0
\$\begingroup\$

I looked today at the datasheets of this unqualified and this qualified product (per the AEC-Q101 automotive standard) and found no substantial difference, but I assume some stress damages only the unqualified one. So, what is the most common qualification problem for a discrete NPN? From my limited experience, I would guess that repeated thermal cycling between extreme cold and hot temperatures causes some package discontinuity (regardless of what is inside the package: NPN, PMOS, CPU, etc), but I wish the manufacturer would then note this nicely in the datasheets (since I will only use this NPN in my climate-controlled lab).

This reminds of those "There is something here known to the state of California to cause cancer" signs I often see which then make no mention of what actually is cancerous. I'd like to decide for myself whether to pay more for maximum safety. So, can somebody help me understand the physical difference between these two products?

\$\endgroup\$
1
  • \$\begingroup\$ So, really, no one thinks I could discern any statistical difference between these two products in a blind test where I have many transistors from each lot, have years of time, and can do anything? \$\endgroup\$
    – bobuhito
    Sep 16, 2022 at 2:59

3 Answers 3

Reset to default
1
\$\begingroup\$

I don't see any difference between the chemical content for qualified or unqualified or the three sites where the chips can be packaged (Dongguan China, Hamburg Germany, Malaysia).

Could be just the quarterly testing done on the Q units. Someone has to pay for all that testing. And if they happen to get fallouts from extreme testing in a lot made in one location or with one batch of materials, those parts can still be sold without the Q.

\$\endgroup\$
7
  • \$\begingroup\$ Well there must have been a fallout/failure or the manufacturer would not have gone through the work of making a second version. But, any speculation as to what specific test failed? \$\endgroup\$
    – bobuhito
    Sep 15, 2022 at 22:26
  • \$\begingroup\$ No, not at all. I bet all the units almost always pass all the tests. But that's just a guess. \$\endgroup\$
    – Spehro Pefhany
    Sep 15, 2022 at 22:31
  • \$\begingroup\$ I don't get it. Why then wouldn't the manufacturer just add "automotive-qualified" to the original datasheet and be done? \$\endgroup\$
    – bobuhito
    Sep 15, 2022 at 22:33
  • \$\begingroup\$ Because then they'd be forced to pay for the testing themselves. By differentiating them, the automotive customer pays for the testing and they can sell the unqualified ones cheaper, to compete with cheaper parts from other suppliers. It's bidness. \$\endgroup\$
    – Spehro Pefhany
    Sep 15, 2022 at 22:37
  • 1
    \$\begingroup\$ @bobuhito I often see automotive parts specified with more restrictive operating voltage or temperature ranges, for example. So they are telling you you can get better reliability, but only if you keep the operating environment more carefully controlled. \$\endgroup\$
    – The Photon
    Sep 16, 2022 at 2:00
1
\$\begingroup\$

AEC-Q parts are typically the same as non-AEC-Q parts, but they go through additional testing (often you can find the details of that testing on the manufacturer's website).

If you require an AEC-Q part for your application you usually just pay the little bit extra for the extra guarantee of reliability afforded by this testing.

\$\endgroup\$
1
\$\begingroup\$

Here is an interesting article that indicates that the qualification has to do with fabrication and screening of products. The qualified/unqualified products are the materially the same. An example is what is called "latent defects". These are defects that under "normal" conditions do not fail, while under stressful conditions such as vibration, temperature, humidity et cetera the defect expands resulting in poor performance or failure before the expected lifetime.

The parts are screened for AEQ, but manufacturing processes may also be improved since the automotive use of semiconductors continues to increase.

\$\endgroup\$
2
  • \$\begingroup\$ I'd like to believe this, but other answers are instead now making me believe that there would probably be no way for me to physically differentiate two lots of NPNs, one from the qualified process and one from the unqualified process, even if I did a 20 year experiment with thousands of NPNs. \$\endgroup\$
    – bobuhito
    Sep 15, 2022 at 23:02
  • \$\begingroup\$ @bobuhito: If the demand can be filled by screening alone, then that is what will be done. If not then production must increase or the process improved to meet the demand. Pure economics. \$\endgroup\$
    – RussellH
    Sep 15, 2022 at 23:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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