Problems with E88CC / 6922 drawing grid current on power-up

[Sassenach]

I'm using E88CCs in a cascoded differential pair as the input stage of a fully-balanced power amplifier. Twin muting relays pull the signal ends of the grid stopper resistors to 0V until the valves have warmed up, and there is an initial time delay to allow the heaters to get up to temperature before the HT is turned on. During this time, the grid acts as the anode of a diode, and current flows from 0V through the muting relays, into the grids and down the tail resistor to the negative bias rail. This is not totally unexpected and shouldn't cause a problem because the grid current flow should cease as soon as the HT rail comes up. At least, that's the theory - and in most cases this is exactly what happens. However, I've found that with certain individual valves, grid current continues to flow for a while even after current starts flowing from the HT rail, and this means that there is an audible "thump" when the muting relays open.

I need to find a solution to this that doesn't involve testing and selecting valves that don't exhibit the problem. Has anyone else come across this phenomenon or got any ideas about how I could get around it?

 

[DF96]

You may be heating the grid so it then emits for a while under 'normal' operation. Try adding small silicon diodes between grid and cathode, so they are forward biased while warming up but reverse biased during normal operation.

The ECC88 and its cousins has a very fine grid, not designed to cope with any grid current at all.

Alternatively, it could simply be that some of your valves are a bit gassy. They may settle down with use, or they may not.

 

[Sassenach]

Yes, I was thinking along the lines of reverse-biased diodes between grid and cathode - I will give this a go. The grid-stoppers are 4K7, so the grid current is hopefully limited to a below-damaging level. It's interesting that most of the valves I've tried are completely okay. I've been using brand new JJ E88CCs, so maybe it's a gassing issue as you suggest. Hopefully the getter will carry on doing its job. Many thanks.

 

[Ketje]

At normal conditions the -Vg is there thanks to the fact that the cathode is at +1...3V above ground,the grid is at 0V.During warm-up and/or without Va the cathode is pulled below ground,causing gridcurrent.Simply put a diode (1N4148)between cathode and ground (anode diode on ground).Now Vgk may go negative but only 0.6V positive.
Mona

 

[DF96]

Read post 1 again, more carefully. This is an LTP with the tail going to a negative supply.

Parallel diode, as I said, is the solution. Not series diode.

 

[Rundmaus]

I powered a 220V DC relay coil from the positive rail, switching on the heater current for the negative rail when it comes up - so the B+ is always up before B-. No power-on thumps up to now...

Greetings,
Andreas

 

[Ketje]

No mistake
Do it like this
Mona

 

[Sassenach]

A diode from the cathodes to ground won't work because the voltage on the cathodes sits at around 1.6V when the HT comes up. (The cathodes are returned to a -15V rail via a BJT current sink that sets the tail current at 12mA). The cathodes sit at around -12V when the grid current is flowing with no HT present.

I've been trying to recreate the fault condition this morning but the grids keep returning to 0V whenever the HT comes on. Typical - I hate intermittent faults!

 

[DF96]

Do it like this

OK, I understand now what you meant. Yes, that would work. Only snag is that it would limit the possible input voltage. Whether this would be a bigger or smaller limit than my suggestion would depend on the details of the circuit.

 

[Sassenach]

Ignore my previous comment - I'd got the diode the wrong way round in my mind! I'll need to do some measurement of Vk over the full signal range to check that it doesn't cause any issues.

 

[wa2ise]

Maybe a 3rd relay to connect the cathodes to a DC bias of a volt or so (whatever the cathode runs at during normal operation) until the heaters are hot, like you did for teh grids. Purpose of the DC bias is to avoid a thump when the relay opens.

 

[Sassenach]

I think I'm finally starting to get to the bottom of this as I have a valve on the bench that is repeatably demonstrating the problem. I fitted the diode between the cathodes and 0V, and with scope probes on the cathodes and one of the grids I see that (as expected) the cathodes are initially clamped by the diode to about -0.7V. As the heater warms up, the grids start to draw current and their voltage slowly drops to the same potential as the cathodes. On a good valve, when the HT comes up, the cathode voltage immediately jumps to about 1.3V to set the bias current at the required 12mA. However, on a faulty valve, the cathodes - and hence the grids - only rise to about -0.5V. Over the course of a further minute or so, the cathode voltage slowly increases, eventually reaching 1.3V. The muting relays open 55s after the heaters came on (and 25s after the HT came up), so with the faulty valve, there is a loud "pop" from the speaker. With a good valve, the grids are sitting at 0V, so the relays open silently.

It seems odd to me that the faulty valve is taking so much longer to warm up (or at least to start passing anode current). Does anyone have any experience of this?

 

[DF96]

Could be low emission. Could be gas needing to be swept up by the getter, which may work better when hot.

If the valve is struggling to produce enough electrons to maintain the anode current then shot noise will be worse, and the cathode will not last very long.

 

[DavesNotHere]

instead of grounding the grid-stop, make it switch to the Anode of a diode whose cathode is connected to the cathode of the tube.

 

[Sassenach]

I have a theory that the cause of the problem is down the the fact that I am running the lower E88CCs with only 80V on their anodes. I'm thinking that if I increase the anode voltage, Vgk will also increase and reduce the likelihood of grid current flowing.

 

[SemperFi]

80V is fine for E88CCs.
The solution is of course not to delay b+ ;-)

 

[djmiddelkoop]

Check the valve for heater-cathode leakage and respect the heater-cathode voltage limits.