¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

An odd bipolar transistor effect......


 

Also the reason why transistor junctions are lower leakage than glass diodes.

Sam Reaves
ARS W3OHM
Owner / Moderator of:
LeCroy Owners Group on Groups.io
Sencore Owners Group on Groups.io
Sprint Layout Group on Groups.io??
Pulsonix EDA Group on Groups.io
LPKF Owners Group on Groups.io
Electronics and Mechanical Hardware Design Engineering Manager
Staff Scientist Andritz Rolls Global Research Center (RETIRED)



On Thu, Nov 21, 2024 at 9:15?PM wn4isx via <wn4isx=[email protected]> wrote:

A lifetime ago, May 1974, a visiting professor to the EE program explained an odd 'side effect' in some transistors.

Photons are emitted when the electrons and holes recombine in the collector/base. And the photons could excite voltage production via PV effect in the base emitter junction.

He demonstrated the effect.

Normally this is a laboratory curiosity and we were told we'd never see any indication in the real world.
I wrote it down in my class notes, underlined and promised to look into the effect...and promptly almost forgot about it. Life was too filled with class work, then marriage, jobs, and hobbies...

?

Today a friend sent me a link...

?

The video is interesting and appears to demonstrate the effect.

Note: My 1980s 2N2222 barely worked, modern 2N2222 and modern 2N3904 worked extremely well. I intend to check every family of transistors I have.

?

I can't think of any practical application, but the effect is interesting.

?

Now to borrow a nightvision (starlight) scope and see if glass Ge and Si diodes emit any detectable light when forward biased.....

And somewhere I have some phototransistors, it will be interesting to see if they emit any detectable light when forward biased.

And will photovoltaic cells emit any detectable light when forward biased.

So many experiments, so little time....


 

On Thu, Nov 21, 2024 at 09:15 PM, wn4isx wrote:

Photons are emitted when the electrons and holes recombine in the collector/base. And the photons could excite voltage production via PV effect in the base emitter junction.

?

Below is written by the guy who created LTspice (Mike Engelhardt), which he wrote in the LTspice user's group about 15 years ago:
?
Bipolar BE breakdown is technically a failure mode also,
because it modifies the transistor doping and the gain
decreases.? BTW, if ground the emitter of an NPN transistor,
leave the collector unconnected, and reverse bias the base
to breakdown, the collector tends to go negative because
the breakdown creates photons which are converted back to
current in the BC junction.? Does anybody know the net quantum
efficiency of that process?? I did a quick check on a 2N3904,
and found it less than 1ppm, which was the limit of my
multimeter.
?
So I guess the photo emission and voltaic effect works either way: BC to BE in one case, or BE to BC in the other.
?
Andy
?


 

On Thu, Nov 21, 2024 at 11:30 PM, wn4isx wrote:
1N34 are pretty bad but they are germanium ...
Just an FYI -
?
I am told that "most" Germanium diodes sold today are not the Germanium diodes they originally were.? Instead, they are Schottky diodes, Silicon-based, packaged with Ge diode part numbers.? I guess the manufacturers figure that as long as they meet the datasheet specs, all is well and good and 'nobody' can tell the difference.? Apparently that is done not just by the counterfeit part makers, but by the legit ones too.
?
It's a bit disappointing.
?
Andy
?


 

On Fri, Nov 22, 2024 at 10:20 AM, wn4isx wrote:
I have 1N34's older then I!
Then chances are (in fact it is almost guaranteed) they are Germanium.
?
But if you bought a 1N34 today, it might not be.
?
Andy
?
?


 

On Fri, Nov 22, 2024 at 10:41 AM, wn4isx wrote:


Alas, it is locked behind IEEE's paywall.
?
:-(
?
Andy
?


 

There are transistors that have multiple emitters on the silicon all coming out to the package pin. (RF transistors)
It would be interesting to compare the red light from one of those to a standard transistor.
I have memory that the hard to get 2N5109 has multiple emitters, but I haven't been able to verify it.
Possibly the 2n3886 also has multiple emitters.
?????????????????????? Mikek
????????????????????