SpaceX Falcon 9/Dragon CRS-2 SpX-2 MISSION GENERAL DISCUSSION

[MP99]

If it is unberthed and CanadArm moves it away and releases, it will gently drift away from ISS due to orbital mechanics.

Cheers, Martin

[Robotbeat]

Would having more than 4 computers on future Dragons & DragonRiders be practical?

Law of diminishing returns applies here. If your expected use scenario does not see you losing more than 2 computers in a single flight then adding more computers doesn't help anything and only adds weight. More so every computer you add adds to the complexity of the hardware which in and of itself add additional failure modes. This was the major problem with clustering engines as a single failure could cause others to fail (say shrapnel).

Adding another computer wouldn't be simple as it might sound.  The "real" solution would be a hardened computer, but that's no simple task either.  Things are usually tightly packed on systems like the Dragon, but if there is any space at all some more shielding might be the easiest way to better reliability. 

More shielding won't stop GCR rad damage...

Much of the worst is caused by solar radiation.

[Comga]

If it is unberthed and CanadArm moves it away and releases, it will gently drift away from ISS due to orbital mechanics.

Cheers, Martin

Right

So if two computers failed beyond recovery, and none did during CRS-1, NASA would have let Dragon drift away passively.  When it got outside the KOS, or some other TBD range, SpaceX would execute a reentry with the remaining computer.  This would have some probability of success, and CRS-1 was far from this limit of disfunction.

Why all the insistence on switching to very expensive, very slow computers when the current redundancy worked?  Rad-hard computers seem like a discussion for the Red Dragon or MCT threads, not CRS-2.

As for CRS-2, what this discussion is supposed to be about, it is hard to see this having any impact.  Perhaps they will automate the resynchronization, with the necessary software testing, but there doesn't seem to be any need for changing the architecture.

[baldusi]

The article also mentioned that they decided not to resynch the computer, because they didn't felt it was necessary, but the fact that it got out of the loop doesn't means it had permanent damage.
I suspect, if I'm forced to make a guess, that the radiation event might have generated some non recoverable latch up. But instead of trying to get it back, they kept it down so they could analyze it later when it returned.

[LegendCJS]

The article also mentioned that they decided not to resynch the computer, because they didn't felt it was necessary, but the fact that it got out of the loop doesn't means it had permanent damage.
I suspect, if I'm forced to make a guess, that the radiation event might have generated some non recoverable latch up. But instead of trying to get it back, they kept it down so they could analyze it later when it returned.

SpaceX wanted to resynch the computer, but Nasa was not in favor of doing that while it was attached to station, so they didn't.  If it had been free flying SpaceX would have issued the resynch command. Moving forward SpaceX plans on making resynching automatic.  Also to note is that SpaceX's expectations for radiation induced electronics trouble were higher than what they actually observed for the mission.

source: talk given by SpaceX employee (senior GNC engineer)

[MP99]

The article also mentioned that they decided not to resynch the computer, because they didn't felt it was necessary, but the fact that it got out of the loop doesn't means it had permanent damage.
I suspect, if I'm forced to make a guess, that the radiation event might have generated some non recoverable latch up. But instead of trying to get it back, they kept it down so they could analyze it later when it returned.

SpaceX wanted to resynch the computer, but Nasa was not in favor of doing that while it was attached to station, so they didn't.  If it had been free flying SpaceX would have issued the resynch command. Moving forward SpaceX plans on making resynching automatic.  Also to note is that SpaceX's expectations for radiation induced electronics trouble were higher than what they actually observed for the mission.

source: talk given by SpaceX employee (senior GNC engineer)

Thanks for reporting that talk.

Was there any discussion / suggestion whether SpaceX attempted a re-sync of the third computer after Dragon was free-flying safely away from ISS?

(BTW, ISTM completely reasonable if this was considered an inappropriate phase of flight to attempt this.)

cheers, Martin

[Comga]

Suffredini's presentation now shows March 3 for SpX-2 berthing on a chart updated 11/13.  This agrees, unsurprisingly, with anik's posts.  (It still has SpX-2 berthing on January 20 on page 5, but that doesn't appear to be as current.)

[oldAtlas_Eguy]

Suffredini's presentation now shows March 3 for SpX-2 berthing on a chart updated 11/13.  This agrees, unsurprisingly, with anik's posts.  (It still has SpX-2 berthing on January 20 on page 5, but that doesn't appear to be as current.)

From the standpoint of pad flow processing, Spacex has until mid to late December to resolve the engine problem, perform the fix if any and ship the first stage to the Cape. Otherwise the launch date will slip again.

[Jason1701]

Suffredini's presentation now shows March 3 for SpX-2 berthing on a chart updated 11/13.  This agrees, unsurprisingly, with anik's posts.  (It still has SpX-2 berthing on January 20 on page 5, but that doesn't appear to be as current.)

From the standpoint of pad flow processing, Spacex has until mid to late December to resolve the engine problem, perform the fix if any and ship the first stage to the Cape. Otherwise the launch date will slip again.

The stage has already been shipped. It was on another thread.

[Go4TLI]

It doesn't really matter if the stage has shipped or not. The essence of the point is correct in that if a fix is required, its effectivity needs to impact that stage and the fix is intensive enough then time most likely is growing short in order to not impact processing for flight and the manifest

[king1999]

http://www.aviationweek.com/Blogs.aspx?plckBlogId=Blog:04ce340e-4b63-4d23-9695-d49ab661f385&plckPostId=Blog%3A04ce340e-4b63-4d23-9695-d49ab661f385Post%3Aa8b87703-93f9-4cdf-885f-9429605e14df

Dragon uses the same design principles as the Shuttle and Hubble.

[mlindner]

http://www.aviationweek.com/Blogs.aspx?plckBlogId=Blog:04ce340e-4b63-4d23-9695-d49ab661f385&plckPostId=Blog%3A04ce340e-4b63-4d23-9695-d49ab661f385Post%3Aa8b87703-93f9-4cdf-885f-9429605e14df

Dragon uses the same design principles as the Shuttle and Hubble.

Very interesting, thank you.

So this seems to imply that there was indeed no non-correctable error that occurred. More so it looks like they have no plan to use rad hardened parts even when going to Mars or anywhere else for that matter. This is very good indeed.

[Jim]

More so it looks like they have no plan to use rad hardened parts even when going to Mars or anywhere else for that matter. This is very good indeed.

That was not said or inferred.

[mlindner]

More so it looks like they have no plan to use rad hardened parts even when going to Mars or anywhere else for that matter. This is very good indeed.

That was not said or inferred.

1. They state that the most important thing is the power usage, memory size, processing speed, and physical size.
2. The second most important thing is workforce availability which allows them to get people who are more used to standard software development (C++, linux) rather than specialized embedded software developers.
3. They also mention that they can't be as agile if each software developer isn't allowed to have a complete flight computer on his/her desk.

All of this implies that going forward in the future this is their development philosophy. You don't 180 your philosophy suddenly for small reasons, you continue to make your philosophy work with the new environment. If the reliability they have now isn't good enough for deep space then they add more redundancy until you get enough. I think this is a pretty strong point to them to continue this way of designing. They won't be switching to rad hardened components.

[Jim]

All of this implies that going forward in the future this is their development philosophy.

No, they were only talking near term, which means LEO and nothing to do with Mars.  They can always return from LEO in a few hours.

[Chris Bergin]

SpX-2 discussion thread guys. All these SpaceX threads to choose from.....

[Go4TLI]

If the reliability they have now isn't good enough for deep space then they add more redundancy until you get enough. I think this is a pretty strong point to them to continue this way of designing. They won't be switching to rad hardened components.

That's the wrong move.  Reliability of a system or component alone is not by itself mitigated by just adding redundancy.  There is an inflection point where more redudancy to compensate for bad reliability adds to overall complexity, cost, etc. 

[Robotbeat]

If the reliability they have now isn't good enough for deep space then they add more redundancy until you get enough. I think this is a pretty strong point to them to continue this way of designing. They won't be switching to rad hardened components.

That's the wrong move.  Reliability of a system or component alone is not by itself mitigated by just adding redundancy.  There is an inflection point where more redudancy to compensate for bad reliability adds to overall complexity, cost, etc. 

This is an information-theory question, and for all intents and purposes, that inflection point is very, very far down.

You can go a LONG ways just adding redundancy and still end up with a much /more/ reliable system even with FEWER reliable components. This is more true with computer systems than it is for other engineered systems.

As long as you're not resetting faster than it takes to sync back up, you have a lot of options available.

(This isn't to say that at some point rad-hard isn't at some point the right cost decision, just that redundancy in computer systems very easily can more than compensate for lack of subsystem reliability.)

[A_M_Swallow]

{snip}
All of this implies that going forward in the future this is their development philosophy. You don't 180 your philosophy suddenly for small reasons, you continue to make your philosophy work with the new environment. If the reliability they have now isn't good enough for deep space then they add more redundancy until you get enough. I think this is a pretty strong point to them to continue this way of designing. They won't be switching to rad hardened components.

SpX-2 is unlikely to have rad hardened computers in it.  SpX-3 could.

There may be other customers for a faster radiation hardened computer.  I suspect that Robonaut needs them.  So will Morpheus landers and MMSEVs.  Some satellites and probes may need one.

As for the computer programmer's desk tops - rad hardened components normally come in a cheaper civilian version.  Build a few 'civilian' computers with the same circuit boards.

[mlindner]

{snip}
All of this implies that going forward in the future this is their development philosophy. You don't 180 your philosophy suddenly for small reasons, you continue to make your philosophy work with the new environment. If the reliability they have now isn't good enough for deep space then they add more redundancy until you get enough. I think this is a pretty strong point to them to continue this way of designing. They won't be switching to rad hardened components.

SpX-2 is unlikely to have rad hardened computers in it.  SpX-3 could.

No that was not what was said. He said its ingrained in how they get developers. They use many different types of processors and they use C++ and Linux. It they get rad-hardened hardware then that suddenly means the hardware (and software to interact with it) is non standard. They now have workforce issues and they need specialized people with knowledge of that hardware.

I'll say it again. They _won't_ be switching to rad hardened processors, not now, not anywhere in the near future. We'll see who gets to tell who "I told you so" when they send spacecraft to Mars.