Marshall JTM30 bias mod problem & IC3 (LM348) problem

[g1]

You need a solution, but we are going by your recollection as there is no documentation of your previous mod.
I'm aboard with what you think you did before except for you leaving R132 alone. Your new circuit in parallel with R132 immediately drops the bias voltage and makes the tubes run hotter, near as I can tell.
But what you are trying to accomplish is to make them run cooler than the stock "class A", correct?
I'm going to guess that when you did the mod before you removed R132 and suggest you try it.

[Tone Meister]

Yep, we definitely want them to run cooler than the stock design runs them. I'm certain all I changed was swapping the 3 zeners for 3 resistors and I thought I had picture documentation to demonstrate, but thus far I haven't found it. I am equally certain I did not remove R132 either of the previous two times, and each time the bias adjustment was within the desired range (55-65% dissipation). Hope I'm not coming across as obstinate on this, but I'm sure of what I did. However, the one difference is the type of pot used.

I think I have another of the 50K pots here that I used, not that it would make any difference. Maybe Juan is right, maybe that Bourns pot is bad. Perhaps it tests fine under no load but craps out when voltage is applied, albeit that is a minimal load.

I still believe that I can't see the forest for the trees. It will end up being something so simple that I am continuing to overlook it, thinking I have it figured out. That's the inexperience showing.

[g1]

Ok, if that's how you recall doing it, but it's not logical that adding an extra load in parallel with the R132 could increase the negative voltage. It should do the opposite, which would make the tubes run even hotter. Unless those zeners load down the circuit a lot more than I think they do.
A simple test is to disconnect one end of your series network. Now what is the DC at C126 ?
This should show if there is some other fault in the bias circuit.
The only other thing I can think of, if you haven't flipped the board over, maybe the other persons bias mod included some trace cutting or jumpers. If you haven't yet, have a look.

[Tone Meister]

I had to pull the board to do complete testing, and there are no traces cut, in fact the board and work was very neatly done - it just doesn't work. With all the components removed, I examined every trace and every soldering pad under magnification and tested each junction for continuity, then reinstalled the two resistors and the pot.

Came across this picture from one of my earlier JTM30 bias mods which shows a common Piher 50K PT6KH503A2020 where the Bourns is fitted now in this amp. The bias works as I'm describing with this arrangement. You can see where the one leg is lifted and not used.



I'll lift one end of the string as you've suggested and measure. I may go ahead and try this pot to see what difference it makes.

[Tone Meister]

Ok, if that's how you recall doing it, but it's not logical that adding an extra load in parallel with the R132 could increase the negative voltage. It should do the opposite, which would make the tubes run even hotter. Unless those zeners load down the circuit a lot more than I think they do.

Yeah, well, that's exactly what's happening too. The tubes are biased hot enough to redplate.

A simple test is to disconnect one end of your series network. Now what is the DC at C126 ?
This should show if there is some other fault in the bias circuit.

OK. First I'm gonna measure R132 with pot set to min and again with pot set to max as you suggested before. Then I'll disconnect one end of the string and check DC at C126. I should have time to get back on this in the morning. Thanks ya'll.

[J M Fahey]

I still believe that I can't see the forest for the trees. It will end up being something so simple that I am continuing to overlook it, thinking I have it figured out. That's the inexperience showing.

We *all* have the WTF!!!! moments.
One extra possibility: maybe you cracked open some track while modding, in that case pot adjustment is irrelevant.
Measure end to end of the resistor string and move pot ... does resistance change?
Yes, even leaving R132 there.
I'm not asking for perfect adjustment at this point, just that there is resistance variation or not.

[Tone Meister]

We *all* have the WTF!!!! moments.
One extra possibility: maybe you cracked open some track while modding, in that case pot adjustment is irrelevant.
Measure end to end of the resistor string and move pot ... does resistance change?
Yes, even leaving R132 there.
I'm not asking for perfect adjustment at this point, just that there is resistance variation or not.

I will perform that check also in the morning when I get back to the amp. As I said a couple of posts ago, I pulled the board and checked all the traces and soldering pads under magnification and also checked each junction for continuity. It's definitely something that damn simple and I just keep missing it.

[Tone Meister]

Was able to get back to the JTM30 last night.

Ended up giving up on the bias pot, but here are answers to the questions by G1 and Juan. All these measurements taken with the bias mod in place:

NO TUBES

• R132 52.8 Kohms (pot @ min) and 61.2 Kohms (pot @ max)
  87 VAC @ cathode of D107
  C126 -20.6 DC (R string connected)
  Disconnect one end of R string C126 -7.1 DC
  Measure end to end of R string for resistance and adjust bias pot - 56.2 Kohms (pot @ max) and 63.8 Kohms (pot @ min)

I ended up removing the resistor string and replacing it with 3 zener diodes. Same problem, and I'll outline the details in the next post.

[Tone Meister]

Abandoned the bias mod and replaced the zener string with slightly larger values over what was stock. The problem is pretty much the same. The power tubes aren't red plating as before but they're likely on the edge. Here's what I did.

Swapped out D107
Removed R string and installed 52v worth of zener diodes (20v > 12v > 20v) stock values were (15 > 9 > 15)

TUBES IN
Cathode of D107 measures 78.5vac
Anode of D107 measures -30.8vdc
Bias voltage at R127, R129 measures -30.1v
Bias voltage at pin 5 of power tube sockets measures -30.1v

Voltages along the zener string:
(0.00) Z101 (-.1) Z102 (-2.6) Z103 (-30.7)

When I soldered in the zeners, I made sure that each component connected to the next by overlapping the leads and soldering them together.

Clearly I'm in over my head here, so I hope these numbers can help one of you pinpoint the problem area. What do you see?

[Tone Meister]

The amp does work, but output through the power amp is very low. The preamp is working fine, sent a signal from the FX LOOP SEND to another amp and no issues in the preamp. However, when putting a signal into the FX LOOP RETURN the output is very weak.

[Jazz P Bass]

I would measure the Volts ac signal at the output tube grids with the amp maxed.
Use a 100mv/ 1K signal in.

If you see a 30Vac RMS signal & the tubes are known to be good, you may have a bad output transformer.

If there is a feedback resistor, I would lift it for the test.

[g1]

R132 52.8 ohms (pot @ min) and 61.2 ohms (pot @ max)

If you don't mean K ohms here, this is a real problem.

[Tone Meister]

If you don't mean K ohms here, this is a real problem.

Good catch, earlier post edited to reflect those readings are in Kohms.

[g1]

Your bias mod will work if you will only remove R132.
Consider the following possibilities:
1) we all like to think so, but none of us have 100% accurate recall. Maybe the previous times the mod worked you removed or just clipped it. Electronically, your bias mod can never reduce idle current if R132 is left in circuit.
2) In the previous time the mod worked, R132 was incidentally out of circuit, maybe due to a faulty component, bad connection, or cracked trace or something.

[Tone Meister]

The bias voltage still ain't in the ballpark with 3 zeners. I swear, this will end up being as simple as "is it plugged in?" but I am still missing the problem.