Pentode Screen Capacitors ? I don't get it !

[HaroldBrooks]

Recently I was reading about screen capacitors used to partially bypass a screen resistor, the kind that are connected to a single ended pentode (in this case a single 6V6) and it's cathode pin no. 8.

I did just that and using a rather large .047uf cap I noticed an increase in Highs. I went further and clipped in three of them in parallel to yield approximately .15uf and even more highs came in with some high mids.

It got me to thinking, how does bypassing a DC controlled screen grid, with no AC signal yield more high frequencies ? I was using some white background noise from my pedals as a means of testing the frequency changes, but this is what is happening.

I thought the role of the screen grid in general was to act as a positive accelerator to the electrons on there way to the plate. How is it that it is influencing the frequency content of an A/C signal in a specific range of frequency, when it's power source is a more or less stable DC ?

Does this happen because of the connection to the cathode ? would it happen if the capacitor was connected to ground directly before the cathode bypass cap ? Does connecting a screen capacitor to the cathode also add an AC component to the action of the screen grid ? I have a feeling that might be it, but I am only guessing.

Also, if you were to FULLY bypass the screen resistor by using a very large capacitor to the cathode, would the resistor still act as a current limiter to the screen ? So far using only the .047uf cap, the resistor still drops voltage.

As always, thanks for your insights !

[Helmholtz]

Please post a drawing with parts' values.

The basic principle is this: Screen voltage controls tube gain. An unbypassed screen resistor allows screen voltage to vary depending on screen current (larger signal > higher screen current > lower screen voltage > lower gain). This causes a current NFB, reducing gain. The bypass cap prevents this NFB above frequencies determined by the values of screen resistor and cap.

More here: http://www.valvewizard.co.uk/pentode.html

[HaroldBrooks]

Please post a drawing with parts' values.

The basic principle is this: Screen voltage controls tube gain. An unbypassed screen resistor allows screen voltage to vary depending on screen current (larger signal > higher screen current > lower screen voltage > lower gain). This causes a current NFB, reducing gain. The bypass cap prevents this NFB above frequencies determined by the values of screen resistor and cap.

More here: http://www.valvewizard.co.uk/pentode.html

Here's the schematic to the amp in question. I've done a few other mods to it, but most recent is the addition of a 1.5k screen resistor and the .047uf screen bypass cap to the cathode pin 8 of the 6V6 :
https://www.seymourduncan.com/forum/...1&d=1517797174

Still not sure how the NFB works in this context. I was assuming the screen resistor worked on more or less DC type voltage only, regardless of it pitching higher and lower do to the requirement for more current. I was thinking it was a more stable element that raised the gain on the tube across the board, not selective as to AC signal content.

Perhaps the screen's action is similar to an that with an AC signal ? and creates a gain or loss overall dependent of frequency ? If that's the mechanism, then I understand the Frequency dependent nature of a screen capacitor as it pertains to changing the tube output, and as a filter of that output.

[Helmholtz]

Just like with grid No.1, the control effect of the screen grid depends on momentary/instantaneous voltage. The momentary screen voltage can be seen as composed of a bias DC voltage (being the screen voltage at idle) minus the AC voltage produced by the screen current. In other words the screen current produces a NFB voltage across the screen resistor just like the cathode current produces a NFB voltage across the cathode resistor.

[Chuck H]

Just to add to the discussion...

Ok, maybe not add. More like raise new questions that most of us (including me) aren't sure about.

The screen resistor is there to limit screen current. So we can bypass it and get some increase in gain, correct? This is actually something I mentioned here a long time ago and a good point was raised. And it's a point I've become familiar with in the course of designing. Bypassing the screen grid resistor reduces screen circuit impedance at the bypassed frequencies and I believe that somewhat negates any operational safety the resistor was providing. So bypassing the screen grids in an an amp that's likely to be overdriven, like many vintage Marshall designs, could lead to the same blown tubes with screen failures that the (now ubiquitous) 1k resistors were there to prevent. I think this is right, but if I'm wrong I'd want to know.

[Jazz P Bass]

I think that the bypassed screen resistor on a pentode tube is purely a preamp tube concept.

[HaroldBrooks]

Just to add to the discussion...

Ok, maybe not add. More like raise new questions that most of us (including me) aren't sure about.

The screen resistor is there to limit screen current. So we can bypass it and get some increase in gain, correct? This is actually something I mentioned here a long time ago and a good point was raised. And it's a point I've become familiar with in the course of designing. Bypassing the screen grid resistor reduces screen circuit impedance at the bypassed frequencies and I believe that somewhat negates any operational safety the resistor was providing. So bypassing the screen grids in an an amp that's likely to be overdriven, like many vintage Marshall designs, could lead to the same blown tubes with screen failures that the (now ubiquitous) 1k resistors were there to prevent. I think this is right, but if I'm wrong I'd want to know.

Here's what I "think".

You are correct, perhaps up to the point were the capacitor starts to filter out a lot of the low frequency related current, the kind that might cause issues that the screen resistor was there to defeat. Generally there is far less energy in frequencies above ~ 2.5khz than below (what I start to call treble), and I believe I am correct in saying that if you size the capacitor to filter the lows and low mids, there is less risk associated with bypassing the venerable screen resistor.

Not sure if this is correct, but it's what I believe at this point, LOL.

Anyone please feel free to step in and correct my assumptions, and I will be most grateful.

__________

By the by, what does everyone here consider "Bass", "Midrange", and "Treble" on a frequency scale ? I will start :

Bass = 20hz to 500hz, Mids = 500hz to 2,500hz, and treble is everything above 2,500hz.

Please feel post your opinions, or perhaps any standard you may know of.

[Helmholtz]

Bypassing the screen grid resistor reduces screen circuit impedance at the bypassed frequencies

OK, I will try an explanation using your terms.

You are right, bypassing the screen resistor lowers the screen circuit/supply (AC) impedance.
Lower impedance means less NFB (= more gain) and higher screen current peaks.

A bypass cap provides an additional AC current path. The impedance of a cap decreases with increasing frequency allowing for more current drain. So the RC highpass filter consisting of screen resistor and parallel acting bypass cap allows for more AC screen current above the corner frequency.

So bypassing the screen grids in an an amp that's likely to be overdriven, like many vintage Marshall designs, could lead to the same blown tubes with screen failures that the (now ubiquitous) 1k resistors were there to prevent.

Correct as well.

As said above, bypassing allows for higher screen current peaks. This can be a problem with high screen supply voltages and high load impedances.

And only an unbypassed screen resistor acts as a "stopper", reducing the risk of HF instability.


Regarding audio frequencies it should not make a difference if the bypass cap is connected directly across the screen resistor or between screen and ground - or between screen and cathode (as long as there is no unbypassed cathode resistor which would introduce some extra feedback).
So it's not clear to me why Merlin recommends to always connect the bypass cap between screen and cathode. He doesn't give a reasoning in his web article (and the pentode circuit examples in his book all show the lower end connected to ground).

But I could imagine that connecting the cap to the cathode might have some advantage regarding HF stability and maybe regarding ripple rejection in some circuits.

[mozz]

Put a capacitor across that cathode resistor if you want some more umph.

[Helmholtz]

As there still seems to be some misunderstanding, I will try to answer your questions explicitely.

It got me to thinking, how does bypassing a DC controlled screen grid, with no AC signal yield more high frequencies ?

The screen is not only controlled by average DC but rather by total screen voltage which may consist of DC plus AC signal on top. The tube internal fields immediately follow any changes of screen voltage.

An unbypassed screen resistor produces an AC voltage drop caused by the changing screen current. This way the tube generates its own screen feedback voltage. Consequently the screen voltage is no longer pure DC.

Does this happen because of the connection to the cathode ?

No.

would it happen if the capacitor was connected to ground

Yes.

Does connecting a screen capacitor to the cathode also add an AC component to the action of the screen grid ?

Only if there is an unbypassed cathode resistor.

Also, if you were to FULLY bypass the screen resistor by using a very large capacitor to the cathode, would the resistor still act as a current limiter to the screen ?

No.

So far using only the .047uf cap, the resistor still drops voltage.

A cap won't change (average) DC voltage. You need to distinguish between DC and AC/signal voltages.
To see the effect of the cap on the AC part of the screen voltage you need a scope.

[HaroldBrooks]

As there still seems to be some misunderstanding, I will try to answer your questions explicitely.




The screen is not only controlled by average DC but rather by total screen voltage which may consist of DC plus AC signal on top. The tube internal fields immediately follow any changes of screen voltage.

An unbypassed screen resistor produces an AC voltage drop caused by changing screen current. Consequently the screen voltage is no longer pure DC.



No.


Yes.



Only if there is an unbypassed cathode resistor.



No.



A cap won't change (average) DC voltage. You need to distinguish between DC and AC/signal voltages.
To see the effect of the cap on the AC part of the screen voltage you need a scope.

Nice, thanks for taking the time ! I think I understand it a lot more now, and it makes sense based on my limited knowledge already.

I have to get a scope, but I'm afraid it will take up too much of my time, as I already stay up late with this stuff and practicing guitar ! Thank goodness my wife tolerates me (so far so good) !