Biasing Marshall VBA. Varying voltages across screen grid

[Bassman Rich]

Help!

I had an output tube blow on my Marshall VBA the other week (It was blowing the ht fuse all the time). So feeling a bit thrifty, I ordered 2 svetlana winged c's, as thats what most seem to advise is the best for this amp, and I figured probably only one, or at the worst 2 valves would be blown.

One valve had no 'glow' from the heater, so seemed the obvious one to replace, so I did the opposite one of the 2 banks of 4, just for symmetry as much as anything.

Now I can get the biassing done well within spec according to the back of the amp instructions, but had some additional info from a very helpful technical guy Dave at Marshall, who suggested that if the voltages across r4 and r11 was above 12mv, it could indicate a further power valve is on its way out, R11 was 14mv, R4 20mv.

Now I have measured the voltage across the 5 w screen grid resistors, and they are as follows:-

Viewing from underneath, top left valve is v5, V5 and V9 are svetlanas, the other 6 are tung sols

v5, 725mv v9, 740mv
v6, 87mv v10, 410mv
v7, 680mv v11, 430mv
v8, 350mv v12, 410mv

The acid test, I played the bass, and it sounded great, and the oscilloscope output waveform of the 1khz sinewave input was nice and clear, no hump or crossover distortion on the sine wave. If I hadn't had the additional advice, I would be perfectly happy, but now i'm not sure what to do. Help!!!

[3below]

If Dave at Marshall is helpful, ask him, see what he says? Looking at the numbers V6 seems somewhat low compared to the others, try swapping it with V10, see if same pattern emerges. Then start looking at schematics and test values.

[MarkBassChat]

If you are in contact with technical guy from Marshall, it's the best source of knowledge. Ask him about it.
Ideally, you should replace all valves because the one that are still good have different parameters than the new one. And you should use a matched set of valves. If you consider that this is too expensive, other possibility is to sent your good valves to someone that has a valve-meter and ask him to select additional two valves with similar parameters. I personally never heard of matching valves by measuring grid current. You need to know what is anode-cathode current of the valve. This can be measured by breaking a wire between cathode and the ground and puting a DVM in this place (and measuring the current). This may be not that easy but in this way you may match valves on your own.
BTW, 87mV on V6 looks bad. This is either a bad valve, or a typing mistake.

Mark

Edited November 14, 2012 by MarkBassChat

[Bassman Rich]

Thanks to all, both here and at Marshall Amps for your input, it has been invaluable.

It seems from Mark above and Dave at Marshall, that ideally I would replace all the valves with matched ones. Given that it cost £80 for 2 svetlana winged 'c' s, this seems a bit extravagant as this is my second, back up amp. I may consider this at a later time.

So I swapped a couple of valves around, and then replaced the one that was 'most' different to the others with my one remaining spare tung sol, and hey presto, all the tung sol ones sit between 350 and 550 mv, the 2 svets are at 725/740 mv. When I get the chance to put the chassis back in the case, I'll give it a good old blast down the rehearsal room.

Ps Mark, that 87mv was not a typo, more like a test lead come off the resistor leg, as when I retested today, it was pretty much like all the others. Sorry to throw a red herring into what has been a pretty tortuous but interesting learning experience for me.

Pps A 30 v rms 1khz signal is so loud in the house through my 4 ohm 2 x 15, the mrs had to run down the garden with her phone to continue her work phone call