2101 Pulse gen repair report


Michael Dunn
 

Well, I might be a bit rusty, but it seems I can still repair things :-) In this case, the versatile 2101 pulse generator from 1970.

I had no idea of its condition, so first off, I wiggled all the socketed transistors, then turned it on, saw *some* sort of output, and checked all the supplies (all perfectly fine after 50 years :-) The two shaft couplings are partly broken…I’ll probably just glue them or something :-(

While playing and familiarizing myself with the gear, it suddenly died. Ugh. I thought this would be easy. The core oscillator circuit was still running though, and I finally tracked the culprit to Q182, part of the pulse forming circuitry. The transistor measures largely open now. Weird…suddenly going from working to very dead like that. I stole the corresponding transistor from the delayed pulse generator. Hopefully there is a suitable sub in my junk box.

Next, a more subtle problem. The negative pulse output had a few volts of negative baseline offset. It took me a while to grok the complex output amp circuitry (it’s still somewhat mysterious to me), but after a couple of false trails, I returned to an early hunch: the final, paralleled, output transistors. Sure enough, one of them (Q352) is bad, showing under 1 kΩ from B to C.

Yay…a mostly working unit.

Finding a TO-39 part for the output transistor might be tricky. I do have quite a few in the junk box, but whether any will have good enough AC specs…??? It does kinda hafta be that package, as the heatsink is made to fit it.

As for Q182, it’s an RF transistor—I may have a suitable sub lying around…

So…transistors. Available selection these days seems pretty sad. Am I right, or not? Especially in leaded (no surprise there). The four types of SS xstrs in the 2101 fall into an NPN/PNP-RF/Switching matrix. I sure hope a PNP switch never fails, cuz I couldn’t find a single example at Digikey, regardless of package…

I should know this, but do we have a Tek parts cross reference somewhere in the files? Quite a few parts in the 2101 have no description beyond “Tek Spec”…


teamlarryohio
 

Hi Michael -- There's a pretty full set of Common Design Parts Catalogs at https://w140.com/tekwiki/wiki/Common_Design_Parts_Catalogs -- should do you just fine.
-ls-


Tom Lee
 

For the output amp, the NPNs are 2N3053 and the PNPs are 2N4036, replaceable here by the venerable 2N2219A and 2N2905, respectively. Those parts are still available cheaply.

Q182 is a 2N3563, replaceable by several devices in the 2N918/PN918/2N2857 family.

Be thankful that it's a 25MHz gen, rather than, say, a 100MHz one. You should be able to find replacement transistors to keep it going pretty much forever.

Cheers,
Tom


Sent from an iThing, so please forgive the typos and brevity

On Feb 3, 2021, at 11:16, "Michael Dunn" <md@cantares.on.ca> wrote:

Well, I might be a bit rusty, but it seems I can still repair things :-) In this case, the versatile 2101 pulse generator from 1970.

I had no idea of its condition, so first off, I wiggled all the socketed transistors, then turned it on, saw *some* sort of output, and checked all the supplies (all perfectly fine after 50 years :-) The two shaft couplings are partly broken…I’ll probably just glue them or something :-(

While playing and familiarizing myself with the gear, it suddenly died. Ugh. I thought this would be easy. The core oscillator circuit was still running though, and I finally tracked the culprit to Q182, part of the pulse forming circuitry. The transistor measures largely open now. Weird…suddenly going from working to very dead like that. I stole the corresponding transistor from the delayed pulse generator. Hopefully there is a suitable sub in my junk box.

Next, a more subtle problem. The negative pulse output had a few volts of negative baseline offset. It took me a while to grok the complex output amp circuitry (it’s still somewhat mysterious to me), but after a couple of false trails, I returned to an early hunch: the final, paralleled, output transistors. Sure enough, one of them (Q352) is bad, showing under 1 kΩ from B to C.

Yay…a mostly working unit.

Finding a TO-39 part for the output transistor might be tricky. I do have quite a few in the junk box, but whether any will have good enough AC specs…??? It does kinda hafta be that package, as the heatsink is made to fit it.

As for Q182, it’s an RF transistor—I may have a suitable sub lying around…

So…transistors. Available selection these days seems pretty sad. Am I right, or not? Especially in leaded (no surprise there). The four types of SS xstrs in the 2101 fall into an NPN/PNP-RF/Switching matrix. I sure hope a PNP switch never fails, cuz I couldn’t find a single example at Digikey, regardless of package…

I should know this, but do we have a Tek parts cross reference somewhere in the files? Quite a few parts in the 2101 have no description beyond “Tek Spec”…












Michael Dunn
 
Edited

Thanks Tom, tho, a few notes:

On Fri, Apr 16, 2021 at 05:17 AM, Tom Lee wrote:
For the output amp, the NPNs are 2N3053 and the PNPs are 2N4036, replaceable
here by the venerable 2N2219A and 2N2905, respectively. Those parts are still
available cheaply.
Those parts are for baseline offset, not output amps per se. No AC requirements, so, yes, fairly easily subbed.

To "fix" the output amp, I just removed the offending transistor of the paralleled pair. The remaining part will be fine, as long as I don't push the output too hard.


Q182 is a 2N3563, replaceable by several devices in the 2N918/PN918/2N2857
family.
Which seem to be obsolete, or close enough to :-( At least here, a modern/available SMD could be kluged in if need be (even so, finding one with suitable DC spex may be tricky). I did find a suitable old part in my junk box (2N5179), though its DC specs may be marginal. I hope it lasts.


Be thankful that it's a 25MHz gen, rather than, say, a 100MHz one. You should
be able to find replacement transistors to keep it going pretty much forever.


Tom Lee
 

Hi Michael,

I expect the same transistors to be good subs for the actual output amp, too, so if the thing blows again, I'd give those a try. I don't know what Tek sorted for, but a (very) crude proxy for ft is 1/BVCBO. Lower breakdown voltage is correlated with higher ft. So if you find the risetime to be marginal, that's one sort you can perform to identify the most promising candidates (assuming you have, or have access to, a curve tracer).

And the 2N918 family includes a great many different part numbers (the 2N918 is sort of the 2N3904 of ~1GHz transistors). Even though through-hole parts are obsolete (or at least obsolescent), I've never had difficulty finding hobbyist-level quantities of devices in that family. Most recently, I bought a handful of 2N5770s from our own Brad Thompson. Those transistors are part of the 918 family. That part, as well as the PNP complement (2N5771), are good to keep in the junk box.

Cheers,
Tom



Sent from an iThing, so please forgive the typos and brevity

On Apr 16, 2021, at 4:39, "Michael Dunn" <md@cantares.on.ca> wrote:

Thanks Tom, tho, a few notes:

On Fri, Apr 16, 2021 at 05:17 AM, Tom Lee wrote:
For the output amp, the NPNs are 2N3053 and the PNPs are 2N4036, replaceable
here by the venerable 2N2219A and 2N2905, respectively. Those parts are still
available cheaply.
Those parts are for baseline offset, not output amps per se. No AC requirements, so, yes, fairly easily subbed.

To "fix" the output amp, I just removed the offending transistor of the paralleled pair. The remaining part will be fine, as long as I don't push the output too hard.


Q182 is a 2N3563, replaceable by several devices in the 2N918/PN918/2N2857
family.
Which seem to be obsolete, or close enough to :-( At least here, a modern/available SMD could be kluged in if need be. I did find a suitable old part in my junk box (2N5179), though its DC specs may be marginal. I hope it lasts.


Be thankful that it's a 25MHz gen, rather than, say, a 100MHz one. You should
be able to find replacement transistors to keep it going pretty much forever.