Big Hopes for Tiny Satellites

shelflife writes: "ST5, according to NASA, will usher in a new era of small, smart spacecraft. Why send a human into space when you can send a computer? And why send something almost as heavy as a UNIVAC if a laptop will do? Compact nanosatellites will have everything you'd want in a full-size, luxury satellite. They will have the attitudinal and navigational capabilities needed to maintain proper orbits, and they will be capable of complex, high-bandwidth communications functions."

Comments

Microsats flew some years ago

[isdnip]

This isn't entirely new. There were "microsats" flown in the 1980s, some sponsored by the Amateur Satellite Corp. (AMSAT), and some university sats like Webersat (from Utah).

With today's smaller and more powerful chips, of course, it's a lot easier to do more in a small package.

Re:Microsats flew some years ago

[jensend]

Well, the microsats sent out from my home state were fairly well one-purpose, one-use machines (amateur radio for webersat, rotation/attitude manipulation with tracking for the JAWSAT)- see here for an optimistic description. See here for other previous microsats. NASA's microsats, according to their page, "carry a wide range of spacecraft services including guidance, navigation and control, attitude control, propulsion, high bandwidth and complex communication functions," some of which are diagrammed on that page and its successor. With the previous story and the other ways in which NASA has exceeded expectations on almost all of their craft in mind, I think this is an idea whose time has come.

Radiation is a big problem, heat too

[Bruce Perens]

There are some significant challenges in building "smart satellites".

Solar radiation is an extremely serious problem for any computer in space. To be rad-hard, chips need to be made of silicon on sapphire, which means a $1 embedded processor suddenly costs twenty thousand dollars. This is not material cost, it's because the economies of scale in production of terrestrial processors are what drives the cost down. Nobody can afford sapphire RAM banks, and thus memories get a flipped bit per orbit, in general. The only way they keep working is that there is a "washing" process that scans memory and does ECC correction continuously. Shielding is simply too heavy to be practical (send up a lead-clad satellite, and your rocket becomes 10 times as large to boost the weight).

Because it's available in sapphire and is flight-proven, the microprocessor of choice for controlling satellites is the 1802. Remember the RCA Cosmac Elf? Most of you weren't born when that was a popular hobby computer

I was surprised to find that the Phase 3D satellite boots up with no ROM. Hardware loads RAM directly from a radio modem. They couldn't afford a ROM they could trust.

Heat is a problem, too. Heat sinks don't work so well without an atmosphere to carry away heat. You have to pipe heat around with heat-pipes filled with a phase-change gas, and then radiate the heat away.Bruce

Re:Radiation is a big problem, heat too

[Bruce Perens]

You have to get the heat to somewhere that you can radiate it. Also, it's sometimes 300F on one side, and -100F on the other side of the satellite, so you pipe heat around just to gain a degree of temperature stability.

Thanks

Bruce

Re:Radiation is a big problem, heat too

[Anonymous Coward]

Actually you don't need to shield the whole thing - just the parts that are sensitive to radiation. There are a lot of studies of radiations on the commerical grade FPGA (surprise !!).

As for ROM vs RAM, wouldn't a Mask ROM be more reliable than RAM ? I would assume you mean 6T or 8T SRAM not those 1T DRAM cells.

Re:Radiation is a big problem, heat too

[Anonymous Coward]

Bruce,

I would like to add that it is quite possible to fly commercial processors in space. It merely requires good system engineering. Space radiation is merely another source of faults. A number of techniques can be used to get much better performance out of the commercial processor than the rad-hard processor while keeping it free from errors and up in terms of availability.

For instance the Iridium constellation uses Power PC 603s. In side by side comparisons, most of the time, the commercial processor is the better choice.

Besides performance and price, schedule is the big concern for using rad-hard parts. Most rad-hard processors that are being developed now (Rad-hard Pentium by Sandia, Rad-750 by BAE Systems, etc.) are way behind schedule and their performance by the time their designs are finished will be too far behind what is current to be useful.

Re:Another good thing about being small

[dragons_flight]

According to this page, ground based stations can track things down to 10 cm (about 4 inches). Admittedly they would be harder to see by an astronaut, but they typically aren't responsible for spotting the things visually. So long as people on the ground keep track of the satellites, it shouldn't be a problem.

Satellite Page

[Anonymous Coward]

Here's a really cool satellite engineering page that lists a bunch of small satellites - micro, nano, and pico. Also check out the humor section down towards the bottom. There are plenty of points that are applicable across many disciplines.

http://www.ee.surrey.ac.uk/SSC/SSHP/

AW

Re:Lots of advantages to being small

[Coryoth]

I believe the benefits of small sats is that you can put several of them on a single launch vehicle, and reduce the launch costs that way. Of course you need a launch system capable of carrying a number of smaller payloads - which there aren't very many of right now.

As far as I know the best for doing that are not US built which is why Surrey Space over in England are making a killing right now in the nanosat category.

Jedidiah

Re:Lots of advantages to being small

[Graymalkin]

Surrey's technologies are still pretty experimental and rely in some pretty difficult formation flying. Nanosats don't carry much in the way of tranceiving equipment and if you read the datasheets they carry uncooled CMOS cameras and basically are good for inspecing other spacecraft they're in formation with. Other formation flying ideas are lots of small birds with one or two tranceivers each that fly in formation and can output at least 10 watts PEP. You could have one sat with an uplink receiver that beams data to the rest of the birds in formation and they downlink different pieces of the signal on their tranceivers. Basically breaking a single comsat into a bunch of small pieces so if one component fails you could deorbit it hopefully and replace it by pushing another bird into the formation or having a backup fill its place. Nobody's really done this yet because formation flying in orbit is still a complex procedure and until recently birds weren't smart enough to navigate themselves. Also keep in mind that Surrey's sats are still all experimental and are using equipment that hasn't been recognized as spaceworthy quite yet. Trying to sell someone a satellite that has a 10% chance of working after a good sized solar storm won't be too effective. I think it is sort of interesting and cool they're powering their birds with StrongARMs.

Re:Space-qualified electronics.

[Anonymous Coward]

Some devices experience SEL (Single event latch-up) and some don't. You can make measurements to determine which ones do and fly only those that don't. Or you can put a latch-up monitor (detects current spikes and resets) on board. Powering down during the refresh cycle would be ludicrous because refreshes happen so frequently. Powering down between uses would work though. Most utilization profiles have a large time of no use with short utilization spikes.

I believe that SEi (now Maxwell) produces a rad-hard 486 also.