A Totally Worthless - But Interesting - Circuit

[PRR]

> Reportedly allowing be junctions to reverse zener degrades beta (hfe) and noise performance

A precision matched pair will become un-matched. A selected hi-Hfe part may stop being the hi-Hfe that you selected. A low-noise part may well become less low noise.

For jellybean switches and most amps, a little zenering never hurt nobody. mediatechnology's trash-act was a waste of 3 cents. It wudda worked fine in almost any use except a lo-Z mike amp input stage.

 

[JohnRoberts]

Apparently the noise and beta degradation can be quite severe, so I'd be inclined to put them in round file, too.

Another tidbit, this damage appears to be reversible with elevated temperature or high current annealing, but that too seems like a little too much effort to rescue a 2n3904..

JR

 

[bcarso]

I revisited Motchenbacher and Fitchen just now, and they cite a loss of 20% of beta for a single reverse-current pulse of 1 mA.

 

[JohnRoberts]

If you revisited M&F you saw the entire discussion.

"The damaging effect is observed to be proportional to total charge flow and the logarithm of time"

".. consequently, a steady reverse current causes the same damage in 1/10th the time shown in fig..."

-------
I am not attempting to predict the amount of damage for a single avalanche event but the extremes described by M&F were large enough to be worrysome, i.e. noise voltage double, noise current 10x. While the focus of discussion was on noise, on the next page a plot of hfe suggested a similar decline in gain.

Indeed 20% reduction if the device was avalanched "only" one time and "only" for 1 sec. My sense is that the damage is cumulative so besides the reverse voltage threshold current and time in avalanche only determines rate of decline not the final state. I don't recall the exact conditions of Wayne's experiment.

For the few pennies, I'd throw them in the round file, and I always clamp across b-e to protect my low noise transistors from reverse avalanche in input circuits.

YMMV

JR

 

[Guest]

So, measurement of b-e breakdown voltage is destructive...

 

[JohnRoberts]

If measured it should be very brief and very low current. Damage can be reversed by elevated temperature or high current annealing. I see little value in knowing that particular data point for low noise parts.

JR

 

[Guest]

[quote author="JohnRoberts"]If measured it should be very brief and very low current. Damage can be reversed by elevated temperature or high current annealing. I see little value in knowing that particular data point for low noise parts.

JR[/quote]

I mean in general. To bring charge carriers back temperature increase is needed. Looks like technology of making emitters is poor.

I saw transistors that had nearly equal collectors and emitters. There were Soviet small signal transistors KT208 and KT209.

However, I would not expect from them to dissipate the same power when collector and emitter were swapped.

 

[JohnRoberts]

I don't know the physics, something about surface states.

There are sundry special transistors, some designed to be operated backwards for lower saturation voltage, but beta is not symmetrical when using b-c as a b-e AFAIK. The reverse avalanche characteristic may be different too.

JR

 

[Guest]

[quote author="JohnRoberts"]I don't know the physics, something about surface states.

There are sundry special transistors, some designed to be operated backwards for lower saturation voltage, but beta is not symmetrical when using b-c as a b-e AFAIK. The reverse avalanche characteristic may be different too.

JR[/quote]

Sure they can't be made equal!

Usually collector is the bottom layer for lower thermal resistance. And it's square is bigger.