Non-rad-hard power MOSFET question - EDN

2022-05-20 21:29:24 By : Ms. Linda Cheng

A client who was somewhat given to wishful thinking wanted to use power MOSFETs that were not characterized for radiation environment service in a radiation exposed power supply application. It was like asking if Play-Doh could be useful for making automobile tires.

In short, to harden or not to harden should never be a question regarding using semiconductors in a radiation environment.

I would like to share a communication with you as follows:

It took a day to get back the answer, but the answer was clear and unequivocal: 

John Dunn is an electronics consultant, and a graduate of The Polytechnic Institute of Brooklyn (BSEE) and of New York University (MSEE).

Yes Radiation and operational semiconductors can be trouble some requiring hermetic metal screening (metal to glass to metal insulation). But the biggest issues come from the effects on plastics (which are long chain polymer constructions) and best advice is to refer to the suppliers data sheets on the material and its long term exposure to radiation and you will need the type of radiation.

Note do not expect the solution to the application to be low cost expect at least 10 times to 1000 times more expensive as some parts will need shielding from the exposure level which usually involves metal casing (not thin) and expect issues with wire insulation and after life disposal and maintenance procedures. Enjoy!

Attempting to use commercial non-radiation hardened semiconductors is like playing Russian Roulette. You may find a device that works but the next time you buy a new batch it may fail miserably.

It is not practical to shield electronics in a space environment and in fact, the shielding may produce may secondary particles like muons that will adversely affect the devices.

TI did compared the TID (Total Dose) tolerance of some common ICs. They found a HUGE variance in the radiation tolerance, as much as a factor of ten for the same device. TID performance is strongly dependent on process variation. Since radiation tolerance is not a consideration for commercial parts, no manufacturer bothers to characterize the parts.

Since commercial parts are manufactured for maximum yield to keep cost low the process for a given device may be tweaked over time to keep yields high. Another factor is that many, if not most, commercial parts are are made in multiple fabs. No two fabs are identical which leads to more variation in radiation tolerance.

Single Event Effects (SEE) are another facet of radiation tolerance. Some devices may perform admirably for TID but fail for SEE.

Unless you are prepared to source parts from a single known fab and periodically test the parts for SEE and TID I would not consider using commercial semiconductors of any kind. Some type of radiation induced failure is almost guaranteed.

You must Sign in or Register to post a comment.