Radiation considerations for electronic design --- books?

Do a Google search on FPGA and radiation exposure, some interesting articles there and approximate failure rates.

Not aware of any books, but NatSemi has published a fair amount on this. Go to:

Barry Lennox

my $0.02: from a buddy who did a lot of smps for space apps, radiation can reduce Vth of power MOS, and even make it negative! So use lots of negative bias (bipolar drive) to ensure they stay off.

Cheers Terry

Isn't MOS banned for space apps due to possible contamination? I know cadmium is because it corrodes titanium.

Terry is speaking of NMOS devices. PMOS devices Vth is increased by radiation, in a negative direction, turning the device off so bias is not needed but higher drive levels are. Cheers + HNY, Harry

Thanks. There's at least one manufacturer of FPGAs who we have experience with and who has the information available.

Best regards, Spehro Pefhany

Will do so, thanks. .

Best regards, Spehro Pefhany

Good point. It appears that Vth can go in either direction (or first down then up with increasing exposure), that breakdown voltage can be negatively affected and that the drain leakage current increases. I have not found anything yet that correlates the effects to the precise construction of the power MOSFET.

Best regards, Spehro Pefhany

Hi, Harry:-

So I see. Do you know if there is any particular type of MOSFET that is more or less sensitive to radiation damage? Can we get this info from the manufacturers? I don't suppose it's a high priority for them to provide it, considering the low quantities of parts that will end up being used.

Best regards, Spehro Pefhany

This stuff probably is banned for that reason:

Best regards, Spehro Pefhany

P.S. There are a bunch of books that sound plausible on Amazon etc., just wondering if anyone found them useful. I know some of you have done designs for space and LEO.

Best regards, Spehro Pefhany

I put the first power MOS into space... hexfets in the Shuttle.

...Jim Thompson

IRF has radiation harden power MOS but you pay X10 more than the same "soft" device. Search their site for more info. A lot of data has been taken by users of rad-hard devices but it is difficult to locate. I have not done a rad-hard job in a while and tend to shy away from them. It like designing back in 1980 because those devices could stand more radiation. Normally the customer gives you a short list of IC's that he has characterized under radiation that you can use, LM124, LM139 and the like. Voltage references tend to be very susceptible. Have fun, Harry

Thought so. I should have been picking your brain over your killer saketinis (sp?) last Friday, but only heard about this project yesterday evening. ;-)

Best regards, Spehro Pefhany

Using off-the-shelf CMOS is tricky. In custom stuff I simply build-in current limiting to avoid destruction unless, of course, I have the luxury of an SOS or SOI process.

Coming back up in a known/desired state is usually the trickiest part.

...Jim Thompson

You could try nasa's klabs.org. iirc they have some studies...

Chris

You have to watch out for specmanship with these statistics.

The average space environment isnt too bad, but there are occasional "cosmic rays" with really high energies, some higher than can be simulated on Earth. IIRC there was one cosmic ray measured at enough energy to have moved a baseball!

So a stat of say 10% leakage degradation at 1000 Rads over 1000 hours isnt terribly meaningful. That number doesnt tell you anything about what happens when one 800MeV particle hits your 74HC04.

It's like saying nobody is affected by traffic accidents as the average person gets hit by 0.00003 of a car per year.

You have to design in error-detection and correction circuitry if you want the electronics to survive a serious cosmic ray.

We have a facility here that may be of interest:

I know that companies have come here to test semiconductors - don't know if any info is publicly available, but you could contact the person mentioned on that site for more info.

Well, the first MOS anyone could talk about, maybe.