Marshall tsl122 Bias problem

[sebagitara]

Today I measured bias voltage with no power tubes inserted again and blowed hot air from the blow drier. When I started the amp bias voltage reading from pin 5 of every power tube was -41V and after I blowed some hot air inside the amp I took the reading on pin 5 again and it was -36V so i guess. I think that the difference is too big and therefore i think zhe cause of the problem is somewhere in the bias circuit also i measured resistance between pin5 and the one next to him, i think it's the screen grid pin and it was -1.5 mega ohm, dont know where that minus came from but i sure think that the resistance between the pins that are normally not connected to each other over anything should have much bigger resitance. So should i drill the board and unatach pin 5 from the board to solve this problem or what? I'm stuck here.

[Jazz P Bass]

Hey, it's your call.
If it works, that is a lot cheaper than a new board.

[sebagitara]

Today i finished with the repair from that site you guys told me, okey the bias seems pretty much steadier then before but now some strange thing started to happen, when i set the volume on e.i. 5 and play the guitar wile measuring the bias, while i'm playing bias jumps all over the place, and when i strum the guitar a little bit harder the bias goes above 120mV for a half of second or something like that, when i stop playing it's settles on the starting value but this things never happened before I really don't know what can be the source of this problem, help guys please i'm stuck here and as i play it further it's starts to rise again even with no signal present

[km6xz]

The bias measurement is the voltage drop across the out devices which are partially turned on with no signal present. That idle current with no signal is what were refer to when saying measure and adjust bias. With a signal, the current pulled will vary directly with the signal amplitude so high current on peaks of output volume is normal and a good sign. After it gets hot, the idle current can increase a little, how much is it increasing with no signal?

[sebagitara]

Ok i'll measure it again through one work hour and i'll post my measurements then, thank you all for replies

[sebagitara]

Okay I did the measurements again and here's what I got :

starting point - both sides 70mV

first measurement after half hour : left (closer to output transformer) - 75mV, right (closer to power transformer) - 94mV

second measurement after one hour : left (closer to output transformer) -77,7mV , right (closer to power transformer) - 97,5mV


and after few minutes i decreased bias to 70mV again and now it stays at that value

[sebagitara]

Well this is very odd. I started the amp and waited 5 minutes for bias to settle down, right side was 57mV and left side 66.2mV and after a while right side got up to 70mV left also to 70mV and then it stopped and remained at 70mV, it increased maybe 0.5 to 1 mV more but that is not so bad, and then i turned the amp off, waited for it to cool down and started it again but this time with bias on the right side set to 70mV and it rised to some value, i don't rember exactly what value, and remained there. I dont know what can cause this things to happen but i suspect on the tubes now. What do you guys think?

[Jazz P Bass]

I would 'set' the bias current, with the amp hot, at 70mv per side.
The bias current of a push/pull output stage is not a critical setting.
There is no point in fussing over it.
The idea is to remove crossover distortion.
At 70mv per side you will most certainly not have any of that.
So if the tubes are not red plating (too hot of a bias) then button it up & play it.

[sebagitara]

Ok I would to that but if i set bias to 70mV it will rise to 90 or 95 mV after hour or so and then it would stabilise itself, through the first hour of work it is constantly rising no matter what voltage i set.

[Jazz P Bass]

What I meant was for you to set it at 70mv's when it is finished rising.

[sebagitara]

Hi all, I just got back from a rehearsal with my band and I did the rehearsal with Marshall. The sound was f****ckin awesome, way much better than the old one with 220k resistors but I the bias thing is still happening it was constantly rising while we played. Right side was going nuts and rising like hell and the left side was pretty much stable. I did every mod on that bias drift repair page and this shit is still happening, though i don't have enough money now to buy new tubes, these ones are pretty much old so I suspect that they could cause the problem, but I must wait and collect some more money to buy a new set of tubes. They're so goddamn expensive here, factory matched quad costs about 150$, and the set I have now is : quad JJ EL34 but these ones are not factory matched, well they are not matched at all to be hones . Please help guys what's the chance that tubes are causing this problem?

[Jazz P Bass]

If you swap the tubes from left side to right side & the problem socket is still misbehaving, then it's not the tubes.
If the problem remains in that socket, then you really do need the newer board.

[sebagitara]

Well before I decide to buy a new board I'm going to do oke more thing. I'm going to take tsl schematic and make a new bias circuit identic to the one on the board but it will be on the separate board i think it can be done because bias circuit itself isn't complicated and i think this could solve the problem without buying a new board from Marshall.

[Alex R]

the problem seems to be caused by HT voltage leaking across to pin 5 and drawing down the bias voltage. It happens quite near to the pin. You would need to take a router or a dremel to it and take all the pin 5 connections and traces right off the board. Even then you still have an electrically porous board. I tried all sorts of things to solve this very common problem with TSL/DSL main boards and I have found a replacement board is the only solution. And it works perfectly; the new Marshall boards are made of better material and I have never had one back once repaired.

I too am a Marshall authorised tech, in the UK, and I promise you that bias drift due to heat on early TSL and DSL models is real and quite a common fault.

[Alex R]

A couple more things about this problem.

It is one of those rare 'standard faults' that you can pretty reliably diagnose from the description of symptoms over the phone etc. My customers tend to assume that any fault will be a 'standard fault' - actually most faults are unique and have to be individually traced, but this one is very widespread with early pale green TSL/DSL main boards.

You can test a board for this fault. Take all the power valves out and connect your voltmeter to one of the pin 5 connections - I usually solder a bit of bare wire to it so the clip doesn't slip. Other lead to chassis of course. Then fire it up and blow a hairdryer at the area of the power valves. If it has the fault you will see the bias voltage go down, not just a volt or two but like 10 volts and more as the board heats up. I guess if you were hacking away at all the pin 5 connections you could use this to see whether you'd cured it. Do remember that all the traces leading to pin 5 connections are suspect, the bias voltage is present on a lot of copper and it's all vulnerable because the whole board is likely to be electrically porous when it gets hot, at least the area that gets hot around the power valves.