, 1999.


Not that satellite - don't know about it - please elaborate.

This one: Once Lost in Space, SOHO is Found; NASA and ESA Struggle to Retrieve It. Physis Today, Oct 1998.

(trimmed... background on the Solar and Heliospheric Observatory...efforts to use ground based radar to find, communicate with it...)

"We had no idea if we would ever be intouch with SOHO again," said Joe Gurman, a Goddard project scientist. "We knew it would be race against time, because without power to receive or send communications and without directio to reorient itself, it wouldn't be long before SOHO's orbit decayed. But if we can restore the spaceraft to its original healthy condition, there is every likelihood that it will provide data to the year 2003....

"The caus eof SOHO's near demise was teh subject of a NASA-ESA investigation board, which released its final report 3 September. The board blamed the controllers who lost contact with the craft during a series of recalibrations that normally take 48 hours, but were were compressed into 24 hours so tht scientists woul dnot lose a day of data.

[Comment - hurry up, make the schedule, don't care about quality - sound familiar, anybody? Think this won't happen again - many thousands (millions?) of times worldwide?]

The erros occurred in turining off one of the three gyroscopes during routine "momentum management", when thrusters are fired to hold the spacecraft steady while a set of wheels, which maintain the spacecraft's attitude by counteracting the external torques, is slowed down. As soon as this procedure was completed, the gyro told SOHO - incorrectly - that the spacecraft was spinning 20 times too fast.

[Comment - bad data input from an automated sensor system - sound familar? Don't think this won't happen next year?]

The system went into a safeguard mode ...

[Comment - automated response to signal - sound familar?]

... called Emergency Sun Reaquisition, which kicks in automatically when an anomaly is detected in the spacecraft's orientation.

[Comment - another automated response to an automatic signal condition - sound familar?]

From there on, the troubles escalated. To save wear and tear, the computer shuts down another gyro while the wheels are braked. But becasue an essential command sequence had been omitted from on-board software during a re-write last year,

[Comment - software error introduced during updated, not found in testing before installation - sound familar?]

the first gyro failed to come back on, unknown to controllers ....

[Comment - failed signal back to controllers from process - sound familar?]

The control team did not check the status of the craft ...

[Comment - just ONE manual error, the first in this sequence of events so far, but based on their previous experience - sound familar?]

... to bring it back on line and used new software that had not been adequately tested and was in conflict with another gyro.

[Comment - their words, not mine, enough said - sound familar?]

When instrument readings did not gibe, the controllers made a snap decision that the first gyro was faulty and turned it off.

[Comment - this is only the second manual error in this sequence of events so far, but it would have also been correct based on their previous experience with good information from the sensors - sound familar?]

Without the gyros, the thrusters fired to stabilize the spacecraft. The false readings triggered continuous firings and SOHO began spinning faster and faster, preventing the solar panels from collecting sunlight and generating power.

[Comment - another automated solution coming from bad input data from automated processes, these firings were NOT manually controlled - sound familar?]

_________

Now, can any reader see why I'm skeptical about systems and process controlers - these WERE rocket scientists, working in isolation and only under mild schedule pressures. And they still screwed up, got screwed up by bad sensors and bad data, and by bad programming, and by bad testing, and by bad assumptions based on their previous experience with controllers and gyros.

The MANUAL controls and the bad programming and the bad data and the bad decisions and the bad schedule pressure caused this problem.

Sound familar? No, it's not Y2K - its a dress rehearsal for year 2000 troubles. Repeated millions of times.



-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), June 15, 1999.


Trying to get Poole's attention.

-- Robert A. Cook, PE (Kennesaw, GA) (cook.r@csaatl.com), June 16, 1999.

Moderation questions? read the FAQ