Could a Teensy survive a trip to the moon?

[jeffreytcash]

I'm currently working on a project where we are planning to brew beer on the moon by the end of 2017 by piggybacking on the lunar rover TeamIndus is attempting to win the Google Lunar XPRIZE with.
Here is an article from PopSci that gives a basic overview of our project, Team Original Gravity.

I am in charge of software development and helping with hardware development, so for our prototype, we are using Teensy 3.6s to process and log experimental data from test brews to a microSD card (Before the project started, I had a bunch of Teensys laying around and I figured they would be easiest way to capture our experimental data.).

Originally my plan was to use the Teensy 3.6 for the prototype, but select some radiation-hardened microcontroller for the real, moon-bound device. However, TeamIndus has informed me that the expected radiation dose is <10 krad for the entire mission, and I learned from this Wikipedia article that "normal commercial-grade chips can withstand between [5 and 10 krad]."

After learning this, I am partial to sending a Teensy to the moon if feasible, as I love working with Teensys and would love to give PJRC some publicity. The real device would not need the microSD slot, so I would probably go with a Teensy 3.2 instead.

This leads me to my question:
Would a Teensy 3.2 or 3.6 survive a trip to the moon, where it will encounter at most 10 krad? Is there any information about the effects of ionizing radiation on Teensys, or Freescale's MK20DX256VLH7 or MK66FX1M0VMD18?
The device will be shielded by multiple layers of aluminum and steel, so the dose will likely be much less than 10 krad.

Our project requires a microcontroller with a good amount of processing power for DSP tasks, a high-resolution ADC, and a high-precision, temperature/supply voltage compensated capacitance sensor, so the Teensy 3.2 and 3.6 really fit the bill. Additionally, if the Teensy is not suitable for the radiation dose, can anyone suggest an alternative that I could look into?

 

[Constantin]

My recollection is that you may run into issues re processor longevity. Doubt anything will be published on consumer grade chips because Astro-rated stuff tends to be special order, gold plated. I seem to recall the shuttle running on ancient technology for that reason.

Hence, my suggestion is redundancy. See if you can implement two or more teensy' CPU as a group, similar to the space shuttle enjoying multiple flight computers. Then use a hardened CPU to decide if the output from the teensy makes sense, reset if it doesn't. (Ie implement some sort of CRC on the operation of the teensy and then communicate via serial)

I'd pay particular attention to the clock, i.e. Wonder if a hardened oscillator may help.

 

[Frank B]

Or use the on-chip oscillator - crystal disabled.
I doubt that the sd card is a good solution. Vibrations during start, and generally, they are not very good and have bad blocks over and over which are disabled/managedby the on-board sd card cpu(yes, an additional cpu...) ...a massproduct, cheapest possible production. The "quality" cards are not really better.
Maybe SPI Flash is a better solution.

 

[Brooks]

Keep in mind that rad is not a "10k and it dies" sort of thing. It's a curve; some devices fail after 1k and some are still alive after 50k. And the failures can be very subtle. You'll want to checksum both flash and ram blocks on an ongoing basis, as much as you can. And build in self-diagnosis as much as possible. Keep multiple copies of the code in flash, and keep running checksums on them. It gets tedious pretty fast. Add as much shielding around your electronics as you can get away with, weight- and space-wise. I was always surprised how short a life-span even space-rated electronics have.