If it is not too difficult, it would be best to remove both grid coupling capacitors until you sort out the problem.
If everything is ok without them, then one capacitor is bad, so replace both with a different type not prone to leakage.
If everything is ok without them, then one capacitor is bad, so replace both with a different type not prone to leakage.
Can they be in place to test for leakage?, I was going to remove the valve, place (-) lead to chassis ground, (+) lead to grid (pin 5) and check for voltage any lower than the -41v that is standard.
I've taken out the bad resistor already so won't have any cathode resistor.
Yes, but watch out for the B+ rising excessively (for the filter capacitors) without the tube in place.
Bear in mind that meter loading may change the true DC voltage at the grid, since it is a high impedance circuit node.
Bear in mind that meter loading may change the true DC voltage at the grid, since it is a high impedance circuit node.
OK, thanks, my filter caps are 500v, b+ is 436v, plate voltage on KT88's is 425v. Not much room for movement here.
OK, well, you know I'm not too sure now but I took the faulty channel output tube out and, firstly, wanted to test B+.(normally 436v)
Voltage jumped to 535v, I think it was still climbing before I switched it off. It was only on a few seconds.
I think all I can do now is replace the coupling caps, the resistors and the valves and hopefully correct the problem without knowing, exactly, what the problem is.
I thought that keeping the driver and one output tube in would have provided enough load but it appears to have registered no load at all.
Voltage jumped to 535v, I think it was still climbing before I switched it off. It was only on a few seconds.
I think all I can do now is replace the coupling caps, the resistors and the valves and hopefully correct the problem without knowing, exactly, what the problem is.
I thought that keeping the driver and one output tube in would have provided enough load but it appears to have registered no load at all.
I did, notice, when the amp was off, that the voltage very gradually reduced and at the same time I tested the driver plate end of the coupling cap which registered (at that time) 410v, the other side of the cap (going to KT88 grid) registered 0v. The good channel did the same, so when there is voltage in the circuit, albeit powered off, the caps do not appear to be letting any dc voltage through.
Is it as simple as that, or it doesn't work like that as its off, don't know.
Is it as simple as that, or it doesn't work like that as its off, don't know.
That's a good sign, but the best test is to check when powered, with tubes, but with no coupling capacitors.
If the bias is ok then, leaky capacitors are likely to be the problem.
It may be simplest to just replace the coupling capacitors with new ones that are known to not leak significantly,
like polypropylene types. Then you can be certain the capacitors will not be the problem.
Of course, as a separate test, you should also try swapping the two tubes to see if the problem follows the tube.
If the bias is ok then, leaky capacitors are likely to be the problem.
It may be simplest to just replace the coupling capacitors with new ones that are known to not leak significantly,
like polypropylene types. Then you can be certain the capacitors will not be the problem.
Of course, as a separate test, you should also try swapping the two tubes to see if the problem follows the tube.
Last edited:
That amplifier was used for many hours , tube sockets can fail , also a "tired" tube could go into a runaway situation when after some time with power on the current goes up until red plating or burning other components . So when that tube is tested the current should be monitored carefully over a period of time .
We'll assume both and I'll be replacing the caps and the valves. We know it's one of them but, for me, it will be better to just replace both at the same time.
A number of mods to the chassis to aid cooling, vents, heatsinks on the resistors and waiting for deliveries mean all this will take a week or so to complete.
Thank you all for your comments, I'll update when all finished and hopefully finish this off with resolution.