Help with Germanium PNP transistors needed.

Status
Not open for further replies.

Babyspark

New Member
Hi, I was given a Dynatron stereo amp SRX26 with one channel not working. Likely failed components are the output transistors AL103 4J according to my research. I tried replacing those on the bad channel with 2 ‘new’ ones but no improvement. I then removed and tested all the transistors with the diode test on my multimeter. The resulting voltages are shown in the picture, hopefully this is self explanatory!
OLD 3 and OLD 4 belong to the good channel and match well.
OLD 1 and OLD 2 also seem to match though with different figures to the above.
Ditto NEW 1 and NEW 2 .
My query is, should OLD 1 and OLD 2 work as a pair. If so the problem must be elsewhere in the circuit, probably the large elec. caps.
The same query applies to NEW 1 and NEW 2.
Would be very grateful for any help
 
You are only measuring individual junctions;
but you are not measuring dynamic Hfe (as would be shown on a curve tracer) or at the very least static Hfe (at a single Vce and Ic value) to properly compare.

Germaniums are noticeably leaky, and the leakage should be measured at both room and elevated temperature.
 
Yes, quite leaky. When checking C-E, tie the base to the emitter and see if that gives you better numbers.
 
As always, randomly changing or testing components is a VERY poor way to try and repair things.

As it's a stereo unit, comparing voltages between the different channels would be FAR more useful. Absolute first thing to do is measure the voltages in the output collectors, an extremely easy job as the metal can is the collector - one in each channel should be full HT, and the other about half of that - pretty well anything wrong in the power amp will disturb the half voltage reading.

You say 'one channel not working', what exactly is it doing?.

Assuming there's no kind of noise at all?, then it 'could' just be dirty contacts in the headphone socket, as this includes a switch to mute the speakers. Stick a pair of headphones in and see if those work?.
 
Well it’s not quite random Nigel, these transistors are known to be a source of such problems and swapping them out would seem to be a reasonable first option in my view. As it happens it made no difference so I am looking further into it.
Channel not working means no sound at all in this case.
I have tried the headphone check and the same problem exists. No sound in one channel.
I have also checked the collector voltages, the good channel is approx 30/15 which is about right as far as I know, the bad one showed the same voltage on both but I can’t remember what it was and as I say I have removed them all now.
 
It's a been a long time since I repaired one of those Dynatrons (and don't recall specific problems with the outputs - but that was long ago) - 30/15 sounds about right (full supply/half supply). Obviously if the other had two readings the same, then they MUST have been 30/30 - and the top transistor can't be S/C, or things would have failed more seriously.

I would suggest refitting the original transistors, and taking voltage reading on all the transistors, it's a fairly standard type of quasi-complementary circuit, and the circuit can be downloaded (it's too large to attach here).

I always considered it a 'strange' circuit, even back when it was new, as it's a perculiar mix of new silicon and old germanium transistors.
 
If it was long enough ago then maybe your transistors were youthful enough to perform properly Nigel ;-)
Anyway as all my pairs seem to be just about matching, in a sense, when tested ex-situ, you‘re probably right about putting them back. I was just wondering if anyone was familiar enough with these components to say if my figures indicate they are fit for purpose.
I shall also look for ‘tin whiskers’ and check the big electrolytics as well as checking all transistors as suggested. Many Thanks.
 

I've never used a diode test range on a Germanium device, they were long obsolete before digital meters came out. But Germanium aren't really suitable for simple static testing, due to their high leakage.

I shall also look for ‘tin whiskers’ and check the big electrolytics as well as checking all transistors as suggested. Many Thanks.

Tin whiskers are a lead free issue, many decades after this was made.

But again, for fault finding you don't "check transistors", you fault find the circuit - measuring the voltages on all transistors would be the place to start.

The entire power amp is a kind of 'opamp', with 100% DC feedback - with the output at +30V the feedback should be turning the bottom output transistor ON to bring it down - the voltage readings should tell you why it's not.

To be fair you're lucky it's not 'self destructed' as DC coupled amps usually do, so you've got the chance to fault find it.
 
Tin whiskers are a lead free issue, many decades after this was made.
Not completely correct.

The old AF115/116/117 series of RF transistors are very prone to them.
In the last year I repaired an old "Transistor Radio" for a friend, the problem was tin whiskers in one of the IF transistors.

Not only that, I cut open the can of the transistor and now I have the pictures to prove it:


For some reason they liked to fill the little can with some goop which looks like silicone grease, I cannot help but feel that the stuff helped to promote the growth of the tin whiskers.

JimB
 

(Very) long term programmed obsolecense.
 

Easiest solution is simply to cut the screen wire off the transistor

I can't say I was ever aware what caused the problem, I just repaired hundreds of them back in the 70's - originally by fitting a new AF115/6/7 and later later on by fitting a new AF125/6/7 (when the originals were no longer available), and finally just by cutting the lead off (after the AF12x series were no longer available as well.

Possibly they used lead free solder inside the transistor, presumably for some reason? - possibly to avoid it melting if you heated it for too long while soldering it in?.
 
I have seen reference to tin whiskers on these amps on another site where the cause was said to be ‘tin’ cased components.

Also the below points to possible problems with NKT transistors used on some models
‘when I screwed the amp PCB down to chassis, one channel cut out, accompanied by a crackling sound. Checking the centre rail, I found it was only about 0.5v(should be 15!) Then the penny dropped. The SRX 26 series chassis uses germanium transistors(NKT212, 717, & 279)which are fitted with cooling clips secured to the chassis by threaded studs. Two of these measured about 30 ohms between the case of the transistor and the collector-hence virtually short-circuiting the centre rail to chassis(Positive in these amps.) The NKT 212, etc., for those not familiar with them, are similar to AC128s and the like’

My amp has NKTs, though 233D rather than the codes above. Can’t find any reference to them but could be (yet) another possibility.
 
Last edited:
We tried that first and it did not work.

Even back when we used to replace them, we always chopped the screen off first, to confirm it cured it, never saw one it where didn't?

Going purely from memory, I seem to recall the pins went collector, screen, base, emitter, with a large gap between collector and screen?.

Worst job was the Mullard radio modules (which this Dynatron uses one of - for AM IF), which used the AF11x series transistors inside - you had to remove the module, then strip it to get to the transistors.

Or maybe the case of the transistor is made from tin-plated steel. They are magnetic.

As are resistor wires
 
Nigel; The screen....
I remember that some transistors had indeed 4 terminals, but I never really understood what the fourth one was all about.

Was the screen connected to the metallic can, to act as a Faraday shield?
 
Nigel; The screen....
I remember that some transistors had indeed 4 terminals, but I never really understood what the fourth one was all about.

Was the screen connected to the metallic can, to act as a Faraday shield?

Yes, it was simply connected to the metal can - the internal short could appear to either of the other three wires, stopping it working as it grounded either base, collector, or emitter. In my experience simply cutting the wire cured the problem, although as a professional service department we replaced the faulty transistor, while we could still get them, or the later AF12x version.
 
So to fix the problem, bring a razor
 
When checking C-E, tie the base to the emitter and see if that gives you better numbers.
i'm thinking he should also do another C-E measurement with the leads reversed from the first measurement and add that to the chart.

if i was fixing this item purely for functionality, i would try using silicon devices and changing the base bias. i would make it a reversible fix however just in case somebody were interested in buying it.
 
Last edited:
I cannot help but feel that the stuff helped to promote the growth of the tin whiskers.
actually, according to the NASA study on tin whiskers, the goop neither helps or hinders the crystal growth, and the crystals can grow through soft glues (like RTV compound), and can also poke holes in some conformal coating materials (much like certain weeds can poke through asphalt walkways). they found that as little as 2% lead in solder will inhibit crystal growth.
 
Status
Not open for further replies.
Cookies are required to use this site. You must accept them to continue using the site. Learn more…