For many applications, you have a choice: use a radiation-hardened part that inherently meets your requirements, or use a cheaper/faster/better-performing COTS part with system-level mitigation (shielding, redundancy, watchdog timers, scrubbing, etc.). The right answer depends on a lot of factors, and reasonable engineers disagree.
Factors favoring rad-hard parts:
- Simplicity. A rad-hard part that meets requirements out of the box means no mitigation design, no additional analysis, and no additional testing of the mitigation effectiveness.
- Destructive SEE risk. For SEL-susceptible parts, mitigation (current limiting + power cycling) adds complexity and recovery time. A SEL-immune rad-hard part eliminates the problem. For SEB/SEGR, mitigation options are even more limited.
- Safety-critical applications. If a radiation-induced failure could compromise crew safety or mission success with no recovery path, the simplicity and reliability argument for rad-hard gets very strong.
Factors favoring COTS + mitigation:
- Performance. Rad-hard parts are typically 2-5 technology generations behind commercial state of the art. If you need the processing power, memory density, or conversion speed of a modern COTS device, rad-hard alternatives may not exist.
- Cost. A $5 COTS microcontroller vs. a $15,000 rad-hard equivalent — even with mitigation hardware costs, the economics often favor COTS.
- Availability. Rad-hard parts have long lead times and limited sources. COTS parts are available from stock.
The analysis challenge: How do you demonstrate that COTS + mitigation provides equivalent (or adequate) reliability? This requires understanding the SEE rates, the mitigation effectiveness, the recovery time, and the system-level impact of uncorrected events. It’s a much more complex argument than “the part is rad-hard to X.”
What’s your program’s philosophy? Do you have a decision tree or criteria for when to go rad-hard vs. COTS? How do you present the trade to a review board?