OT - Curiosity's lack of memory

[Tim+]

This story intrigued me.

See the last paragraph. " Due to capacity limitations on the craft, the rover launched with only the software needed to land. Once on the planet, new software had to then be installed remotely."

Is that likely? Of all the things to take to Mars, surely in this day and age, the weight penalty of a bit of memory is probably one of the least significant things?

One the one hand I suppose it makes sense as the landing software is redundant after landing but it seems to be putting a lot more reliance on a good upload link.

Tim

[John Rumm]

That NASA report[2] does not mention memory capacity:

although the rest of the storey seems to fit...

Elsewhere[1] they say:

"On-board memory includes 256MB of DRAM and 2 GB of Flash Memory both with error detection and correction and 256kB of EEPROM. This onboard memory is roughly 8 times as capable as the one onboard the Mars Exploration Rovers."

Keep in mind that it was not designed "in this day an age", but possibly a significant amount of time before. I don't know how long space exploration equipment lead times are, but I would be surprised if they are any quicker than military ones, and in those circles its not unknown for memory technology to actually go obsolete before the design of the thing is even finished!

(2GB of milspec, rad hardened flash is not bad!)

[1] [2]

[Davey]

This might help:

[John Williamson]

I'd expect all the software to be stored in a ROM on board, but with different sets of software loaded at different points in the mission to increase reliability. While the weight penalty of extra RAM is minimal, every gramme counts in spaceflight, installing all the software permanently will increase the chances of corruption, and will also increase the power draw of the computers, so increasing the dead weight of the rover to handle this.

There is also likely to be an ultimate backup procedure to load software from a signal sent from Earth. The corruption, according to NASA, is in a segment of the Flash memory attached to one of the computers, so they may well end up just sending a difference file to rewrite the corrupted bits, and may also have to map some RAM as unusable. They may also be able to feed data from the uncorrupted flash on Computer B if that uses the same OS, although good practice would suggest that the computers use different operating systems to increase reliability.

[mark]

Plus it would have to search for a wi-fi hotspot to download any new software.

mark

[The Natural Philosopher]

I think I saw that one of the problems was radiation damage to pre-installed software..

The option was to load what was needed and check it, once it had in fact got there.

[Mentalguy2k8]

And I bet the first message it sent back was that Adobe needed updating

[Mentalguy2k8]

Pretty amazing when they can upload new software from Earth to a computer on Mars, yet I can't get a wifi signal in the bog.

Pretty trusting too, if it's anything like the gadgets I've encountered, it's probably sitting in a Martian desert flashing the message "this update can only be installed via USB cable".

[Mark]

Na. It's not running Windows.

[Tim+]

Thanks. An interesting read.

Tim

[Tim Streater]

"Missing keyboard - hit any key to continue" more like.

[Mark]

I've seen that actual error message.

[The Natural Philosopher]

ah, but if you had a couple of billion you would.

well there is a subtle difference between 256K of hand coded assembler written and tested over 10 years by professionals and 64MB of bloatware written in PYTHON by a graduate who hasn't a clue.

>

[John Rumm]

Ah, that will be the crap radio shielding then ;_

[Owain]

Yeah, the assembler doesn't talk TwitBook API

Owain

[geoff]

In message , Tim+ writes

Don't forget that design for this was prolly fixed 10-15 years ago and was then based on proven technology

Rocket science just ain't rocket science

[geoff]

Good old silicon on sapphire - they just don't make it like they used to

[Windmill]

If the upload link is no good, the rover is just scrap metal anyway.

[Brian Gaff]

Its interesting that this is flash memory's the other surviving rover has also had issues with the same kind of memory. The second chain, now in charge will probably fix things for a while, but it does seem that radiation can upset this sort of memory if it is exposed. Certainly the journey to Mars for Curiosity did have a few glitches due to sun activity, and one must assume that some damage has occurred. apparently a new ram health testing system is in work, presumably to mask out bad or suspect areas. As you say one would have though more spare might have been taken. I was also a bit perturbed by the breaking of a small tank of chemicals that allowed th testing overns to be flooded by the leaked chemicals. Would one not make sure all fluid containers were made either more robust or at the very least so the leak could be routed away from the instruments?

Brian

[Brian Gaff]

I don't think the issue of errors is really an issue as there will be rock solid error protection involved. However it does take some time to do it. I understand that there was of course some minimal software for using the science instruments on board, and in fact the memory must have survived the trip or things would have gone wrong. I'm sure though we all know that memory can fail due to earlier partial damage, so it could indeed have been compromised in flight, and has now become unreliable. Brian