Measured reverse base damage of NPN

Were they new transistors out of the box? Can the graphs be explained by something other than physical damage to the structure like charge-storage effect or some kind of wear-in process that would occur when a new transistor is used in any configuration?

Forward-active beta varies for different amounts of constant current thru the base-emitter junction there's nothing unusual about that. AFAIK it doesn't matter which direction the current is going to affect it.

Yikes.

Cheers

Phil Hobbs

recovers.

...

Don't have a heating chamber at home, using the one in kitchen for electronics causes distortion and our sauna at home has a hard limit

next to plastic.

hFE curves before heating (hFE~=105 for Ic=1mA, Vce=3V)

The performance drop at low currents is especially gruesome!

After heating (hFE~=177 for Ic=1mA, Vce=3V)

Heat cures!

Yes, they were out of the box. The effect was visible even after hours no backwards BE current, I would expect any charge storage to clear within that time - on the other hand heating seems to help.

Here's one paper I found about the subject - it mentions 'interface states created in the oxide over the emitter-base junction during the stress' as being one generally assumed reason for additional base current.

I guess the interface/surface states could be cleared up by heat ?

So - add a heater or cover your bases!

Pease says someplace that Widlar used to use his cigar on metal can transistors for exactly this.

Cheers

Phil Hobbs

Fun... here's a silly question. Can you cure 'em with C-E voltage/current heating?

Heal thyself, physician.

George H.

ge

run

There was reputed to be an HP pulse generator which worked much better afte r the service engineer stubbed out his cigarette on the output transistor.

Other people's cigarettes worked equally well.

You may find that the amount of measurable damage depends on the product of time and revers current, not unlike radiation damage or ionic contamination.

It is well known (since before the Apollo mission) that the damage can be annealed out.

Yeah it just seemed like an awful large droop for such a small amount of charge which is why I was trying to explore whether it could be accounted for by a measurement error.

The resolution of the latter graph is not that good, but to me it looks like the 1uA current causes damage slower.

downloads.hindawi.com/journals/apec/2001/053209.pdf

Good work! Last year I did a few simple experiments too and noticed the same rapid initial decline in beta which then tends to level off. But I never thought to look for changes in breakdown voltage.

piglet