'Spirit' Communications Emergency

"JimO" wrote in message
...
Problems -- give 'em a day or two to work them out...



Rover suffers anomaly on Martian surface
Mission managers report loss of data from Spirit
http://msnbc.msn.com/id/3855168/
MSNBC staff and news service reports

Updated: 12:41 p.m. ET Jan. 22, 2004

An "anomaly" of an unknown nature cut off data transfer from the Spirit
rover on Mars, mission managers said Thursday.

The news came a day after NASA said the rover was out of contact because of
bad weather at a radar transmission site in Australia. On Thursday, project
manager Peter Theisinger told reporters at NASA's Jet Propulsion Laboratory
that the loss of data did not appear to be due to the weather, but due to a
"very serious anomaly on the vehicle."

Repeated attempts to contact the rover, using direct Earth links as well as
satellite relays on NASA's Mars Odyssey and Mars Global Surveyor, were
unsuccessful, he said.

"There is no one single fault that explains all the observables," Theisinger
said. Among the possibilities could be a software glitch that caused the
rover to reset itself, or a power surge, or a temperature-related hardware
failure, or perhaps even a cosmic-ray hit, he said.

He said Mars Global Surveyor did make contact with the Spirit rover's radio
during one pass, but the telemetry received contained no meaningful data.

"It was sending a random pattern of zeroes and ones," deputy project manager
Richard Cook said. "What it means is that the radio was on but the computer
wasn't sending information over to it."

Toward the end of Thursday's news briefing, Theisinger passed along word
that managers had received a preliminary signal indicating Spirit was still
functioning on the Martian surface. If confirmed, that would mean the rover
had detected what it thought was a software fault.

The anomaly was a surprising twist in Spirit's mission, coming less than
three weeks after its landing.

Spirit landed on Mars Jan. 3 on a two-pronged, $820 million mission to find
out whether the now-dry planet was wetter and hospitable to life long ago.
Spirit's twin, Opportunity, is scheduled to land on Mars on Saturday.

"JimO" writes:

Toward the end of Thursday's news briefing, Theisinger passed along word
that managers had received a preliminary signal indicating Spirit was still
functioning on the Martian surface. If confirmed, that would mean the rover
had detected what it thought was a software fault.

Yet they mentioned earlier that they didn't get a communication from
it at a certain time (late last night?) that they should have gotten
if it (or the software?) was running in fault mode. And the random
data communication they got off an orbiter was only about 20% as long
as normal. Given that, it was no suprise to hear them imply that the
problems don't match any problems they have considered during the
development.

They seem to be in wait-and-see mode, but I wish some reporter had
thought to ask if it's possible to send reset commands of some sort
and about their ability to replace potentially-corrupted software via
the low-gain antenna. (As long as I'm wishing, I wish JPL would share
more details to begin with, even if it means less time for reporters
to ask how long a Martian day is or how someone feels.)

On Thu, 22 Jan 2004 20:38:04 GMT, (Gary W.
Swearingen) wrote:

They seem to be in wait-and-see mode, but I wish some reporter had
thought to ask if it's possible to send reset commands of some sort
and about their ability to replace potentially-corrupted software via
the low-gain antenna.

That is exactly what I was thinking, when it sounds not unlike what
would happen if your home computer crashes and stalls.

So I was wondering all during this press conference if either this
rover's backup system or themselves could issue a command to have it
restart.

(As long as I'm wishing, I wish JPL would share
more details to begin with, even if it means less time for reporters
to ask how long a Martian day is or how someone feels.)

Yes. Not many people at these news conferences these days, where I
feel that tomorrow that a whole load more people will be there.

Someone please ask the "reset" question, when if it can be made to
restart, then this may solve their problem.

Things like this tend to head towards worst case situation though.

In the end you have to ask the question... Could this same thing
happen to Opportunity as well? Should Spirit be lost, then moving
Opportunity off the lander may well go a whole lot faster.

Cardman
http://www.cardman.com
http://www.cardman.co.uk

Crazy idea:

Is it possible for a stray/garbled signal from Earth to be
misinterpreted by the rover? I guess there's a CRC mechanism for
uplinks but... any thoughts?

On 22 Jan 2004 16:04:28 -0800, (Remy Villeneuve)
wrote:

Crazy idea:

Is it possible for a stray/garbled signal from Earth to be
misinterpreted by the rover? I guess there's a CRC mechanism for
uplinks but... any thoughts?

Very unlikely.

What most people overlook is that serious data corruption can be
considered as random noise. And as we know that random events can
given time produce valid data, then so can seemingly valid commands
pass CRC checking.

I see that myself in downloaded files now and again, when CRC checking
is not that absolute.

So while it is very unlikely, but I would not exclude the possibility
that Spirit got a command that it should not have.

If that could have caused this problem I just don't know, when as far
as I know Spirit would not do the lemming off the cliff trick. I could
be wrong on that one though.

Still, you have seen the landscape. So what can Spirit do to hurt
itself? Baby in a playpen.

Cardman
http://www.cardman.com
http://www.cardman.co.uk

(Remy Villeneuve) writes:

Is it possible for a stray/garbled signal from Earth to be
misinterpreted by the rover? I guess there's a CRC mechanism for
uplinks but... any thoughts?

Yes, it's possible, but like every other known risk, it should be a
small risk, with commonication errors probably much lower than most
risks. It was said at the briefing that the data is protected by ECC
code that correct single bit errors and detect double (often even
numbers of) errors per data block and I suppose there's other,
CRC-type checks too. And they didn't transfer much data up before the
error. Previously uploaded data (particularly software) had been
pretty-well checked out in the last many days. Cosmic ray problems
should be just as low risk.

I'm speculating that it was some circuit board problem that was bad to
start and got worse from the bouncing and explosive devices and
finally gave up when jostled by a wind gust or whirl or temperature
change (it was afternoon local time). Maybe some memory has been
lost, as the communication systems seem to be working OK and the
computer is working well enough to answer the phone, but it's not able
to send data to the communications systems. Probably caught in some
low-level "safe" mode, occasionally trying to exercise software that's
gone and going back to safe mode. But let's hope it's something
temporary. They might be able to reload software code and data, even
mapping out bad memory hardware, if needed.