3 ms·

> A radiation hardened equivalent of a $20 FPGA can be something like $20,000 Has anyone actually tried putting up non-rad hardened equipment to measure how th

> A radiation hardened equivalent of a $20 FPGA can be something like $20,000 Has anyone actually tried putting up non-rad hardened equipment to measure how they perform? The Mars helicopter wasn't RAD hardened and used off the shelf parts & succeeded and the Mars atmosphere is not thick enough to meaningfully block the amount of cosmic rays hitting the surface. I think NASA doesn't do a good job sometimes tolerating risk and then everything is treated as needing safety-levels of risk mitigation without considering that a 1/100th cost reduction will not generate as much in parts failures.

dr_orpheus 2y ago
Yep, they do! I had some of this discussion on a thread talking about the Mars helicopter here that Goddard does a lot of radiation testing on commercial chips. https://news.ycombinator.com/item?id=39175423#39182421 https://news.ycombinator.com/item?id=39175423#39182421 Lots of the new space, and smaller satellite companies use a lot of commercial parts. A lot of the flight data has shown even better results than the radiation testing (possibly due to added stress of testing at higher rates vs low rates over longer mission duration). Generally speaking most of this is in LEO with a pretty low radiation environment. Whereas the area around Jupiter is one of the worst radiation environments in the solar system due to the radiation belts (like the Van Allen belts on steroids). This page on the Juno Radiation Vault says the spacecraft is exposed to an anticipated 20 Mrads of radiation. Whereas spacecraft in LEO are exposed to 0.1-10 krads per year depending on the orbit. Also a fun fact, this is with Juno trying to limit exposure to the radiation belts as much as possible. [1] [0] https://en.wikipedia.org/wiki/Juno_Radiation_Vault https://en.wikipedia.org/wiki/Juno_Radiation_Vault [1] https://en.wikipedia.org/wiki/Juno_(spacecraft)#/media/File:Juno_trajectory_through_radiation_belts.png https://en.wikipedia.org/wiki/Juno_(spacecraft)#/media/File:...
- vlovich123 2y ago
  Ok. That's a good point about the radiation belts. I hadn't considered that Jupiter's massive magnetosphere captures & concentrates a huge amount of solar energy & Juno is very close to it. Thanks for the additional info!
fecal_henge 2y ago
Lots of work on this in the High Energy Physics community. Big experiments can design their own rad hard silicon, but everyone else has to test. Lots of space rated electronics is also qualified at far lower levels than we need. Upside ia that people are happy to share what they find: https://arxiv.org/pdf/1912.01742.pdf https://arxiv.org/pdf/1912.01742.pdf https://twiki.cern.ch/twiki/bin/view/Main/RAD-HARD-COMP https://twiki.cern.ch/twiki/bin/view/Main/RAD-HARD-COMP
dr_orpheus 2y ago
> I think NASA doesn't do a good job sometimes tolerating risk and then everything is treated as needing safety-levels of risk mitigation without considering that a 1/100th cost reduction will not generate as much in parts failures. I do absolutely understand this impression of NASA. But I also think it gets inflated because the highest profile NASA missions that you hear about in the news are the most expensive and least risk tolerant missions. But there is pretty large spectrum in terms of cost caps and risk tolerance to NASA mission classes. I think generally in order of descending cost/risk tolernace it is: Human Spaceflight, Flagship (i.e. JWST, Mars Rovers), New Frontiers (Juno falls here), Discovery, Explorer, Mid-Explorer (MidEx), Small Explorer (SmEx), Venture. For an example in the Venture class you can look at something like CYGNSS. Constellation of 8 spacecraft to better understand dynamics of hurricanes by looking at ocean wind speeds. This is done by mapping doppler delay of reflected GPS signals off of waves in the ocean. Important science, super cool technology with mostly automotive grade parts. ~$150 million for the whole mission that lasted about 7 years.
chgs 2y ago
If you’re spending nearly $1m on something to launch a camera do you really want to take the risk, so you can launch for 950000 instead of 970000?
- vlovich123 2y ago
  It’s not risking the entire mission. Just the part the camera is should be risked. And from that perspective attach a camera to each thing you do and you’ve got a good shot of it working most of the time rather than spending outsized amount of resources to guarantee success. This is very similar to how SpaceX does rockets by the way so even if the cameras are useless you have a good shot at learning more things.
  - chgs 2y ago
    The $1m isn’t for the mission Think of it another way, would you pay $99 for a camera which will work or $97 for a camera which might work.

dr_orpheus 2y ago

Yep, they do! I had some of this discussion on a thread talking about the Mars helicopter here that Goddard does a lot of radiation testing on commercial chips. https://news.ycombinator.com/item?id=39175423#39182421 https://news.ycombinator.com/item?id=39175423#39182421 Lots of the new space, and smaller satellite companies use a lot of commercial parts. A lot of the flight data has shown even better results than the radiation testing (possibly due to added stress of testing at higher rates vs low rates over longer mission duration). Generally speaking most of this is in LEO with a pretty low radiation environment. Whereas the area around Jupiter is one of the worst radiation environments in the solar system due to the radiation belts (like the Van Allen belts on steroids). This page on the Juno Radiation Vault says the spacecraft is exposed to an anticipated 20 Mrads of radiation. Whereas spacecraft in LEO are exposed to 0.1-10 krads per year depending on the orbit. Also a fun fact, this is with Juno trying to limit exposure to the radiation belts as much as possible. [1] [0] https://en.wikipedia.org/wiki/Juno_Radiation_Vault https://en.wikipedia.org/wiki/Juno_Radiation_Vault [1] https://en.wikipedia.org/wiki/Juno_(spacecraft)#/media/File:Juno_trajectory_through_radiation_belts.png https://en.wikipedia.org/wiki/Juno_(spacecraft)#/media/File:...

vlovich123 2y ago
Ok. That's a good point about the radiation belts. I hadn't considered that Jupiter's massive magnetosphere captures & concentrates a huge amount of solar energy & Juno is very close to it. Thanks for the additional info!

fecal_henge 2y ago

Lots of work on this in the High Energy Physics community. Big experiments can design their own rad hard silicon, but everyone else has to test. Lots of space rated electronics is also qualified at far lower levels than we need. Upside ia that people are happy to share what they find: https://arxiv.org/pdf/1912.01742.pdf https://arxiv.org/pdf/1912.01742.pdf https://twiki.cern.ch/twiki/bin/view/Main/RAD-HARD-COMP https://twiki.cern.ch/twiki/bin/view/Main/RAD-HARD-COMP

dr_orpheus 2y ago

> I think NASA doesn't do a good job sometimes tolerating risk and then everything is treated as needing safety-levels of risk mitigation without considering that a 1/100th cost reduction will not generate as much in parts failures. I do absolutely understand this impression of NASA. But I also think it gets inflated because the highest profile NASA missions that you hear about in the news are the most expensive and least risk tolerant missions. But there is pretty large spectrum in terms of cost caps and risk tolerance to NASA mission classes. I think generally in order of descending cost/risk tolernace it is: Human Spaceflight, Flagship (i.e. JWST, Mars Rovers), New Frontiers (Juno falls here), Discovery, Explorer, Mid-Explorer (MidEx), Small Explorer (SmEx), Venture. For an example in the Venture class you can look at something like CYGNSS. Constellation of 8 spacecraft to better understand dynamics of hurricanes by looking at ocean wind speeds. This is done by mapping doppler delay of reflected GPS signals off of waves in the ocean. Important science, super cool technology with mostly automotive grade parts. ~$150 million for the whole mission that lasted about 7 years.

chgs 2y ago

If you’re spending nearly $1m on something to launch a camera do you really want to take the risk, so you can launch for 950000 instead of 970000?

vlovich123 2y ago
It’s not risking the entire mission. Just the part the camera is should be risked. And from that perspective attach a camera to each thing you do and you’ve got a good shot of it working most of the time rather than spending outsized amount of resources to guarantee success. This is very similar to how SpaceX does rockets by the way so even if the cameras are useless you have a good shot at learning more things.
- chgs 2y ago
  The $1m isn’t for the mission Think of it another way, would you pay $99 for a camera which will work or $97 for a camera which might work.