NASA to reformat Opportunity rover's memory from 125 million miles away IT administrators at NASA's Jet Propulsion Laboratory are preparing to perform an interplanetary remote wipe and reboot on its Martian Opportunity rover after the decade-old explorer began to get a bit senile. The latest view from Mars sent by Opportunity The latest view from Mars sent by Opportunity Opportunity's …
They only thought the rover would last 90 days, so I suspect flash endurance wasn't a really a consideration.
But look at it like this: That flash card has been working in a challenging environment for 10 years. If my server SSDs survive that long here on planet Earth, I'll be more than happy.
You remember the Macs with the PowerPC architecture? Like that but silicon on sapphire, thick gates, big transistors, running at 20MHz. ECC memory. They used the commercial version, running at around 200MHz, in a number of printers long after Apple had gone over to Intel.
But almost certainly not running Mac OS 9.2
For Curiosity, we read beyond the paywall:
The control software onboard the spacecraft consists of about 3 MLOC. Most of this code is written in C, with a small portion (mostly for surface navigation) in C++. The code executes on a radiation hardened CPU. The CPU is a version of an IBM PowerPC 750, called RAD750, which is designed for use in space. It has 4 Gbytes of flash memory, 128 Mbytes of RAM, and runs at a clock-speed of 133 MHz. About 75 percent of the code is autogenerated from other formalisms, such as state-machine descriptions and XML files. The remainder was handwritten specifically for this mission, in many cases building on heritage code from earlier Mars missions.
But also....
....Second, we introduced a flight software developer certification course, focused in part on software risk and defensive coding techniques. Every software developer is required to complete this course and pass the exams before they can touch flight software. The course covers the coding standard’s rationale, as well as general background on computer science principles and the basic structure of spacecraft control software. Some of this material is also presented to more senior managers at JPL to secure a common knowledge base regarding the challenges of mission-critical software development (although in the latter case, the material is presented without the pressure of an exam at the end).
Excellent! Glorious Freedom from the web craze.
Forget this proper technical stuff - they should watch more telly - that would tell them all they had to do was whizz over to Spirit, use the sampling arm to remove it's SSDs and magically simply slot them into place, bearing in mind that ANY component fits ANY slot in the movies. Yes, I know it's possible this may not work because at the last minute an actual human needs to take a hair pin from tasty sidekick to "create a bridge" (and why is it always "create a bridge"?) but it's got to be worth a try
"For solar-powered planetary explorers, dust is a killer. "
What strikes me as odd, is that knowing this someone made the decision to send them up anyway and write off the rovers after 90 days, rather than fit a wiper-blade on the solar panels.
It doesn't take a rocket scientist to, oh wait....
I often wonder if there was a lone scientist in the room at briefings that said "you know guys, if we were to fit a wiper blade, we could keep these things going for at least 10 years", at which point he'd be laughed out of the room.
...and if they're not happy with the mechanical reliability of a big moving object, how about a couple of "sonic shakers" around the panels? All those Youtube videos showing granular stuff neatly grouping into various thin lines on speakers suggests it might even be possible to "wipe" dust off using the right vibration sequence.
The worry was that the dust would like Moon dust be charged and the vibration trick wouldn't work all that well with charged dust, also ultrasonic vibrations would put stress on the connections between the panels and the rover which could introduce a whole new selection of faults. So even if it could work the problems in hardening the rest of the vehicle to tolerate the vibrations make it not worth the effort. Probably.
As for wipers I suspect they didn't want to scratch the cell surfaces which is what wipers in a dry dusty environment will inevitably do.
Which to me still doesn't make sense. They expected to lose the rovers after 90 days due to dust build-up, so there was a planned multimillion $ write-off from the start. A few scratches are here nor there in the grand costs.
Unless the plan was to waste that money, so now someone at NASA has had a 10 year egg on their face :)
Which bit, the wasting money, or the 10 year egg on face?
If the latter think about a group of scientists in a room right now.
Scientist 1) "I think this new Mars project will last 6 months."
Scientist 2) "What would you know, your last estimate was out by a factor of 40!!!"
So I'm back to the first part. A whole lot of money was deemed to be acceptably wasted which I find a bit strange.
Unless the idea was "Hey we're struggling to find funds to keep us in a job, lets send an RC car to Mars, and we'll tell the government that it will only last 90 days, then don't forget to act surprised when it keeps on going, they'll have to fund us for decades!!!"
PS I'm not belittling the spectacular achievement that we have invaded a whole planet with robots. I'm just surprised that the best brains of the planet couldn't come up with a way to keep the robots working past 90 days in the first place, or that no-one figured out that the weather wasn't all that bad in the first place.
Have you looked at the top deck of these things? Do so and report back with the number of wiper arms you'd need. Take into account that you're not wiping the windscreen of an automobile; the design of the deck requires a plethora of protrusions, PV cell interconnect wiring, etc.
"As for wipers I suspect they didn't want to scratch the cell surfaces which is what wipers in a dry dusty environment will inevitably do."
Scratching the surfaces wouldn't matter that much, but moondust is amazingly sticky (attempts to wipe it off suits simply didn't work) and the assumption was that mars dust was the same, so it wouldn't work.
Some kind of brush/puffer might work (assuming a compressor can be fitted into the robot chassis that won't itself be clogged up with dust) but the bigger problem is mechanical endurance in such a hostile environment - the failed wheel motors and ripped tyres being good examples of this.
Given the beetle shape of the rovers, I was minded to suggest that the panels could pivot to vertical and shiver slightly, but that's another level of mechanical complexity.
"Not scratching the surface" is kind of a silly argument. If the option where to either A) Have the entire surface covered in dust and lose the rover, or B) Have the solar cells get scratched up and lose 10% of their efficiency, I think it's obvious what the better option would be. I'm sure the dust buildup has a far greater effect on performance than scratches ever would.
Of course, the real reason was likely that their goal was to make a rover that would last at least 90 days, and within that time frame, dust buildup probably wasn't their biggest concern. During the initial planning stages, they may have figured that other components would be more likely to fail first, like from dust damaging motors and such. In that sense, the other components might be considered to have been overengineered, seeing as they held up far longer than necessary. And they probably did consider the possibility that gusts of wind might clear off dust, even if it wasn't something that was guaranteed.
In any case, actual "wipers" probably wouldn't have been the best solution anyway. Some sort of tiny air compressor or compressed air canister would probably work better. Or even just a clear plastic sheet covering each panel like a screen protector, that could be peeled back once with an attached wire in the event that dust became an issue.
"...write off the rovers after 90 days..."
It doesn't work like that. You don't assume that your car will stop working after 2 years, but when you buy it, they only say that it will work for 2 years (after that you are on your own and have to pay if something happens).
What happens, is that the specs call for a reliability figure and a time (e.g. my last project had a reliability figure of 99.999999% for 15 years of life). Everything is then designed for that lifetime/reliability. So in the case of the rover, there would've been "somepercentage%" probability that it would last 90 days. If it died before that, someone would've been sent to the American equivalent of a Siberian weather station (Guantanamo?). The expectation would be that the mission would last longer than 90 days, but you need to draw a minimum line in the sand to start with.
Everything would've been set up and funded to operate the rover for at least 90 days, and after that, additional funding and manning would be provided if it still worked, and the resources were not needed elsewhere (e.g. time on NASA's deep space network.
The wiper blade idea has been mentioned in other posts, but the dust is not like Earth dust. It is extremely gritty and will scratch anything it's dragged over. Also, if it is charged, that won't necessarily help anyway, because it could just jump back over the cells when the wiper blade has passed over.
...and have done it with the MER rovers several times with no problems, the idea of optimizing and reformatting a system from 125 million miles out scares the hell out of me, when I think of the times a simple OS update on my desktop has blown up in my face.
...and have done it with the MER rovers several times with no problems, the idea of optimizing and reformatting a system from 125 million miles out scares the hell out of me, when I think of the times a simple OS update on my desktop has blown up in my face.
Or the poor thing is stuck there for eternity with "Press any key to continue" on its console.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
