PL 400 Rebuild

[Gepetto]

Correct, I'm just being chicken. And re-checking my old notes, I apologize, the new caps did raise my B+ 5 volts! I see 82 VDC with wall @120vac. I must have bumped the dial when taking first after cap readings. Lots going on here with sick black lab taking my time...

Mike

You are getting proper and normal bulk voltages Mike. 80-82 at 120VAC is the nominal you should see.

 

[Magic Smoke]

You are getting proper and normal bulk voltages Mike. 80-82 at 120VAC is the nominal you should see.

Yes sir,

I had mentioned in an earlier post, incorrectly, so apologise to you too.

BTW, I'm not sure I'm doing this right so pulled both rail fuses to outputs to measure for the 22 volts and others on the control board. I checked and pulled left pot and compared both bias transistor with meter so will get familiar with layout when time permits. The left channel was still randomly pegging meter gently and different time period with output fuse(s) disabled. I will be replacing board so practising how to troubleshoot this thing. I removed some humongous solder blobs that were very close to other..blobs. On the component side I found one blob of solder stuck to resistor lead, but held for 30 years or so. Might have been Sue A. first day on job, lol.

cheers

 

[Magic Smoke]

I have the "A" version of the control board. This thread, http://forums.phxaudiotape.com/showthread.php/4784-Phase-Linear-400-issue-help!/page21, had some interesting things to check concerning my left channel bias problem. Are the boards close enough for trouble shooting?

Thx, Mike

 

[Magic Smoke]

Q1, Q2 are probably getting a little wonky. 34 mv of offset is a good clue. Replace the old electrolytics on that board. They are a ticking timebomb.

Passing time till I get safety bonus, I'm pulling caps and transistors from left channel control board to read, test. Learn to set up to read gain on some of the transistors.... and curious to find why I can't adjust bias on left side.

Halfway thru I might have found problem, on a simple straight wire. The gray wire end that connects to bottom driver, right half, was burned and hanging by a few strands. I'll continue to test the rest, clean up board and look harder at things.

 

[Magic Smoke]

Got the no bias adjustment on left channel fixed. D3L was open with the burned grey, left ch. out, on backplane. However, now has DC offset drifts.
If Joe's board were not in then I would shotgun the caps and as many transistor I could find for both channels.

Q1 & Q2 looked ok but bigger gain differences with Q3 and Q4 using my $17 Arduino tester, see imbedded pics. I was not sure if these needed to be matched..

mike

 

[laatsch55]

There is a very easy trick on matching Q1,Q2. Measure basic resistance between Base-Emitter and Base-Collector of Q1, Q2. The closer the BE jct's and BC jct's between Q1,Q2 the lower the offset will be. You need a few extra to come up with a match , but it's worth it. Fred Longworth on AK called this a first order Beta match and it really works. No curve tracer , no gain reading , no problem. I matched TIS-97's like that fopr years and I coulkd consistently get offset to less than 5MV with a resistance match of 2 points or less.

For example...I'm testing two 97's right now...


Transistor #1---BE--729, BC---727
Transistor #2---BE--728, BC---726

The closer the numbers are between the BE, BC jctns in EACH transistor the more gain it will have. The closer the BE, BE and the BC,BC jctns are to each other, the lower the offset will be.

Those 700 numbers represent ohms in the continuity test of my Fluke 179. In standard resistance mode the numbers are in the 500K ohm range. I'd have to read up on my meter about what the difference in the two modes are.

With a difference of only one ohm BETWEEN the BE, BC jctns that could be a difference in the temp of the tranny's too. So once I had screened a dozen or so to be that close , I would lay them out together on a surface that could hold all of them and let them sit for ten minutes and test them again. DO NOT HOLD THEM WITH YOUR HAND. After the ten minute period I would pick them up by the lead with the mini-grabbers only. Body heat transfer will skew the reading....

 

[laatsch55]

I could usually get an exact match between BE,BE and BC, BC jctns with a dozen or so transistors to start with...

 

[Magic Smoke]

There is a very easy trick on matching Q1,Q2. Measure basic resistance between Base-Emitter and Base-Collector of Q1, Q2. The closer the BE jct's and BC jct's between Q1,Q2 the lower the offset will be. You need a few extra to come up with a match , but it's worth it. Fred Longworth on AK called this a first order Beta match and it really works. No curve tracer , no gain reading , no problem. I matched TIS-97's like that fopr years and I coulkd consistently get offset to less than 5MV with a resistance match of 2 points or less.

snip

Thanks for that! If I knew what I was doing I'd be a far less dangerous guy. I'm coming back to this board after Joe's board up and running and get spares for the old board in order.

cheers

 

[laatsch55]

Very Welcome. And after 10 minutes both tranny's were 730--BE, and 728--BC....a very simple but effective matching procedure for the input pair. As far as matching outputs, I would stick with Joe's procedure in trhe WOPL archives....

 

[Northwinds]

Very good info for all builders there Lee!!! It's the minute things that can make all the difference in the end of just how good these will sound when done. Takes a lot of care and pride to match things up like that. Most folks probably just put shit in and don't test for close tolerances

 

[laatsch55]

Thanks Ron, it can and does make a difference . Of course with the "E" boards this is a moot point now....With the OPA 2134's replacing the differential input pairs offsets are now.000 to .001, really hard to achieve with discreet components...