Page 1 of 1

Flash corruption in Space

PostPosted: Mon Jan 11, 2010 2:09 am
by aek
Well, look at what's up there right now -- Libertad-1 (34 days) and Delfi-C3 (623+ days) have Pumpkin CubeSat Kit FM430 Rev B & Rev C, respectively, and have reported no reboots at all ... so it does not necessarily follow that CubeSats in typical CubeSat orbits suffer dramatically from rad events.

Most CubeSats are single-point designs -- i.e. a single failure (e.g. in the controller of an SD Card) can bring down an entire (sub)system (e.g. the SD Card). So you have to evaluate where it makes sense (if at all) to "harden" a CubeSat. You could "harden" by doing TMR on the SD Card, for exampple, but you're still subject to SD Card controller failure.

As for the BSL, unfortunately it's not too easy to use, as it references Tx/Rx pins that are not UART pins. But, in theory, if you could fly a hardened uC/external WDT that could "talk" to the BSL, then you could in theory re-flash the MSP430 with code that was stored in this uC/external WDT.

But CubeSat data seems to suggest that existing CubeSats are "adequately reliable" in LEO orbits.


Re: Flash corruption in Space

PostPosted: Mon Jan 11, 2010 11:11 am
by ankushdesai
Hey Andrew,
Regarding the doubt I had asked before. Its being observed that the memory corruption due to soft errors is very high in space. And since we are not using any radiation hardened devices what are the posibilities of having corrupted code in space ( Corruption of Flash code/Data). Also can you suggest upon the remedies for the same. Also what are the method because of which Cubesat is safe against it.
Also we were planning of programming the Fm430 with BSL. any suggestions on the same.