0
\$\begingroup\$

What problems, if any, are to be expected when using 2N4037 metal can PNP transistors that were installed 40 years ago?

I can't divulge much about the design, but transistors are in a voltage regulator that was a spare that sat on the shelf for nearly 40 years, with short yearly tests. The other components (metallized mylar caps, carbon comp resistors, diodes) all seem to still hold their values within tolerance and work as expected. There are no electrolytic caps in the design.

Are there any common failure modes or pitfalls with these parts at this age and usage level?

\$\endgroup\$
  • \$\begingroup\$ Was it tested recently? Any issues present? \$\endgroup\$ – anrieff Feb 25 '20 at 19:38
  • 1
    \$\begingroup\$ 1. "NB: This is 'out of my head' - it makes sense to me based on 'stuff I have seen' over decades but has no formal basis: Hermetically sealed" components may have been 'challenged' over that time period. "Baking it gently" may be in order. Not so got as to damage any components, but towards the upper end of its rated storage temperature range. Even maybe in two stages - first a somewhat above ambient for hours to some days, then hotter. || 2. Rub the cans gently with fingertip and fingernail end. Do they have a rough 'lumpy' feel. Tin whiskers will grow in well under that time and vary ... \$\endgroup\$ – Russell McMahon Feb 25 '20 at 20:37
  • \$\begingroup\$ @RussellMcMahon, Thanks for the input. Tin whiskers will grow without the unit being powered? I've seen tin whisker issues with RoHS solder, but in more or less constantly powered devices. I am no tin whisker expert. \$\endgroup\$ – schadjo Feb 25 '20 at 20:50
  • 1
    \$\begingroup\$ @schadjo Yes - tin whisker growth is often (always?) independent of powering. Here are many examples - powered and otherwise. | \$\endgroup\$ – Russell McMahon Feb 26 '20 at 6:40
2
\$\begingroup\$

These are hermetically sealed cans, so they're not much subject to environmental attack. Migration of the dopants over time can change the geometry of the layers, aging germanium transistors would sometimes improve in gain and frequency response with aging as the base layer thinned. That, plus migration of the metallization for the electrode attachments are the main worries, but these are fairly physically big dies, and if they'be been stored at a reasonable ambient, the change will be minimal.

MIL-HDBK-217 prediction of failure rates probably don't mean much at the extremes of projection of life and failure rates, but all the factors lead to the obvious conclusion - the transistors should be fine. It's the carbon resistor's I'd be more worried about. Film resistors would be better for extreme durability, but if the design is tolerant of some drift in values, it may not matter.

\$\endgroup\$
  • \$\begingroup\$ I agree that carbon comp resistors are the most suspect parts here, but if they all appear to measure correctly with a typical ohmmeter (we're using Fluke 287s), what other "invisible" problems might appear with ancient carbon comps? \$\endgroup\$ – schadjo Feb 25 '20 at 20:56
  • 1
    \$\begingroup\$ There's the possibility of cracking of the carbon and the overmolding material that might cause them to behave erratically under thermal cycling, but if they've had little thermal cycling so far from storage, they should be OK. 217 table 9.1 puts the base reliability as pretty good, but the temperature and power factors increase the failure rates much more dramatically than film or wirewound. \$\endgroup\$ – Phil G Feb 25 '20 at 22:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.