Transistors - die pictures

[Noopy]

Next "merchandise project" is a LTZ1000 coffee pot. 

Done: https://www.redbubble.com/de/i/tasse/Referenzspannungsquelle-LTZ1000-von-Richis-Lab/63614214.9Q0AD 

The LTZ1000 clock looks quite interesting... 

[Noopy]

Today I have a Dual-J-FET for you, the DN1682:






It´s a Dual-J-FET witch matched transistors.




It looks like they have cut the wafer in two steps. Often hat is done to reducing damage in the active area.




A normal J-FET.
45 61 probably gives the die a unique number for matching the right parts.


More pictures here:
https://www.richis-lab.de/FET06.htm

 

[SilverSolder]

What's the purpose of the black goo around the dies?

[T3sl4co1l]

Weird, it's common-gate?

Wait... are those not actually the can?  And if they were, there wouldn't be any need for epoxy in the bottom.  Could those be... gate pins, coined flat for mounting the dies on?  Then the epoxy underfill is not so much to insulate (the hermetic seal should be fine) as to control vibration (so it's cemented down).

Hm, is the can floating, then?

Tim

[Noopy]

What's the purpose of the black goo around the dies?

It hold´s everything together. The black potting holds the pins in place.

Weird, it's common-gate?

Wait... are those not actually the can?  And if they were, there wouldn't be any need for epoxy in the bottom.  Could those be... gate pins, coined flat for mounting the dies on?  Then the epoxy underfill is not so much to insulate (the hermetic seal should be fine) as to control vibration (so it's cemented down).

Hm, is the can floating, then?

There are seven pins: 2x Gate-Drain-Source and one for the can so you can connect the can to whatever you want.
The flattened pins are the gate pins and are connected with a bondwire to the edge bondpad.

 

[David Hess]

Weird, it's common-gate?

The connection to the gate is through the bottom of the die, so the die has to be insulated from the can.  Hybrid construction would take advantage of this to have only two wire bond connections on the top.

[exe]

Today I have a Dual-J-FET for you, the DN1682

I'm surprised, I expected to see two weirdly-interleaved fets on a single die.

[Noopy]

I'm surprised, I expected to see two weirdly-interleaved fets on a single die.

That depends on what you want: Here the thermal coupling is worse but the capacitive and resistive coupling is much less than with one die.

[Noopy]

Do you know "Elektronska Industrija"? It was an Yugoslavian company and they also built a 2N3055. 




The transistor is remarkable similar to one generation of the the RCA 2N3055: https://www.richis-lab.de/2N3055_05.htm




There is some fiber glued to the die. 




An of course it´s glowing! 
The breakdown voltage of the base-emitter-junction is quite high (-17V) because of the hometaxial construction.
Even with 1A there are some parts still dark due to the inhomogeneous structures.


More pictures here:
https://www.richis-lab.de/2N3055_13.htm

 

[exe]

Wow, I didn't know 3055 were so popular. So many companies were producing them...

[serg-el]

http://www.155la3.ru/kt818_19.htm

Russian 3055

[exe]

btw, you reminded me that I also have a few bjts to open, and, coincidentally, I had a hacksaw at hand. How do you say I can make them glowing?

[Noopy]

The 2N3055 was the standard power transistor and was widely used. Back in the day there were a lot of semiconductor manufacturers... 

The base-emitter-junction is glowing in breakdown. Just connect a negative voltage to the base-emitter junction. HF-transistors often break down at 6V but I had an old Siemens hometaxial first breaking down at 70V (!). Normaly you see something around 10-20V. Be sure to limit the current and you have to cool the part or be fast: 20V*1A=20W 

[exe]

What is this white stuff on the crystal? How to remove it? I tried IPA, got some mild results. Will try a pcb cleaner in the evening.

[Noopy]

You can find many different protection coatings. Often it's a clear coating or nothing at all.
White stuff sounds like a silicone. That's quite hard to remove. Silicone cleaner can help but doesn't dissolve the stuff. In the end you have to peel it away. A plastic screw driver or something like that can help. You have to be careful not to damage the bondwires.

[exe]

Here is the transistor. Probably, the best picture I could make. The camera is slow to take pictures, so at 10A it burned out faster than "shutter" was able to take picture. This picture, is, afaik, at 5A. At 10A it blinked, and then I saw little sparks on the crystal. I think voltage was about 20V, and it became hot very fast. After that it rejected glowing, and only produced some sparks and smell. Poor bjt, RIP.

[Noopy]

10A is a lot of current! 
Sometimes I go up to 1,5A but with more current death comes fast... 

[exe]

Sometimes I go up to 1,5A

Ouch

[Wolfgang]

Noopy, what about pulsed measurements ?

[Noopy]

I do manually pulsing.   
Up to know that was sufficient.

[Wolfgang]

Maybe, but you could use a lot more current if the duration is kept really low.
And its an interesting start problem for a computerized PSU / photography project. 

[exe]

BTW, my power supply can do pulses, I just thought I'll be quick-enough to do it manually. I was wrong. At 10A there was really little time to do the shot. I think it was about a second or less before brightness dropped, then sparks showed up.

[magic]

power dissipation = voltage drop · current

Just saying

[Noopy]

And its an interesting start problem for a computerized PSU / photography project. 

Absolutely! 

BTW, my power supply can do pulses, I just thought I'll be quick-enough to do it manually. I was wrong. At 10A there was really little time to do the shot. I think it was about a second or less before brightness dropped, then sparks showed up.

power dissipation = voltage drop · current

200W 

[Noopy]

The SL113 is an old Silicon-HF-Powertransistor built by Halbleiterwerk Frankfurt Oder (60V, 400mA, 40MHz).




The "datecode" KB says august 1961 but there is a catalogue that states the development was finished in 1965.




Now that´s a strange wiring...




...and the bondwire is quite close to the bottom plate.




The die is attached on a round plate which is attached to the bottom plate of the package.




The die is 2,6mm x 1,6mm.
You can spot some minor damages.




The surface of the emitter (and only the emitter) is noticeable rough. I don´t know why... 




 
10V 10mA




20mA




50mA
The current distribution is not very uniform.




100mA


More pictures here:

https://www.richis-lab.de/Bipolar29.htm