06-08-2020, 07:14 PM
Yes, the strange/unexpected behaviour is why I've explained it so extensively - I'm hoping others won't have to go through the same pain I did. Whether it's whiskers or just a heavily leaking junction is almost irrelevant - the symptoms are similar, and the AC128 was the cause. It's tempting to look at the schematic and see that the AC128 is being used well within its limits in that VAS position, so look elsewhere first, but I'm always suspicious now. Later Hacker sets used a BC159 or similar in that role - I've no idea why the AC128 was chosen, apart from familiarity, perhaps. When I've run out of AC128s, I'd have no hesitation in using a BC178 or similar - even a BC558 or BC327 if looks aren't important. No modifications necessary.
The output cap is only bootstrapping T3's collector load. If the cap was included in the global NFB loop, then that loop would go some way to mask any issues caused by that cap. Personally, if the ESR was reasonable, I'd have a go at reforming it - in my experience, slight leakage in those Philips caps usually improves with use.
Trevor, I agree that lifting the screen lead is not elegant. It only works while the whiskers are in contact with just one of the electrodes. As soon as other whiskers contact one of the other two electrodes, then you've got a short that goes via the case. So it's a short term solution at best - I prefer to think of it as a confirmation exercise. Not that I bother with that these days - if there's any sign of misbehaviour, all the AF11x get zapped before I do anything else. In the case of the TO1 types, then yes, if not mounted on a non-isolated heat sink, then they are normally OK until the second electrode meets with whiskers - which of course completes the path between that and the first electrode to meet whiskers via the case. But it depends...
As I think I've probably stressed enough, there are so many faults that go away after zapping, even if the transistor didn't seem to have much, if any, of a short to case, because the whiskers can recede from the heat of soldering, or perhaps handling. It can be really confusing or even unsatisfying to remove, test, find no significant leakage, zap anyway and reinstall, but then find the set works perfectly. However, I've stopped questioning it, and now just accept that you might never find conclusive evidence that it was a whisker, but if zapping works, that's prima facie evidence that a whisker was indeed the cause, and that's good enough for me. Of course, it depends on the circuit - so the mixer/oscillator will generally be more fussy than the final IF, and outputs are less likely to be troubled by weaker whiskers than the preceding VAS. The only times I've been led towards a whisker in an output transistor, it's been a dead short that creates quite a spark when you do the zapping - any fine whiskers are usually blown away instantly at first contact by the available current at that part of the circuit. Usually.
Mike's fault could be either (whiskers or junction fault), but we'll probably never know for sure. My previous experience is screaming "whiskers", but I'm trying to keep an open mind, so I'll give both equal odds at this stage.
Anyway, I'm glad my previous experience of troublesome AC128s in the VAS position helped to speed up the fault-finding process. You're welcome...
The output cap is only bootstrapping T3's collector load. If the cap was included in the global NFB loop, then that loop would go some way to mask any issues caused by that cap. Personally, if the ESR was reasonable, I'd have a go at reforming it - in my experience, slight leakage in those Philips caps usually improves with use.
Trevor, I agree that lifting the screen lead is not elegant. It only works while the whiskers are in contact with just one of the electrodes. As soon as other whiskers contact one of the other two electrodes, then you've got a short that goes via the case. So it's a short term solution at best - I prefer to think of it as a confirmation exercise. Not that I bother with that these days - if there's any sign of misbehaviour, all the AF11x get zapped before I do anything else. In the case of the TO1 types, then yes, if not mounted on a non-isolated heat sink, then they are normally OK until the second electrode meets with whiskers - which of course completes the path between that and the first electrode to meet whiskers via the case. But it depends...
As I think I've probably stressed enough, there are so many faults that go away after zapping, even if the transistor didn't seem to have much, if any, of a short to case, because the whiskers can recede from the heat of soldering, or perhaps handling. It can be really confusing or even unsatisfying to remove, test, find no significant leakage, zap anyway and reinstall, but then find the set works perfectly. However, I've stopped questioning it, and now just accept that you might never find conclusive evidence that it was a whisker, but if zapping works, that's prima facie evidence that a whisker was indeed the cause, and that's good enough for me. Of course, it depends on the circuit - so the mixer/oscillator will generally be more fussy than the final IF, and outputs are less likely to be troubled by weaker whiskers than the preceding VAS. The only times I've been led towards a whisker in an output transistor, it's been a dead short that creates quite a spark when you do the zapping - any fine whiskers are usually blown away instantly at first contact by the available current at that part of the circuit. Usually.
Mike's fault could be either (whiskers or junction fault), but we'll probably never know for sure. My previous experience is screaming "whiskers", but I'm trying to keep an open mind, so I'll give both equal odds at this stage.
Anyway, I'm glad my previous experience of troublesome AC128s in the VAS position helped to speed up the fault-finding process. You're welcome...
