Do manufacturers test transistors?

[Ptownkid]

So I was testing a bunch of bd139's from fairchild...and one of them of the four I was testing wouldn't give me a reading unless i reversed the test leads I had going to the emitter and base...does this actually happen is is there something else going on here?

 

[Dusty Circuit]

I don't think the machine that connects the dice to the pins make mistakes of that kind but perhaps the text was printed on the wrong side.

 

[Ptownkid]

Even the three indents on are the same side as the text...I'm kinda baffled...

 

[JohnRoberts]

Stuff happens...  In very high volume manufacturing individual finished products are not tested but the process is managed by SPC (statistical process control) where sampling and extrapolation to the whole keeps the process running optimally. In the early days of 6 sigma research they determined that 100% testing actually damaged more parts than not.

You may have a mis-branded part, unlikely to be wired wrong but who knows. Testing with a curve tracer could better reveal what's inside that transistor. I'd be inclined to just discard it.

JR

 

[Rochey]

Also make sure your buying from a reputable source. I've seen some terrible counterfits.

 

[Ptownkid]

They came from digikey, and when I reverse the wires I get the same Hfe as the other three without reversing wires...go figure.

 

[1954U1]

Even the three indents on are the same side as the text...I'm kinda baffled...

I second what JohnRoberts said..
If I were you, I would not trust the transistor.
Or at least I'd try it for other parameters too, Vbe, leakages, Vce sat.. a curve tracer would be ideal.
But too much work for only 1 cheap piece IMHO.

P.S. To be honest, if I were you I'd open it  ;D

 

[Ptownkid]

yeah, I have no intention of using the little bastard...

 

[okgb]

And ,  of course
anything from china , QC is extra

 

[evm1024]

Don't know about transistors but in IC we tested every die on the wafer so as to get rid of the bad parts. The cost of packaging was a significant part of the uP and memory we were makings total cost to manufacture.

I suspect that the cost of packaging on a transistor is several times the die cost.

SO some testing was done. It has been a while since I worked in the field but I also suspect that it was all automated at this time.

--Ethan