Weird howl--Marshall

[g1]

The idea of shielded wires connected to plates does not sound good to me, even if the shields are disconnected.

[Helmholtz]

The idea of shielded wires connected to plates does not sound good to me, even if the shields are disconnected.

Why?

It makes a lot of sense to shield plate wires (as well as grid wires, of course) as this reduces radiation of electric signal fields and thus reduces unwanted coupling.

A non-connected wire shield has no effect at all.

[g1]

Why?

It makes a lot of sense to shield plate wires (as well as grid wires, of course) as this reduces radiation of electric signal fields and thus reduces unwanted coupling.

A non-connected wire shield has no effect at all.

I was worried about stray capacitance as the fault may be due to oscillation.

[Helmholtz]

I was worried about stray capacitance as the fault may be due to oscillation.

A non-connected shield is just a floating, non-connected capacitor of some 10pF, not changing fields or potentials.

A grounded shield is an equipotential surface at ground potential which doesn't radiate or couple as it doesn't carry signal voltage. It reduces stray capacitance by grounding the electric field around the wire.

[Chuck H]

Marshall had a couple of JMP models where they would tack the plate and grid leads side by side for an inch to an inch and a half on the chassis with goo, hot melt, or other. Plate connection of the shield has been kicked around here a few times. I haven't seen it for myself, but Ken Fischer supposedly did it on at least one TW Express. One of my designs uses a tiny capacitor from plate to grid on the input triode. Now THAT would be a more likely problem. If the cap ever failed short instead of open there would be significant voltage on the grid,.. And the guitar! At some point I'm going to experiment with just twisting the plate and grid leads together for a few turns, which seems safer.

[The Dude]

Class Y cap maybe?

[Chuck H]

Class Y cap maybe?

I considered something like that, but my practical side says no. Rather, make the amp work without the need to spec a special part. Still something to keep in mind though since there are a few users of this amp out there. If my eventual "work around" doesn't suit their ears then a class Y cap it shall be

[Valvehead]

Im going to swap them out anyway as they are a different color than the rest...it cant hurt...the stock wires were not shielded . Ill post back if there is a difference in the hum

[Helmholtz]

One of my designs uses a tiny capacitor from plate to grid on the input triode.

A capacitor between plate and grid in a "grounded" cathode stage should have the same effect as wiring a capacitor having a capactance of G times the plate to grid capacitance between grid and ground (Miller effect). G is the voltage gain of the stage.
The effective grid input capacitance and the grid stopper (+ source impedance) constitute a low pass filter.

[Valvehead]

Changed those wires---hum still there. I went back through the amp and checked all grounds

[Chuck H]

A capacitor between plate and grid in a "grounded" cathode stage should have the same effect as wiring a capacitor having a capactance of G times the plate to grid capacitance between grid and ground (Miller effect). G is the voltage gain of the stage.
The effective grid input capacitance and the grid stopper (+ source impedance) constitute a low pass filter.

Yep. And nope. With the cap from plate to grid the affect of the cap varies with that triodes gain. In a tube amp that can be significant with voltage sag. So when the amp is tanking voltages there's less HF roll off. To my ears this makes for a nice dynamic. When the amp is cleaner, and would typically be harsher, there's a little softening of the top end. When the amp is pushing hard there's less HF attenuation muffling the much wanted harmonics.

I first saw this done in a couple of Marshall models from the 90's. Then later I had a Fender Prosonic that included them in the design. The Marshall values were too high and really squashed the tone. The Prosonic used the circuit on two or three of the gain stages, which seems excessive also. My design uses a single 4.7pf cap. If multiplied by what is probably the average gain of the triode I guess that would be like a 280pf cap to ground. Sort of like adding an extra 10' of cable I guess. Except for the added dynamic element. Which is small, but does give the option some merit over a simple shunt to ground I think.

[Chuck H]

Changed those wires---hum still there. I went back through the amp and checked all grounds

Sorry Helmholtz and I are getting a little sideways on your driveway

You've made some headway. Is there still excessive noise or hum? I've seen hum chasing go from fixing excessive hum to an attempt to eliminate ALL hum. Even in designs that are prone to a little noise or hum. That doesn't mean things can't be improved further. Just that there may be a point of diminishing returns for the effort.

[Helmholtz]

In a tube amp that can be significant with voltage sag. So when the amp is tanking voltages there's less HF roll off.

Interesting idea.
What is the stage gain difference between idle and full power in your design? Over 10%?
It's typically less than 10% for a plate supply change of 50V.

[HaroldBrooks]

Sorry Helmholtz and I are getting a little sideways on your driveway

You've made some headway. Is there still excessive noise or hum? I've seen hum chasing go from fixing excessive hum to an attempt to eliminate ALL hum. Even in designs that are prone to a little noise or hum. That doesn't mean things can't be improved further. Just that there may be a point of diminishing returns for the effort.

I've gotten caught up in reducing hum and then trying to eliminate every shred of it. Problem started when I got a single 60's Valco amp (Gretsch 6156) that was dead silent, even after I boosted the gain a bit by changing a few things. That's a hard act to follow, and not to be expected from every single amp all the time.

Hum always sounds much worse in a quite room in the dead of night, on a scale of 3 to one. I had another amp that still has some hum and I almost talked myself into not bringing it to our last Gig. Turns out we needed it as the other guitarist's amp cut up the night before, and he didn't have a back up. I lent him the amp, and once we were in the Pub you couldn't hear any hum at all ! and this is early in the day. Ordinary background noise has a way of "Cancelling" out hum on it's own, it masks it.

But then there's the issue of Low Freq Hum modulating upper frequencies that doesn't sit right with me. It depends obviously on the amplitude of the hum, but low frequencies have a lot of energy by the time they are heard.

Still, there's a time to cut it free, and accept some hum in any tube amp. It makes absolute sense to me... (during the day that is).

[Chuck H]

Interesting idea.
What is the stage gain difference between idle and full power in your design? Over 10%?
It's typically less than 10% for a plate change of 50V.

Did I mention the difference was small I build these on a bench. I don't test them in a lab. But I'll guess that the gain difference could be 10%, but probably not more. That would seem like a lot. So the real world dynamic of the plate to grid cap would be like a grid to ground cap shifting some between 250pf and 280pf. Pretty insignificant when I look at the numbers. But there are a lot of interesting things happening in a tube and I don't doubt that voltages can bounce high as well as low. And then there's the recovery time of the filters, etc. In other words, there's enough variables in operating conditions that I think the plate to grid cap can make an audible, if subtle difference from a grid to ground cap. I don't have adequate technical knowledge to analyze it on paper well and I don't have one open on the bench right now or I'd just test it.