Hardened Spacecraft and aircraft electronics

Wow! I just asked for an ES of a small computer and it wasent even hardened and they wanted $30k for it? This guy said it's about $70k for a hardened one! If there are surplus hardened computers out there that are not to big we want one! In fact anyone with aircraft or spacecraft hardened gear laying around please contact us we need that gear! We are going to have to make do with what we can find. I would not give $70k for a computer if it was hard as diamond! This is why we are not in space yet! So think hard about weather or not your really going to use that hardened amplifier or raido or whatever gaget your using for a paperweight or doorstop or whatever. Lets put it in space where it belongs. We really really need your help here!

Monroe
Team Prometheus

monroe@teamprometheus.org

monroe you definitely don't need hardened electronics. your orbit will decay well before you can expect any SEUs from random radiation. as long as you aren't unfortunate enough to have a solar storm while you're on orbit and it somehow gets you into a non-resettable state you won't have any issues. you only need hardened electronics if your mission has to survive solar storms.

EDIT: i don't know why you'd need a flight computer either, just use commercial microcontrollers and/or fpga's. contrary to popular belief these are plenty fast and robust to run in a space environment and give the functionality you need for a short amount of time as long as your thermal subsystem is reasonably good enough to keep the temperature within a reasonable band.

terra, for the N-Prize mission that might be true. Our first mission in orbit will more likely not be the N-Prize attempt. We also need the hardened gear for our station keeping platform that will spend a lot of time in near space and serve as a satellite replacement as well as other uses to get money to fund things (we hope). The near space platform comes before the ability to launch from there. There are other reasons as well. What do you think the DOD thinks about unhardened hardware and the ability to safly abort a mission if need be. Ect... We think of every angle over here esp. thoes angles the happier they are the happier we will be.

Monroe

So you don't have any goodies?
Edit: Real flight computer's are not that complicated. The ISS uses 5 redundant 80386SX processors (with math co's added) at 20Mhz. Part of the reason is the die size. Its a Saphire chip too. Sram memory.
Yep, Hard drives don't work either contrary to popular they are not hermeticaly sealed and no air bearing no drive the heads hit the platers.

My understanding of hardened processors is that one of the features is hardware lockstep of each calculation inside the CPU, more or less doubling the transistor count. This is used in banking chips like the Fujistu Sparc.

Well you would do a poormans version of this in software using two CPU's or running everything twice on one.

However Intel Xeons should be bringing hardware locksteping between processors to 'cheap' servers in the near term. They already have the ability to duplicate or triplicate their memory banks to protect against radiation corrupting the ram. (in case multi-bit ECC isn't enough)

Thats good information thank you. I also understand the latching voltage is higher and takes more power on hardened chips. Any chip on any board that was intended to go into space (Nasa and missiles and navy ships ect...) has hardened chips of all kinds. Thoes are the chips we are after. We can build a simpler system from them but they need to be hardened for our platform. If possiable. Hey, if we can't get them we'll figure another way. The information you just gave me is positive.

Monroe

i'm not sure what kind of radiation you'd be looking at on your station keeping platform, it is low compared to space hardware and the atmosphere is dense as well. perhaps this would be a good thing to measure when you're up there. i'm just not convinced that radiation hardening is necessary when you can use TMR and error correction to protect your unhardened chips. also i'm pretty sure DOD isn't who you're concerned with not approving of unhardened electronics - it'd be FAA and FCC i'd imagine. as long as a failure wouldn't result in a heavy weight falling from 100,000 feet up in any circumstance (by a mechanical parachute backup perhaps) i'm guessing they don't care as long as there's not any chance your transmitters go haywire and mess up electronic communications in some frequency range, or pose a hazard to airspace.

regardless, i'm sure it'd be nice to have but hardened electronics are some of the most unneccesarily expensive components for short-duration space missions. i really think you should try your best to avoid using them.

on a side note, loral has flown an unhardened, not space rated, camera system on one of their recent satellites to film the unfurlable reflector on it. unfortunately, they turned it off after IOT but that was still 2 weeks on orbit in GEO with no problems, in addition to surviving thermal vac and EMI testing, etc etc. as long as you're ok with maybe losing stuff in a big solar storm i'd say 2 weeks in GEO is probably equivalent in radiation terms to over a month in the upper atmosphere. and good luck keeping your balloon up there for a month and not losing it (not suggesting this is impossible, just harder than making unhardened electronics work).

The plan for the platform is min. 3 months to work as a backup system for satellites in orbit. This system is designed to work as a fast deploy for military and commercial applications as well as for scientific endeavors. I don't see why you’re against hardened equipment for a commercial application. The fact is we may have to open our own plant to produce the needed hardware and offer it cheaper. I have been and my father has been involved in the semiconductor business from the start and a small fab is not outside of our capability's we are going to have hardened electronics one way or another. In fact this is a way we can diversify our efforts to ensure our place in space exploration. Hardened chips are needed that’s a fact.

Monroe

well if you're building your own that's completely reasonable! i'm against them only because they're overpriced. reducing the cost of hardened electronics would be a big contribution to the industry. i was just thinking shorter term for the actual n-prize, since that seemed to be what you were looking for parts for.

Terra, I'm glad we had this conversation and glad I asked. This is starting some new thinking and a small fab is a good Idea for the industry. We could start by having the lithography done and do the implanting and deposition ourselves. At least now we are looking at that option and I like how that looks for us. We can build our own implanter and deposition equipment for function and specialization. "I think it's clobberin time for these guy's!" I think we need a clean room in the garage anyway No I'm serious this may be important for the industry and good for our business plan.

Monroe

I do not know enough about sapphire substrate, implanting that or how much it cost for wafers.
I wonder how hard it would be to grow our own.
Charge me 70k for a computer I don't think so! And we shouldn’t be paying that much that's a huge problem and one that can be remedied.

EDIT: Some of the assumptions I made here are incorrect however the Idea is sound and we are looking into it right now.

Here is a correct explanation of what we are talking about:

Silicon on sapphire (SOS) is a hetero-epitaxial process wherein a thin layer of silicon is 'grownâ€

You could submerge a processing board in coolant as well. That would significantly boost its protecting at a mass cost. The Mass cost would have to be offset by loswer system cost, compactness, robustness and a better thermal envelope.

There are very good points about the thermal issues and we will address that in the near future right now we need to reduce this cost problem. This is really nuts and really we can do something about it. Thanks for the input we need all the input we can get.

Monroe

Like So

I was thinking about making a 9 Inch box (one inch thick) with a Via dual Nano board. That gives you two physically seperate CPU's. Fill that little box with non-conducting Oil and you have a pretty secure little box.

If you are really really good at making your own semi-conductors then engineering a custom Bridge chip to lockstep two or three cheap CPU's in hardware with seperate ram for each would be pretty cool.

But that is how you get to $70,000.

Well I'm not a chip designer maybe a chip maker. I did look into a small computer encased in thermoplastic with pheltner cooling for speed several years ago, These devices may be useful for dissipating heat to a medium to be carried to an exchanger(like helium). The fluid in your box would still get hot and I don't think the heat could be conducted away good enough in space without an exchanger but it depends for sure. An entire computer would require the fluid to be passed to an exchanger im sure.
I do not at this time know enough about waste heat dissipation in space and I'm counting on the new Aerospace Engineers on the team to figure that out and explain it to me as we develop those systems. Stew is one of those guys.



Monroe