Valve buffered IGC - volume pot causes DC offset, crackling

[AndrewT]

Originally posted by Franz Gysi
your heatsinks for the LM chips are much too small!

LoL

[analog_sa]

There may be some good news though. Your tubes don't look much like ECC88/6922, although i have never seen Chinese ones. Quite possibly they are ECC83 or similar which will explain the weird biasing.

[gaplessophile]

Well, as it sounds likely that my valves are faulty, which is the only thing I can think of too, I think I'll order a new one. Any particular guidelines for not getting ripped off on eBay again? (I thought it was a bit strange that my "military grade NOS" Shuguang valves cost only a few US dollars each - and they go this cheaply all the time. The seller did oddly have 100% positive feedback though.)

Thanks so much for all the help! I'll post back as soon as I've got a new valve.

I'm looking in other threads for recommendations, but if anyone has any advice on what brand of 6922 valve to get, I'd love to hear it. Looking for a warm, smooth sound.

[analog_sa]

I don't think your valves are faulty - they are just not 6922s. Being conservative i like seeing a recognisable type stamped on the valve: ECC88, 6DJ8, 6922, 6N1P, 6H23 will all do fine.Why not start a new thread in the tube section with a close up pic of the internal structure and see what people think. You will get very little feedback in this section. As for warm and smooth i am not sure 6922 do that at all.

[leadbelly]

Originally posted by gaplessophile
Well, the centre of the valve supplies and the negative rail of the heater filament supply is connected to the central ground, along with the centre of the chip amp supplies, and this central ground is then connected to the centre tap of the transformer (only one transformer for both valve and chips), and to the chassis ground and to the earth wire in the power cord.

Is this okay?

That's your wiring mistake right there. You can't short the negative leg of the bridge feeding the heater supply to the centre tap of the transformer feeding the bridge. Remove this wire.

[Franz Gysi]

That's your wiring mistake right there. You can't short the negative leg of the bridge feeding the heater supply to the centre tap of the transformer feeding the bridge. Remove this wire.

Wrong: the filaments voltage has (optimally) to be set to the cathodes potential, therefore avoiding a parasitic diode between the cathode and the filament.

Absolutely wrong is not to connect the filaments voltage to any potential, floating.

For higher voltages the max Ufk is defined in the tubes datasheet.

Franz

[analog_sa]

Originally posted by leadbelly


You can't short the negative leg of the bridge feeding the heater supply to the centre tap of the transformer feeding the bridge.

Ehm, why?

Not that any of this matters for the incorrect biasing.

[leadbelly]

You are both wrong. Read what he is doing: he is connecting the same set of transformer secondaries to 4 separate bridges, and grounding the centre tap. If the negative rail of the filament leg is ungrounded, the filament supply does NOT float. It's reference voltage is set by the the centre tap of the transformer secondary, which is already grounded. By lifting this short, the heater supply circuit becomes +/- 3VDC. The way he has wired it, the -3VDC leg is shorted to ground, causing all these problems.

[gaplessophile]

Sorry - wasn't clear. The filament supply does have its own separate transformer. The other (main) transformer powers the valve plates+cathodes (as in the diagram) and also the LM chips.

[gaplessophile]

SOLVED! It was the tube. I bought a brand new JJ E88CC and it worked perfectly first time. That's the last time I trust slightly dodgy looking Ebay sales.

Thanks so much to all who helped!