Rising grid bias voltage

[Buddha's Guitar Tech]

Having an issue with a Fender '65 Deluxe Reverb Reissue.

Loud volumes result in a cut-off - sounds like the power is totally drained; ie not a loose connection. Signal comes back very quickly, which led me to believe the power supply was being totally drained and taxed.

Of course, when i took it apart, it stopped doing it!

But i took some voltage measurements anyway and the supply does dip, as you would expect, but nothing too drastic. What DOES happen, which is new to me is, as the volume goes up (and power draw increases) the grid bias voltage CLIMBS. ie going from -48 to -61.
i'm guessing that what WAS happening was the bias voltage climbed to a point where it cut off the 6V6s momentarily.

Figures look like this:
idle: 430 plates, -48 grid
volume up: 367 plates, -61 grid.

We are talking 6V6s here. i tend to think the idle voltage is pretty high. i am in the UK, this is an export amp.

There's also a bit of solid state circuitry in there that doesn't appear on any schematic i can find - it's a PR239 model.

Never tested for rising bias before, is this normal or odd?

[tubeswell]

6V6s like to idle between 20 to 30 volts - depending on plate voltage and current. Not to say that it couldn't possibly be higher, but in this case I'd say your bias voltage leaves it way too cold. If the bias voltage itself is changing of its own volition, then at I guess I'd say it could possibly be a flaky resistor in the bias supply's voltage divider that is suffering from temperature related over heating and drifting. Try some other resistors there and see.

[Phil_S]

Of course, when i took it apart, it stopped doing it!

OK, so think about this. What's different when you've pulled the chassis?

1) Maybe there is a loose component in the bias circuit that was jostled back in place? That suggests a bad solder.

2) Maybe there was something in there shorting it that fell out and you didn't see it?

3) Maybe there was heat accumulating and in the open air whatever was stressed isn't?

The bias circuit doesn't have that many parts. Consider replacing all the resistors. I 'd suspect #3 is the problem. More likely a resistor than a cap. This will also address the broken solder problem if it exists. The good news is that there is no tone running on the bias circuit, so just get good quality resistors -- the type of material (carbon film, metal film, etc.) doesn't matter much, if at all.

[Buddha's Guitar Tech]

Thanks for the replies. It's a bit of a funny one.

bias voltage rising, it's only on the tube side... The two, quite normal 220k resistors in the bias circuit... Under test, one end of each resistor's voltage rises, the other stays the same. It's the tube side (ie the grid pin) where the voltage rise seems to come "from." Just wondering if this increase is what's causing the cut off. So far, it's the most reasonable explanation. Is this grid current?

On the transformer winding side, no change in voltage.

Don't think heat is an issue, it happened about a minute after switching on.
Shall check for loose stuff.

With the plate voltage at around 430-odd, the plate current is a mere 6mA, but it still sounds nice and warm, and the idle dissipation is around 3w per tube. It's conservative maybe, but even with the original 'liberal' bias, it was dissipating something like 11 watts per side. And it was cutting out then too.

Overall, something just doesn't seem right. There are also 2 extra PCBs not in the schematic, they seem to be the trem LFO, as there is no "roach" where there is space marked out for one on the main PCB.

[Buddha's Guitar Tech]

Hang on a minute. Just refined something and some bells are ringing.

Biased to 8W plate dissipation (still sounds high to me though)

As expected, when volume cranks, overall voltage drops. This time, so does the bias, just a bit. Fair enough.

i'm noticing now that the bias voltage rises (ie -38 to -50) only when the 6V6s start to clip. This makes some sense i think. Perhaps grid current is simply there, but not really the problem. Dunno yet.

Annoying thing now is that under idle, the screens are higher voltage than the plates. Wondering if i should put in a different resistor before the screens to compensate. (there is a choke in the circuit)

[VacuumVoodoo]

Check coupling capacitors between phase inverter and power tube control grids for incontinence ie current leakage. Or just replace them without checking

[Buddha's Guitar Tech]

Already checked, nothing wrong there.

When the valves go into clipping, ie positive grid, doesn't grid current kick in? If so, it would seem normal.

[tubeswell]

Already checked, nothing wrong there.

When the valves go into clipping, ie positive grid, doesn't grid current kick in? If so, it would seem normal.

Yeah grid current limiting kicks in as the grid goes towards 0. But blocking distortion happens when the electrons can't leak back through the grid leak resistors fast enough to discharge the (extra) accumulated +ve charge on the coupling cap(s). And it happens under louder playing conditions.

What size are the coupling caps?

Are there grid stoppers on the 6V6s (to inhibit grid current)?

[Buddha's Guitar Tech]

Thanks.

Coupling caps from PI are both 0.1
They're ok, no undue DC getting through.

Usual fender 1K5 on the grids.

i now don't think there's anything wrong with the bias circuit. Deer in the headlights - good learning day.

i'm still wondering about the screen voltage though. The amp sounds good, but the screens are above the plates.

Wondering should i bother trying to change it with some extra resistor(s) or let it go, as it must have always been behaving like this.

[Colossal]

i'm still wondering about the screen voltage though. The amp sounds good, but the screens are above the plates.

Wondering should i bother trying to change it with some extra resistor(s) or let it go, as it must have always been behaving like this.

This sounds odd to me; the screen voltage being above the plates...that can't be right. Did someone do a cap job on the amp and then reverse the B+ leads from plates to the screens and vice versa?

[Buddha's Guitar Tech]

Pretty sure it's all stock. Clean as a whistle.

[Colossal]

Pretty sure it's all stock. Clean as a whistle.

Hmm, well that is odd then. FWIW, usually when troubleshooting, I'll measure the voltages at every single point in the amp; cathodes, plates, screens, and document them on a schematic with the "expected" voltages written at the same points (or I'll put the expected values vs. measured/actual in Excel in two columns and compare them, looking for anomalies or anything that stands out).

Have you measured the B+ at each power supply cap and resistor node? (I apologize if this has been discussed previously...). How about the integrity of the tube sockets or a blob of solder where it shouldn't be, grounding when getting hot? Anything strange with the screen connection? It seems like the high voltage on the screen is the shell casing...just need to find the smoking gun.

[jjman]

My screens are 3 volts higher than the plates on my '71. I think the bias reading during signal is not "valid." During clipping the grid stays around zero for some time but is not prevented from swinging lower and lower on the bottom half as you increase the signal. So the "bias" reading you're seeing may be an "average" of the ac signal voltage.

[Buddha's Guitar Tech]

Cheers guys.

Having some time away from it now, i remember something similar with a champ i built. The resistance in the OT primary bumps down the plate voltage. This amp has 204/210 ohms each leg of the OT primary, with a 4 volt (or so) drop at 20mA (or so).

So it stands to reason that you'll get more loss there as the screens pull so little current. HOWEVER, with all the other 12AX7s in this amp, you'd think you'd get more current pull - but aha! There is a choke in line, with very little resistance, so you're not going to get much of a drop when it comes to the screens. (as is the case here)

i figure what i'll do is shove a sine wave in again, under dummy load, check plate AND screen voltages as the volume rises, see if the screens fall below plates under PLAYING conditions. If it clears up then, great.
If not, i'll have to put in way bigger screen resistors. After all, the screens are there to prevent secondary emission (s.e.). Surely if they're at a higher potential, they will only ENCOURAGE s.e.

Any thoughts?

[tonewood]

mid 70's champ from the flea mkt. Got it going and 6V6 has 353v on the plate and 360 on the screen. Sounds like it is working fine but seems like the scn would be attracting more electrons than it should.
The schematic for the AA764 says plate =342 and scn =340.
The GE 6V6GT data says 315v max for both plate and screen.

like Buddha's Guitar Tech said-
Any thoughts?