Author Topic: Help identifying a unijunction transistor  (Read 1301 times)

0 Members and 1 Guest are viewing this topic.

Offline Michael-VK2BEATopic starter

  • Contributor
  • Posts: 48
  • Country: us
Help identifying a unijunction transistor
« on: August 25, 2023, 12:37:03 am »
I have a power supply (HP 0950-2261 / Computer products XL130-3604E) from an HP logic analyzer that I'm debugging. (It works (in a fashion) at 110V when configured for 220V, but not when strapped for 110V)
I have a reversed engineered circuit of a similar Boschert supply that identifies this as a 2n2647. I can only find datasheets listing this as being in a TO18 or TO92 package, not the one shown in the picture.

Can anyone recognize this device or confirm that it is indeed a 2n2647.
(BTW, I'm pretty sure this is not the problem)
 

Offline Tomorokoshi

  • Super Contributor
  • ***
  • Posts: 1212
  • Country: us
 

Offline floobydust

  • Super Contributor
  • ***
  • Posts: 7756
  • Country: ca
Re: Help identifying a unijunction transistor
« Reply #2 on: August 25, 2023, 03:39:33 am »
I have 2N2646 and they are in a small transistor package, as 1855-0010 (GE).
The pic shows a custom module 190042 Rev. 2A 9810 pins 1, E, 2 and a strange hole in the PC board at the middle pin.
I had guessed perhaps it's a clipped HP part number like 1900-0042 or 1900-4200 but results did not look promising. 1900-0042 (TRW) A5X1046 is another rare beast, likely a JFET.
 

Offline Kim Christensen

  • Super Contributor
  • ***
  • Posts: 1819
  • Country: ca
Re: Help identifying a unijunction transistor
« Reply #3 on: August 25, 2023, 04:38:53 am »
I wonder if they've simply replaced the old obsolete unijunction transistor with a module when building the later models. Does your PCB look like this one? Arrow pointing at where I think your module would be:
 
The following users thanked this post: fzabkar

Offline Michael-VK2BEATopic starter

  • Contributor
  • Posts: 48
  • Country: us
Re: Help identifying a unijunction transistor
« Reply #4 on: August 25, 2023, 03:21:46 pm »
I wonder if they've simply replaced the old obsolete unijunction transistor with a module when building the later models. Does your PCB look like this one? Arrow pointing at where I think your module would be:

Yes, that's it.
 

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 2453
  • Country: mx
Re: Help identifying a unijunction transistor
« Reply #5 on: August 25, 2023, 03:26:48 pm »
Very likely it is a PUT, with included resistors in the assembly to provide an identical stand-off ratio of the original, and very likely obsolete, unijunction transistor.
 

Offline Kim Christensen

  • Super Contributor
  • ***
  • Posts: 1819
  • Country: ca
Re: Help identifying a unijunction transistor
« Reply #6 on: August 25, 2023, 05:11:16 pm »
Yes, that's it.

I think if the module is behaving like a unijunction transistor, then you can call it good.
 

Offline Michael-VK2BEATopic starter

  • Contributor
  • Posts: 48
  • Country: us
Re: Help identifying a unijunction transistor
« Reply #7 on: August 25, 2023, 11:06:13 pm »
Agreed.

The problem is that when strapped for 110 V I get only about 2V on the 5V output and nothing on the other outputs. (no load).
In this configuration (110V), the input rectifier and capacitors are acting as a voltage doubler (so the DC voltage on the primary is ~300V).
When configured for 220V (just a bridge rectifier), the DC on the primary is only ~150V but I get the correct output voltages (however it cannot take the rated load ... for obvious reasons)
This is how the unit was configured when it came from eBay (wrong voltage setting) 8-(
As received, the Logic analyzer was working but the oscilloscope board was not. Upon opening I saw that the cable to the oscilloscope board was missing. When reconnected, the additional load was enough to cause the PS to come out of regulation (dropping the 5V to 3.5V).
The behavior is a little odd because the secondary must be fine and if it's something like the over voltage devices breaking down then I would expect the input fuse would blow. I replaced the two big input capacitors with no improvement (and all the capacitors I replaced, I checked and found them fine). I just got an isolation transformer so I can start doing some debugging on the primary.
 

Offline Kim Christensen

  • Super Contributor
  • ***
  • Posts: 1819
  • Country: ca
Re: Help identifying a unijunction transistor
« Reply #8 on: August 25, 2023, 11:57:11 pm »
Not sure if you found this document in your search, so I'll include it here:
https://www.radiocollection.be/images/restaurations_img/HP54503/supplyHP54503.pdf
 

Offline Michael-VK2BEATopic starter

  • Contributor
  • Posts: 48
  • Country: us
Re: Help identifying a unijunction transistor
« Reply #9 on: August 26, 2023, 12:07:19 am »
Not sure if you found this document in your search, so I'll include it here:
https://www.radiocollection.be/images/restaurations_img/HP54503/supplyHP54503.pdf
Yes, that is where I saw that the UJT is marked as 2n2617.
There is an excellent Boschert maintenance manual on Bitsavers ... "Single Stage (Flyback-Discontinuous) Power Supply" that describes how the topology works and how to test it. It is not for this exact unit but its very helpful none the less.
 

Offline Michael-VK2BEATopic starter

  • Contributor
  • Posts: 48
  • Country: us
Re: Help identifying a unijunction transistor
« Reply #10 on: August 30, 2023, 08:52:21 pm »
So, it seems that I was not providing the minimum load when testing.
I had a 200mA load on the 5V output but upon reading the HP1660 service guide I see that this should have been around 2A.
When loaded at 2A the supply runs correctly when provided 110V and strapped for 110V (i.e. voltage doubler enabled). I was led down the garden path because a smaller load worked fine when the input voltage low.

While investing this issue I reversed engineered the over voltage protection & reference board which, for this supply (HP 0950-2261 / Computer products XL130-3604E), which differs somewhat from the supply in the HP 54503A scope that is referred to in this thread.

I'll attach it here just in case its useful to someone else fixing the same supply.

Michael
 
The following users thanked this post: Tomorokoshi, floobydust, fzabkar


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf