2SA798 Double PNPs - The Real Scoop

[c.coyle]

2SA798 double PNPs. The threads on here are kind of vague about their reliability. Many say they replace them, but I haven't found much on actual failures.

I'm asking because I picked up a receiver with these things for differential pairs. DC offset is 3mv/6mv and all is well. If I was going to keep it for myself, I would leave them in. But, it's going to be a gift, so I don't want it coming back in 6 months.

Keep 'em or replace them with back-to-back KSA992s or the like?

 

[chazix]

I can offer some anecdotes from personal experience, for what they're worth.

The most annoying anecdote came from a situation much like yours. There was no sign whatsoever of problems with the 2SA798s in a power amp I was working on for a chum, as long as it was on my workbench. And that was a pretty long time, and included days of final test running. As soon as the owner put the amp back into service, it started making snap/crackle/pop noises. Since my friend lives a thousand or so miles away, I had to ask a surrogate to do the rework. (Thank goodness for an older, wiser brother who's also a friend of the amp's owner!) At least there was very little doubt in my mind about what was wrong.

In my experience, the usual symptom of faulty 2SA798s is noise, at least initially.

Once, I put a known noisy part on a curve tracer. When run up to collector voltages that should have been no sweat for the part, it stopped making curves that looked anything like a BJT. So, curve tracing might be a useful go/no-go test.

Cheers,

chazix

 

[restorer-john]

I second the use of a curve tracer, but the instrument is a device that is a little obscure for the majority of hobbyists and they have a learning curve as well as a significant investment. I would think the best option for the majority of people would be to buy a bag of 100 suitable transistors and match pairs of them with a cheap transistor tester, put some heatsink compound on the flat faces and zip tie them together.

I had a 2SA979 (similar) the other day giving me grief on an SX-1980. I still have some original NOS, but they can range from perfect to noisy too. IME, black lead disease is the biggest killer of them.

 

[JURB]

That noise is usually intermittent so the curve tracer might tell you nothing.

 

[sregor]

I'll throw in my 2 cents. Input pairs, in general, seem to degenerate when there are other issues in the amplifier which put rail voltages on the output. So If DC is low and unit hasn't had any service problems, then the input pair are usually fine. However if output problems, I would nearly always change the input pair while I'm at it. I don't recall seeing just a bad input pair with no other problems, but never did kind of study on the problem.

 

[restorer-john]

That noise is usually intermittent so the curve tracer might tell you nothing.

The noise is often related to transistors that have deteriorated. Often that manifests as a much lower sweep voltage than rated on my curve tracer to cause breakdown or AWOL curves.

I have a bag full of T092s that were intermittently noisy (Akai GX-xxx RTR) and only the curve tracer showed the problem. A normal Tr tester doesn't help.