Transistors - die pictures

[Noopy]

These two transistors have the same collector

That's right... It would have been enough to integrate a second collector pad instead of a whole transistor... 

[exe]

My friends, look what I found: P503, MP-16B, MP14-A and 2T803A (the big can) . The first one seems to be one of the first germanium bjt produced in the USSR. Accroding to http://www.155la3.ru/p501.htm, they were developed in 1958-1959, and were in production for 5-6years. Rumors said they were discontinued due to extremely low yield of 1.8%.

It seems my P503 is not doing well or I'm measuring it wrong: . What should I do with it? Shall I keep it for children, donate somewhere, or open it?

PS This one was donated to me in late nighties by an ex-EE engineer among with many other old parts. It was sitting in a storage room until last year when I found it while traveling to my home town. I brought a few Soviet bjts back in hope to put them into use. Not sure how to use them. Build a distortion pedal for the guitar?

PPS I wanted to measure the diode drop, so I used a dmm that has test voltage of 3.2V. From my measurements, the forward voltage is 0.244, the reverse is 2.7, so it seems they broke down. According to datasheets, those germanium transistors are only rated up to 3V of reverse voltage.  Does it mean I killed them all? 

PPPS the year of production of P503 in question is 1963.

[Noopy]

I quite sure you can't kill such a transistor with the diode tester. The current should be low enough.

Perhaps your transistor tester has a problem with the probably higher leakage currents?

[Noopy]

Today I have an older part for you, the 2N1561, a high-frequency-germanium-diffusion-alloy-mesa-transistor. 




Nice packing. 




TO-107




Absorbent cotton? 




The die is 0,57mm x 0,67mm.
It seems that the sawing of the dies was a bit of a problem. The edges are a little bit splinted.




I assume they had a p-doped substrate, diffused a n-doped base-layer on top of it and then placed a p-dopant and a n-dopant on top of it. After some baking yout get under the p-dopant an emitter and you can use the rest of the material to contact the emitter. The n-dopant gives you the base-contact.
Then some MESA etching and you get a nice clean (ok a bit shaky  ) base-collector-edge.


More pictures here:
https://www.richis-lab.de/Bipolar17.htm

 

[exe]

Quite some effort to produce a transistor. No wonders they used to be expensive.

[Noopy]

Today I have again a more simple germanium transistor for you:
ACY38 built by COSEM (France) which merged later with SESCO to SESCOSEM (=>Thomson=>SGS Thomson=>STMicroelectronic)






 
Paint stripper is good to remove this silicone oil based thermal paste. 




Transistor is covered with some silicone stuff.




Unfortunatelly the wires were pulled of while removing the silicone.
But here you can see an interesting thing. The surface under the indium pill looks like it was machined. Actually that´s quite reasonable. A smooth surface gives you a better junction than a lot of roughness (different base thickness and varying doping).




Emitter...


More pictures here:

https://www.richis-lab.de/Bipolar18.htm


 

[Noopy]

Well that´s a really old 2N3055:




Who knows Solitron? 




Nice! You don´t need a datasheet to connect the transistor. 






A classical transistor structure.
The solder is not applied evenly. That doesn´t look very promising.
The emitter connector seems to be dangerously long. It looks like it can short to the base... 








And breakdown light of course!  => -16V / 0,5A




It´s probably a hometaxial transistor. The surface (emitter) is partly etched down to connect the base.
You can see that the etched base area is smoother than the emitter area.


More pictures here:

https://www.richis-lab.de/2N3055_06.htm

 

[Wolfgang]

Hi Noopy,

nice, but what you see here is not a second breakdown, but the breakdown of the base emitter diode.

[Noopy]

Thank you Wolfgang! 

Of course that´s not a second breakdown! I have corrected the text...
I had a long day... 

[Wolfgang]

Hi Noopy,

in fact it would be nice to see a second breakdown, but I guess you need a high-speed camera to get this on film.

Thanks for your good work !
 
  Wolfgang

[Noopy]

I would need a high speed super-macro-camera with enough focal distance to protect the camera. That sounds expensive! 

Thank you for the recognition! 
I still have some interesting parts in stock. 

Best regards,

Richard

[T3sl4co1l]

Also 2nd breakdown probably doesn't emit any visible light, so you'd need a high framerate IR camera at that!

Tim

[magic]

That being said, you could still try basic avalanche breakdown of the BC junction. Just need a resistor-limited HV generator.

[Noopy]

Also 2nd breakdown probably doesn't emit any visible light, so you'd need a high framerate IR camera at that!

Tim

I had in mind a second breakdown with high power. Getting that on a film would be pretty impressive even without IR sensitivity. 

That being said, you could still try basic avalanche breakdown of the BC junction. Just need a resistor-limited HV generator.

I once tried a breakdown of the BC junction of a transistor with low voltage rating (can´t remember which one). But I wasn´t able to see any light.
Perhaps that´s because most of the BC junction is covered under "a lot" of silicon. 

Unfortunatelly higher breakdown voltages generate more heat so the transistor dies very quickly... 

[magic]

Did you verify that there was current flow?
I tried measuring breakdown voltage of some BC857 once and found it to be over 100V, which was the limit of my generator. Basically, I applied 100V and nothing happened whatsoever.

Perhaps that´s because most of the BC junction is covered under "a lot" of silicon. 

That's possible, although usually the edges of the junction reach the surface.

[Noopy]

Yes I tracked the current. I finally found one that broke down <80V. I think it was a small HF Transistor.

Perhaps I should try it again... 

[Noopy]

With "Perhaps I should try it again..." I wanted to say "I have to do that instantly!" 

I took the 2N2857 with a maximum collector-base-voltage of 30V:

https://www.richis-lab.de/Bipolar16.htm


Beginning at 57V(!) the current through the collector-base-junction reaches 10µA.





=> 50µA
You can see a small glow in the upper right corner of the base rectangle.




=> 100µA
The lower right corner starts to glow a little.




=> 250µA
Now we see the corners! 




=> 500µA
The left corners start to glow.




=> 1mA
The electrical field is highest in the corners of the rectangle and so the corners are illuminated first.




=> 2mA




=> 3mA
Then the right edge is lightening up.




=> 5mA / 75V
I can´t go any higher. Probably with more power loss (>375mW) the 2N2857 would die anyhow.
It seems the area under the emitter contact is dark. The emitter isn´t connected but perhaps/probably the emitter contact somehow affects the electrical field. 


 

[T3sl4co1l]

The emitter most definitely affects things -- C-E has a lower breakdown than C-B, effectively the C-B leakage is multiplied by hFE so runs away at a lower voltage (Vceo vs. Vcbo).  In the transition region between these two extremes, breakdown depends on e.g. B-E resistance; and breakdown can become exceptionally noisy, so much so that it can effectively switch "on" in a fraction of a nanosecond!

(I've observed avalanche switching in every transistor I've tested; the width of the region (in terms of what range of R_BE works reliably) varies, with most epitaxial parts being poor.  2N2369 is the classic, but 2N3904 also seems to do well.  Power transistors don't handle any more current before burning out, perhaps because avalanche only occurs in a small part of the whole junction, and isn't able to propagate over the full width.)

If the glow persists in Vceo or Vces mode, I wonder if it has any physical significance -- sufficient light output could perhaps cause the junction to ionize much sooner, and more broadly, than it would due to carrier diffusion alone!

Tim

[Noopy]

If the glow persists in Vceo or Vces mode, I wonder if it has any physical significance -- sufficient light output could perhaps cause the junction to ionize much sooner, and more broadly, than it would due to carrier diffusion alone!

I just did a short test of Vceo.
At ~45V over emitter-collector and ~10mA you can´t see any light.

I would have guessed to see at least some glowing... 
Perhaps the breakdown ist more chaotic and distributed over the whole die... That would be a good explanation for  more noise... Just speculating...

Have to try Vces but no time left right now.
I assume the collector-base-junction will glow in that case but since a lot of current is flowing to the emitter I won´t be able to raise the current high enough to make the glowing visible before the transistor dies.

[magic]

Another factor is beta: at 10mA collector current, only some 0.1mA may be flowing by means of breakdown and the rest because of the normal current gain of the transistor.

[Noopy]

Another factor is beta: at 10mA collector current, only some 0.1mA may be flowing by means of breakdown and the rest because of the normal current gain of the transistor.

I quite agree with you! 

[Noopy]

Hm... Somehow I managed to kill the 2N2857...
Unfortunatelly I can´t say what went wrong. I was in a hurry. That´s never a good thing.   

[magic]

Dunno, looks like you shorted out a PSU with that poor thing

Reminds me of that time I tried to weld a thermocouple with a laptop brick.

[exe]

May be it's only bound wires are broken. Is it possible to solder new wires to the die?

[T3sl4co1l]

The die is discolored.  That's not just ash on top, that's an obliterated what-used-to-be-a-transistor.

Tim