Telecommunications Satellites - Who's in Charge?

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

What about embedded chips in telecomms satellites? Do they also need to be checked & replaced? Is this ongoing? Who's doing it? If they fail, imagine the mess! (telephones, computers, pagers, GPS, defense, etc etc etc.) Please discuss.

-- Kay Killam (dkill@exis.net), January 03, 1999

Answers

Kay,

I've been wanting to know about the satellites, too. Is this a case of inaccessible chips, like the ones in North Sea oil platforms? Have there ever been satellites in orbit that had to quit being used because their technology was so out of date?

Maybe engineers knew technology would change and designed satellites to somehow merely pass along information instead of processing it. I've never heard anything factual on this issue.

Do any of you know?

-- Kevin (mixesmusic@worldnet.att.net), January 03, 1999.


In my admittedly worthless opinion, satellites probably need to be continuously synchronized with ground stations. They may have clocks on board, but in many cases it seems reasonable that current time is set from the ground. Bad news?: It's onboard software better have 4-digit years. Good news?: maybe the 1978 setback trick would work.

My impression is that the designs of alot of onboard processing in the aerospace biz is often 10-20 years behind street level. (The Shuttle's computers (3) are the size of refrigerators, quite primitive by current standards, but run incredibly reliable software)

-Lewis, who got excited about the DC-X.

-- Lewis (aslanshow@yahoo.com), January 03, 1999.


The shuttle may be OK, but how about MIR? Yeah, buddy, I sure would want that gig ....

-- jhollander (hollander@ij.net), January 03, 1999.

# # # 19990103 Kay and Kevin: The US Navy is in charge of the afflicted satellites. I watched, with incredulity, the live ( thank you, C-SPAN! ), Congrssional hearing in April, 1997. To paraphrase the Navy reps: they had no intention of fixing the satellites; they were taking appropriate action within the armed forces for work arounds; everyone else--tens of millions--using GPS devices were on their own to work out the problem. Generally, the problems are the (a) "modulo 1024" and--again--(b) lack of human foresight problem. (a) An integrated counter in the GPS systems counts elapsed weeks, using the value in algorithms that crank out an element used for determining relative geographical locations. Unfortunately, the counter--as specified by the US Navy--can only handle increments of 1 ( one ), in a range of 0 ( Zero ) through 1023 ( = 1024 ). Upon reaching 1023, the next increment of 1 ( one ) rolls ( resets ) the counter to 0 ( Zero )! The ramifications of this ( expected ) counter behvior is that upon roll-over, the raw calculations will generate the undisired result of causing correct location calculations on the earth as it was positioned in the 1970's! Cute, huh?! Any devices with the, heretofore, special purpose, standard chips used to render coordinates need to have a new chip that will compensate/offset for this soon-to-be erroneous signal. Duh! Some examples of segments impacted by this ground navigation systems flaw: Tanks, missiles, aircraft, ships, boaters/hunters/hikers, trucks, automotive, farming ( agriculture ), handhelds, etc. (b) The lack of human foresight in specifying 1023 weeks counter, alludes to the shortsightedness of the designers to the _expected_ life of the satellites. At least that's what the US Navy testified to. The bottom line, though, is that as new satellites were developed, the itty-bitty detail of the counter limitation got lost on the developers in the overall picture. Bottom line: This August 21-22 ( midnight ) rollover defect is similar to the characteristics and shortsightedness of the 2-digit ( instead of 4-digit ) year amiguity problem. Human nature, ineptitude, and the proclivity of thinking that cost is more important than doing something right the first time leads to "unforeseen" ( NOT! ) and "unpredictable" ( NOT! ) consequences. *sigh* This GPS event will compound and amplify sour sentiments of societies toward computerized systems in the future. Some URLs of interest: http://www.bluemarble.net/~storageu/y2k-a118.htm

http://www.navtechgps.com/supply/y2k_issues.asp

http://www.leica-gps.com/y2000/year2k-week.html

I hope this is an understandable explanation of the GPS roll-over situation? Regards, Bob Mangus # # #

-- Robert Mangus (rmangus@mail.netquest.com), January 03, 1999.


Hi Folks,

has anybody accumulated a list of civilian GPS systems which are y2k and GPS-rollover compliant? had been looking at the possibility of one or two of these for some folks who aren't too swift with a map and compass (still the best way to do it, IMHO), but have been unable to find anything close to a comprehensive listing of affected and unaffected systems.

Thanks, Arlin

-- Arlin H. Adams (ahadams@ix.netcom.com), January 03, 1999.



GPS ground terminals built before (I think) 1995 are affected by the rollover. It will cause an error read out for the one millisecond that it hits 0. The next millisecond will accurately determine the position. In effect at zero, the satellite will think it's taken 18 years for the signal to reach it, so the position will be miscalculated. The GPS bug is a non issue. Stop sending out bad info Bob.

Communications satellites just pass info. No, they will not be able to check all on board embedded systems. But satellites have built in redundancy. And since 90% of chips are not a problem, satellites will probably not fail on 1/1/00.

The ground stations perform TTM for satellite health. If the ground station can't get to the satellite, it can still perform its mission. Satellites are pretty much autonomous. They will stay in orbit for a long period without any TTM since they move according to the laws of orbital mechanics and drag is their enemy.

Troll Maria

-- Maria (anon@ymous.com), January 04, 1999.


Maria,

>But satellites have built in redundancy. And since 90% of chips are not a problem, satellites will probably not fail on 1/1/00.

But what is the nature of the satellite's redundancy? Where it consists of multiple copies of identical hardware or software, redundancy won't help when there's a Y2k flaw.

-- No Spam Please (anon@ymous.com), January 04, 1999.


jhoolander -

"The shuttle may be OK, but how about MIR?"

Last I heard, the current crew on MIR was the last one, so chips won't matter.

Rick

-- Rick Tansun (ricktansun@hotmail.com), January 05, 1999.


So, do satellites merely pass information along, or do they process it? Is date keeping done only on the ground, or in the satellite too?

I became convinced Y2K was a serious problem around the time of the Galaxy IV failure in the spring of 1998. Do any of you know links on satellite compliance, as opposed to GPS problems?

-- Kevin (mixesmusic@worldnet.att.net), January 05, 1999.


Kevin, Communications satellites just pass info like a conduit. Early warning sats process info. Weather sats process info. They don't care what year it is for station keeping. They use a point in space (Aires) for their positioning since the earth rotates. All calculations of their orbital parameters come from this "stationary" point. So comm sats will continue to function for quite some time after 1/1/00. I assume weather sats are also OK since they take pictures as they rotate. You see that on weather reports. They use the spin as gyros to maintain pitch, yaw, and roll.

Troll Maria

-- Maria (anon@ymous.com), January 05, 1999.



Troll Maria,

If only life were as simple as you make it out to be. Some comsats are "dumb" repeaters, others are not (ie DSCS II versus DSCS III). But that is hardly the point. The "package" on any space born platform is only part of the issue. All satellites contain on board health and status systems which perform vital functions such as orbital adjustment, power consumption, and platform attitude control. These feedback and control systems all use embedded controller technology and are very ancient in their design (Dept of Defence Sat Comm System III has less than 10K of memory using some ancient chips). If you lose attitude control, you usually lose the bird. Have you ever tried to recover a three axis stabilized comsat in a free spin? I have. (And as a side note, Aries is not the point of reference, it is a sun sensor that is used to hold position.) All these feedback loops integrate over time so yes there is the potential to have problems.

Truth is, no one is sure if the birds will remain stable and operating properly. And while errors may be recoverable, the use of limited on board fuel to correct attitude or orbital anomolies will result in a shorter satellite lifespan. Keep in mind there a a lot of satlleites besides Comsats and GPS anyway. The packages on some of those birds are highly complex in and of themselves.

It will be a very interesting ride.

Big Toe

-- Big Toe (nospam@nospam.com), January 05, 1999.


OK Big Toe, most sats that spin out of control do so at the point of injection into orbit when they deploy objects such as their solar panels. Sats already in orbit do not go spinning out of control unless they are hit by another object. Drag (more than any other component) will affect the sat's orbit and require the ground station to do a burn. Embedded chips are the big question with Y2K and how the on-board systems will shut down. I didn't mean to imply that sats are not complicated, they certainly are; but the mathematical formulas that keep them in orbit are nothing new (dated around the 16th century) and furthermore do not depend on years let alone the year 2000.

BTW most of the world population do not care about DSCS II or III or any other generation of DSCS or Milstar. Most of the world population do not use these encrypted comm birds.

Troll Maria

-- Maria (anon@ymous.com), January 05, 1999.


Here's the military link for GPS

http:/ /www.laafb.af.mil/SMC/CZ/homepage/y2000/index.html

They claim "GPS is certified", but details are accessible to mil. officials only.

Selected status sumary though, and you get this:

- All JPO-procured receivers are Y2K/EOW compliant
- All GPS satellites are Y2K/EOW compliant
- Satellite ground support systems are not Y2K compliant
- Control Segment is not Y2K/EOW Compliant
- Rollover test plan and facilities available to all DoD
- GPS JPO is widely publicizing both problems
- GPS performance on schedule for EOW/Y2K compliance



-- Chris (catsy@pond.com), January 05, 1999.

Troll Maria,

Sorry to disagree but I have seen birds spin out of control for many reasons including attitude control system failure as well as personal errors by ground controllers. Look at an embedded failure as a machine based personal error. As to the math being around forever, sure, but the solution of the equation requires an integration over time and that time hack comes from the chip. Some lazy engineers have used existing code within chips for this integration function and never bothered to drop the date from the calculation even though it only required a much smaller time interval.

Whether or not people are interested in DSCS III, MILSTAR, DMSP etc is beside the point. The civilian sector took their lead from the military designers and build the commercial versions along the same lines. In fact, without the DOD buying power for custom development, I would hazard a guess that there is a higher probability of generic PLC use in the civilian birds and hence a higher likehood of poorly designed integration functions. (And I would put forth that everyone in the US should be concerned with whether or not our DOD birds work, as our national defense and missile warning string cannot work to a great extent without them.

Big Toe

-- Big Toe (nospam@nospam.com), January 05, 1999.


Oooh Bob, I love the way you talk dirty to me! Do it again. Oooh baby.

Get a grip you moron! What the fuck do you know about satellites? Where did you get your expertise? The net? Oh my, I'm impressed. I've worked for the government in space operations for about five years and have a Masters degree in Space operations.

Troll Maria

-- Maria (anon@ymous.com), January 06, 1999.



Well dang! That explains everything. The gov. hires trolls, no wonder we're such in deep dung. Did you get your masters degree in a gov. school too?

-- Bored (yawning@home.com), January 06, 1999.

Troll Maria,

Just curious, you at Falcon or Onizuka and can you even say what bird you work on? My guess is, if you have five years in, you're probably a Captain out at Falcon in either MCC-1 or MCC-2 (please say your not an MCS puke). Orbital Analyst, Planner Analyst, Mission Commander, Ground Controller (or do they even use those titles anymore)? I like to know the point of view of those claiming detailed systems knowledge on the platforms.

As for me, I can remember when Onizuka was called Sunnyvale and Falcon was not even a hole in the ground yet. Been the around the block a few times and I'm willing to wager you have some real fun ahead of you come 2000. Don't plan on taking any leave.

Big Toe

-- Big Toe (nospam@nospam.com), January 07, 1999.


Well Big Toe, I'm with you. My experience dates back to when Falcon wasn't even on the five year plan. I worked on SPADOC (that's A, B, and C). And I never claimed to be an expert! Not once. I'm like everyone else on this forum I have an opinion. I worked on the GPS program some 17 years ago during its infancy. I worked on the Milstar program. My degree is from UCCS. I was in the Air Force, left as a Captain.

Troll Maria

-- Maria (anon@ymous.com), January 07, 1999.


Troll Maria,

I was at Sunnyvale (84-85) and Falcon, (85-90) working DSM, ARTS III, GPS, DSCS II, DSCS III, DMSP and other "stuff". Followed on at Hanscom (90-95) when I took the VSI and ran. Maybe we crossed paths. Ever hear of the Emu's?

Big Toe

-- Big Toe (nospam@nospam.com), January 07, 1999.


No it doesn't sound like we crossed paths. I was at Sunnyvale back in 82. Worked as a contractor at Falcon (93-95) on a training simulator. I seem to be a bit older than you. I can remember when early warning operated out of Ent AFB (now the Olympic Training Center) with binoculars and tin cans and string (just kidding). Nice to see someone with experience try to answer some of these questions.

Troll Maria

-- Maria (anon@ymous.com), January 07, 1999.


Thank you both - Toe and Madam "Maria the Troll". Your tech has been great, and is appreciated.

Assume somewhere between 10 and 60% of the satellites in use now behave "badly" - for whatever reason - after power is lost in several areas down here. Who and how can they be recovered if there is no ground power/lights/cooling or HVAC/control programs down here?

What are the alternative ground stations, and what are their emergency control plans?

Axis control (based on a faulty ground-based program) was the failure mode of the Galaxy satellite - and I think that satellite is still OOC.

-- Robert A. Cook, P.E. (Kennesaw GA) (cook.r@csaatl.com), January 07, 1999.


Say, did any of you know Robert Berg from MIT ?? He did a lot of work researching/teaching/everything to do with satellites. He was a wonderful man! He was our hospice patient. Any memories of him, anybody?

Ashton & Leska in Cascadia, who loved Robert Berg and his family

xxxxxxx xxxxxxx xxxxxxx xxxxxxx xxxxxx

-- Leska (allaha@earthlink.net), January 07, 1999.


Well, some 15 years ago I did research for the government as a consultant. I researched the DoD's satellite systems (excluding the intel ones) and commercial commsats. Put together the space operations network which included all owners and operators, ground stations, and the connectivity to SPADOC (land lines as well as sats). It also included NASA and NOAA stuff. The document was classified because it pointed out vulnerabilities. For the most part, ground stations are duplicated around the globe. NORAD can operate for 30 days in "button up" mode; that's without access to the outside world (no trips to the grocery store). Falcon has generators also; not sure how long they last. Not sure about Intelsat or Inmarsat but I can't believe that they would chance a power outage during critical TTM functions, so I have to believe they have backup generators. Since satellites are very expensive (into the billions), owners and operators build in a lot of duplication. The duplication is not a second computer (for the most part) but a second design to perform a backup function. They literally can't afford any failures.

I'm sorry Robert I can't answer your question specifically because I don't know that much about commercial stuff and my experience is dated. But I believe from what I do know, that power will be supplied to ground stations during an outage. (Was that the heart of your question?)

Troll Maria

-- Maria (anon@ymous.com), January 07, 1999.


Troll Maria,

Probably never crossed paths but if you were on a simulator you maf have been inside a place called the TMF? I was the lowly Captain who worked with IDI Corp to build that place if that's were you were.

As to ground stations can't say what all the commercial world looks like, but the Air Force is still using IBM 3083's for all the remote tracking station and TTS functions. You know, the same ones the FAA is using for ATC that IBM says won't work come 2000? Plenty of rendundency in the systems but they are all the same configuration. If the core processor doesn't work, all the backup systems in the world won't help. Let's just say I am glad I am not in that business anymore.

Big Toe

-- Big Toe (nospam@nospam.com), January 07, 1999.


Geez, my first job as a government contractor was to test missile warning messages on, you guess it, those FAA consoles. What dinosaurs! A friend who used to be a commander at NORAD tells a story that foreign dignitaries came to tour the mountain. When the tour ended, they asked the US officials, so where are the real centers? They couldn't believe we used such antiquated equipment; they thought we were hiding the centers. Soon after NORAD began upgrading its systems.

Sometime in the early 90's the "blue cube" received funding for an RTS moderization program. I was out of that world at the time but I gotta believe that they replaced those relics during the upgrade.

We (my previous employer) installed the CMTS (Cheyenne Mountain Training System) on the second floor of the TMF. I think I needed IDI sign off of my clearance. Don't they maintain the facility?

Anyway what are EMUs?

Troll Maria

-- Maria (anon@ymous.com), January 08, 1999.


Troll Maria,

Actually the IBM 3083's went in as part of the RTS upgrade (known as ARTS III). As to the TMF, IDI was replace by another company in 1989. Emu's, why we were the flight to emulate. Just the unit mascot for echo flight out at MCC-2. Not suprised no one has heard of them.

Actually, considering your time in SPADOC you probably had more of a chance to run across my wife. She ran the missile warning string tests back in 93-95 (another AF officer). We were at Hanscom at the time so lots of TDY.

Big Toe

-- Big Toe (nospam@nospam.com), January 08, 1999.


Yeah but if they replaced these systems in the 1991-1992 time period, and figure funding & spec's were two-three years ahead of that, it is likely that Y2K "prevention" would not have had a high priority at the time.

It's likely that these would indeed need checking, if not remediation - hardware and software sides.

-- Robert A. Cook, P.E. (Kennesaw GA) (cook.r@csaatl.com), January 08, 1999.


Troll Maria & Bigtoe;

I would like to ask you two if you could answer the original question for us weenies.

Last I heard Bigtoe said it could be a problem with the birds and T. Maria said "nah it ain't to worry."

I know that the things will PROBABLY stay reasonably in orbit, although I can also see how some won't and will go phffft. But ops wise? What's your best SWAG?

Who is right? Enquiring, although somewhat simple, minds want to know.

S.O.B.

-- sweetolebob (La) (buffgun@hotmail.com), January 08, 1999.


SOB,

You'll love the answer (not). I really don't know. Like everyone who has dealt with complex systems I can define several identifiable failure paths that would lead to an outage, and also know there are many I can't even begin to anticipate. But will some happen? Probably. Will they reach critical mass? Maybe.

By way of example. The GCS (ground control segment) for DOD satellites is comprised of a series on remote tracking stations (RTS) located around the world. These sites are typically what is called a bent pipe for channeling the telemetry back to the Mission Control Center (MCC)located elsewhere (specifically Synnyvale CA, and Colorado Springs, CO for the birds I'm familiar with). This can be done either via land line or a telemtry relay satellite (ie TDRSS). Within the MCC the telemtry can either be bundled for forwarding to the agency responsible for using the data, or it can be used for command and control of the actual orbital platform. Commands issued from the MCC are retransmitted to the RTS for uplink to the bird. Within that entire chain you have power, telecommunications, mainframe platforms, slaved antenna arrays, and the actual command and control software. This is before you even get to the actual space born asset and its embedded systems.

Most commands sent to satellites are somewhat time critical and the support window is usually very short as the satellite continues in orbit past the view of the RTS. Should a command only be partially processed (ie begin thruster burn, but the stop thruster burn command is never received)odd things tend to happen. Should one those odd things be the start of a tumble in space you have a real problem. As the satellite wobbles around on orbit it becomes very hard to restablish an antenna link as the antenna itself sweeps into and out of view rapidly. This is how many satellites are "lost".

The same thing can happen within the embedded systems. Should a sun sensor loose lock on our own little star, the satellite can begin a search to try and reaquire its beacon. But if that loss of lock was due to a faulty integration over time of the sensor value the attempt to acquire will most likely fail. The bird then goes into some type of safe mode and waits for ground control input. It is possible to load firmware patches from the ground provided you know exactly what went wrong. Unfortunately, the use of onboard fuel is what defines a satellites lifespan and so every "adjustment" to the platform shortens its useful life.

From the standpoint of Y2K being considered as a design parameter in those assets, I can say with authority, it never crossed our minds at the time. That also means I am only making educated guesses based on what I remember out of a bunch of technical specifications, not a definitive systems engineering assesment of an actual failure. Humble opinion: We will experience a combination of GCS failures during rollover as well as sporadic embedded systems failures onboard some birds that will result in a sgnificant reduction in the useful lifespan of those platforms after recovery from the outage.

Big Toe

-- Big Toe (nospam@nospam.com), January 08, 1999.


Big Toe;

Thank you for your answers Sir.

I know some of the guzinners/cumzouters of these things myself and I appreciate your honesty in this.

We could use a whole lot more answers like this to a lot of the other questions we have. That would go a long way towards lowering the FUD factor.

Thanks again Sir.

S.O.B.

-- sweetolebob (La) (buffgun@hotmail.com), January 08, 1999.


Moderation questions? read the FAQ