Power Tubes' total input capacitance - pF to nF

[Tillydog]

I was talking about Marshall, take a look here, 5x5k6 grid stoppers:
http://www.ampwares.com/schematics/mars ... 0_100w.pdf

It's not a mistake, 6100 30th 100W heads have 5x5k6 grid stoppers.

I'd bet that the second 5k6 grid stopper on V11 is a mistake - there is no part number associated with it. Both the schematics you posted are different revisions the same drawing number : 6100-63-04. Has anyone seen the 5th grid stopper in real life?

From an AC point of view, it makes no difference if the 47pF capacitor across the PI output is before or after the coupling caps - (they are > 2 orders of magnitude bigger in value). It more likely depends on where is it most convenient to mount the 47pF capacitor.

The normal reason for increasing grid stoppers would be to avoid blocking distortion - maybe if each PI output is driving 2 grids through 1 coupling cap, each needs twice the value grid stopper as would be necessary driving only one grid through the same coupling cap?

Input capacitance of tetrodes & pentodes is much less than you might think because there is (nearly) no Miller effect.

If an output stage uses NFB, small value caps between grid and ground would usually be to improve the feedback loop stability ("slugging the dominant pole"), and would usually cut off well, well above audio frequencies.

Different grid leak / bias feed resistors in conjunction with the PI coupling caps give different LF cut-offs.

Just several of my 0.02p

Andy

[roberto]

Thank you andy,
yes, I've seen the 5th grid stopper in real life, this is what I was talking about. We talked about that also on al italian forum, because another moderator has noticed that extra resistor.

I know there's nothing that could explain why it sounds different.
This is the reason I'm asking.

[roberto]

Here you can find the DSL100 schematic of the 1997, when grid stoppers were 220k. 2003 version had 5k6 grid stoppers.
http://www.mif.pg.gda.pl/homepages/tom/ ... cm2000.zip

The 220k could be due to poor quality EL34s of the period (remember that Marshall switched to 6L6 due to this reason), so to limit the bias excursion ratio to B class while overdriving the power amp.

Looking at this link:
http://www.ampbooks.com/home/amplifier- ... excursion/

I can see that a standard Marshall 22n 220k 5k6 configuration has a bias excurtion ratio of 0.59, with a bias excurtion time of 1.62 ms and a bias recovery time of 6.34 ms.

A modified 22n 220k 220k has a bias excurtion ratio of 0.24, with a bias excurtion time of 3.92 ms and a bias recovery time of 6.34 ms.

[katopan]

Just wanted to follow up that I ran a simple SPICE sim of an EL34. 1k to 100k for the grid stopper showed no upper end difference in the audible range.

[roberto]

Thank you for your time katopan,
this time I haven't used SPICE, because it'n not a classical grid stopper cutoff. I've tried it again, and it's more evident when I raise the volume, so could be connected with bias excursion while playing (I've to try lower values of the second bias capacitor to see if the effect enhances).