where: Cin = input capacitance
Cgk = grid-to-cathode capacitance, composed of the internal tube capacitance plus the stray capacitance
Cgp = grid-to-plate capacitance, composed of the internal tube capacitance plus the stray capacitance
A = stage gain
Now, what's the total input capacitance of EL34s 6L6s EL84s 6V6s and so on?
You can clearly ear the difference between a 1k5 to 2k7 to 5k6 grid stopper resistor on EL34, or even a 2k2 to 2k7 change in 6L6 amps. So it's absolutely in the audible 20-20kHz range.
If you search on datasheets you'll always find something below 100pF, and a 2k2 grid stopper will give you a 720kHz cutoff.. but the capacitance has to be at least one order of magnitude higher.
Roberto you have to take into account the Miller effect in the power amp because the output stage is operating as an inverting gain stage. In this case the typically small grid-to-plate capacitance (Cgp) gets multiplied by the gain of one output tube. If we were to operate our amps with cathode follower arrangement (OTL) then you don't run into the Miller effect. In this case we would have to use much larger grid stoppers to squash RF noise.
Miller effect in power pentodes is negligible because of action of the screen grid and on top of this their usually low voltage gain less than 10.
EL34 has input capacitance 15pF, add another 10pF for wring for your calculation of upper bandwidth limit, no Miller. Gridstopper has a different effect once power tube gets pushed to overdrive and starts pulling grid current.
Aleksander Niemand
------------------------
Life's a party but you get invited only once...
affiliation:TUBEWONDER AMPS Zagray!-review
Perhaps you are a mutant with bat genes?
Will you hear this in a double blind test? Autosuggestion is a powerful trick a mind plays on itself.
I can hear the grass grow. In winter, under the snow.
Last edited by VacuumVoodoo on Thu Apr 12, 2012 8:23 am, edited 1 time in total.
Aleksander Niemand
------------------------
Life's a party but you get invited only once...
affiliation:TUBEWONDER AMPS Zagray!-review
VacuumVoodoo,
Please don't start flames posts, this is a clean thread in which I would like to understand the reason of this audible roll-off effect.
I'm sure you will ear the difference too, if you try.
Roberto, even though I often sound sarcastic I am serious. Don't take me wrong.
A general statement about grid stoppers on power tubes (a technical term is "anti parasitic resistor") without additional information doesn't say much.
What circuit, SE or PP, class, biasing scheme?
What else is there between power tube control grid and driver/PI? Shielded cable? If so, what kind, how long, where is it grounded?
How large is bias feed/ grid leak resistor?
What about OT matching to power tubes?
You may have a combination of OT properties (interwinding capacitance, mutual inductance) with tube plate/screen dynamics (screen grid resistor too small?) causing a resonance at high enough frequency to be affected by input HF filtering.
Such resonance will affect high end response of the amp. Low value of antiparasitics placing input HF cut-off close to the resonance could be one explanation of what you are experiencing. But riding the HF resonance on a power amp can be dangerous to the power tubes.
So you see, it's not just a question of that little resistor in series with control grid.
Aleksander Niemand
------------------------
Life's a party but you get invited only once...
affiliation:TUBEWONDER AMPS Zagray!-review
I've noticed in different amps with different configurations.
Latest one is a 460V B+, 4xEL34 Push Pull on a plexi OT, 47k (+25k lin pot) 4ohm tap NFB, standard plexi PI at 400V, 47n to 150k bias resistors, 100n 10k lin pot presence, 1M log 4n7 depth.
Two channel preamp, plexi-like clean and 4GS+CF lead.
4x1k screen resistors.
4x1k5 and 4x2k2 grid stoppers have more hi-mids than 4x5k6.
Similar sonic changes with a 500V B+ 4x6L6GC on IGPW SLO100 OT,
everything else similar except 4x470R screen resistors.
4x1k5 sound different than 4x2k2 or 4x2k7.
I've also asked to five people who don't know about effects on sound,
and they told me exactly what I ear.
VacuumVoodoo, please try and tell me what happens.
Could be useful to have a comparative frequency analysis on it.
Can anyone do something similar?
Keep in mind that I am new at this hobby.I love these post because they motivate me to do some reading and speed up the learning process.
I can't quote copyright material,but from what Richard K. shows that the gridstopper affects the gain at middle frequencies but does not have much effect at low or high frequencies if I understand correctly.
I suggest getting his books if you can.They are not that expensive.He also has some good stuff on this website:
Thank you Mario,
it affects mids but not highs? Do you know how?
I know Marshall used 220k also in some recent builds (I'm talking grid stopper on power tubes, not bias resistors).
In some other Marshalls there is a resistor in series to the 5k6 grid stoppers (it's not a mod, it's stock from Marshall). I've tried somethink similar at it seems something like a comb filter.
I read this quickly(too quickly) last night,and made a mistake.It is the grid leak that affects the gain at middle frequencies.Let me read some more and get back to you.
I agree you can hear the difference, but the miller cap values don't line up with that. The capacitance is fully modelled in the Spice valve models. Might be interesting to run some simulations and see how (if) the resistor values you've mentioned effect the freq response curves and try and work out where it's coming from. I'll see if I get a chance to do that over the next few days.
