On my right channel issue, I've noticed that the right set of EL34's had a bit of 'ringing' noise to them. So I measured the offset which was about -67V (with no headphones plugged in) vs about +2V offset on the left. It wasn't fully warmed up when I measured this though. Balance was close to perfect (0.1V) if that says anything. I tried swapping left & right sets of EL34s, and using 6DJ8s instead of JoaMats solid state input. No change though. Could the sand responsible for handling the offset just be bad? Or something else?

I had an early problem of this sort that turned out to be a J79 had blown. That was possibly the last gasp of collateral damage from fake C3675's. But that particular J79 took about a year to go pop.

Was it Q24 or Q25 in your case that had blown? Or did you replace both of them?

Can't remember which one. I probably reported it somewhere in the bowels of this thread. But I definitely only replaced the blown one. Hasn't missed a beat since (around six years).

Did some searching of this thread, doesn't seem like you mentioned which J79 it was. Any way to tell which J79 might be bad without pulling them from the board?

You might try the diode test method described by Kerry in this post. If lucky you get a hint which component that is the faulty one.

Edited May 15, 20179 yr by JoaMat

I've replaced the J79 closest to the front panel now. Seems to have done the trick, offset is now normal on the right channel. No extra noise so far either.

Thanks for the help Craig & JoaMat.

Edited May 15, 20179 yr by s_r

Result!

Congrats!

Sometimes taking the bulldozer approach by change extra parts is the easiest path    I've done this when I've been stumped as well.

Really glad you've got everything sorted out.

@s_r: When you say no extra noise, do you mean that your T2 finally has a noise floor you're satisfied with?

With the solid state input, yes. Both tube sets I have, even the 'low noise' philips, are intolerable to me.

Guess I spoke too soon. The exact same behavior is happening again (-70V offset, distortion in the right). Something else causing the J79 to have issues?

try using the diode tester on your DMM to individually test/compare all of the J79 with power off

Edited May 18, 20179 yr by justin

Tried that just now, and the values between all four seem to be within the margin of error. The two on the right channel measure pretty similarly, while the left two differ moreso, strangely enough.

Either a battery, or one of the many current sources?

 

Stax block diagram.docx

The batteries measured fine last I checked, the K216s beside the J79s all measure similarly to each other. Replace the LF353 maybe? No change after swapping out the LF353 either.

Edited May 18, 20179 yr by s_r

One channel isn't oscillating at some high frequency?

Haven't checked that yet. Where's the best place to measure that? Could try the +200V output of the battery, or the final output, like the measurements I did some months back.

Edited May 18, 20179 yr by s_r

If it is oscillating you'll find it everywhere. You'll need an oscilloscope to find out though.

a little late to the party but i have what i think are some real 2SC3381-BL on the way. "Like" this post if you think you'd be interested in some. i'll post them up for sale somewhere once i test them

I'm up for some 2SC33810-BL, Justin!  Hook me up, pleaseee... 

I am too!

I have a few but don't know if it is authentic 

 

Edited May 19, 20179 yr by Whitigir

So after setting up my scope to see if there was any oscillation, the distortion and high offset disappeared. Afterwards the distortion would show up after a few hours of it being on, and it's sometimes cured by just unplugging/replugging the headphones. Feels like I'm chasing after Schrodinger's distortion

Whatever the case, I've put heatsinks on both the 3840s and LSK389s now.

7 hours ago, Whitigir said:

I am too!

I have a few but don't know if it is authentic 

 

has anyone ever found a fake part in that package? seems like it wouldnt be worth the effort to fake 

Those look real. That said, I am aware of fake 2sk389, but then they used relabelled 2sc3381, so pretty doubtful they would fake a 2sc3381 when they would more likely label them as 2sk389... I did see some questionable 2sc3381 with flat legs, but the ones pictured are squared, which matches the spec sheet.