Output Tube Grid Stoppers - Why 47k suggested for preventing parasitic oscillation ?

[walkman]

Hi All,

I have been searching for the source reference that states 47K should be used to help prevent parasitic oscillations of Grid Stopper for output tubes.

One reference i have found is Tremaine Audio Cyclopedia.

While smaller values can and are often used and their selection may be for the purpose of preventing blocking distortion in guitar amplifiers, why is 47k the value suggested to prevent parasitic oscillations?

I’m not questioning if 47k is the best value to use in any or all situations but curious to know why 47k?

Does it relate to cut off frequency or perhaps some other reason?
Was it simply a convention that became adopted? Or a maximum?
Or a good starting point, considering grid leak requirements of various tubes?

We see many amplifiers may have 1k5 or 5k6 values, and the statement that high mu output tubes may require a greater value, perhaps 10k or greater.

Does the amount of negative feedback influence what size the grid stopper should be or needs to be?

Do single ended designs need different considerations?

Hope this promotes some discussion. ;-)

[trobbins]

Tremaine's description is likely the 1969 2nd edition - perhaps quite a few revisions from the original 1959 edition.

My view is those comments are just used as a simple way to describe the application, and not a reasonable technical discussion or insight or explanation. Any amp can exhibit weaknesses, especially when related to large population manufacture and wanting to mitigate warranty and credibility events for a wide range of part and environment variations and aging. Increasing the grid stopper can certainly lower frequency and could be a panacea for known or inadvertent instability issues. And then there is the tonal and overdrive aspect - which likely supports why some manufacturers continue with a particular circuit value setup.

[eschertron]

And note the specific tube under discussion is a 5881. Is the tube type significant when choosing grid stoppers? I'd think so.

[walkman]

And note the specific tube under discussion is a 5881. Is the tube type significant when choosing grid stoppers? I'd think so.

Yes, that is part of what I’d like to know, as I have also seen 47k used in SE circuits for 6BM8.

Maybe there is a technical reason, however I suspect the value was selected either as general enough for a range of circumstances or that a specific usage has become generalised.

Note the images are from the third printing 1974, which I own. The schematic is an example and there is no statement in the text that the value of 47k is devised specifically for this circumstance / circuit

[Helmholtz]

I never read anything about an optimum grid stopper value in tube literature and manufacturer application circuits typically show lower values.
So I agree that the author likely just decribed a design example here.

Input capacitance of power pentodes/tetrodes is typically below 30pF. So 47k grid stoppers would still result in a cutoff frequency above 100kHz.

It should be observed that the grid stopper adds to total grid-to-cathode resistance which has an upper limit.

[pdf64]

Some discussion of grid stopper value selection in this thread https://music-electronics-forum.com/...ion#post696572

[Helmholtz]

If there's a desire to use a very large grid stopper in order to mitigate blocking distortion (I'm sure this wasn't a concern in the old days when power amps weren't meant to be overdriven), it could be split into 2 resistors, one before the bias feed/grid leak resistor, the other one going to the grid.
This is to avoid exceeding the Rgk limit.

Con: Will somewhat reduce available grid drive depending on resistor values.

[Helmholtz]

Regarding my post #5 I forgot to consider that the PI source resistance (typically 30k to 40k) adds to the grid stopper, so cutoff frequency will be lower but still well above the audio range.

BTW, while very large grids stoppers can mitigate blocking distortion, they tend to make grid clipping more aggressive (>hard clipping).

[alexradium]

there is a fairly good explanation in the second preamp book by Blencowe

[walkman]

there is a fairly good explanation in the second preamp book by Blencowe

See the discussion at https://www.18watt.com/viewtopic.php?t=17950 , MerlinB references Radio Engineering, Terman, 1937 p231-239.
Where there is some reasoning around frequency cut off and el84.

This again makes 47k grid stoppers seem excessively large.
Maybe this was simply a practice by one set of designers or other may be the case, however Tremaine has included this large value as an application to address parasitic oscillations.

Some other reading I have done appears to indicate that large amounts of negative feedback may provide an incentive for keeping grid stoppers small.
https://www.diyaudio.com/community/t...ations.329473/

And this thread where I wonder if the 47k may have been a value more typical for some preamp triodes that then fell into use for output tubes in some radio circuits where there may not be any or large amounts of negative feedback https://www.diyaudio.com/community/t...values.267847/

Are there any good examples where large grid stoppers have been used on output tubes specifically for tone-shaping ?

Large grid stopper to tackle blocking distortion may not be enough argues Rod Elliot https://sound-au.com/valves/analysis.html#s5

”Suggestions to 'cure' the problem include increasing the size of the grid stopper resistor (slows the process and reduces high frequency response, but not much else) and/or reducing the size of the coupling caps (reduces bass response). This changes nothing, but allows the blocking condition to recover faster. These techniques fail to prevent grid current and coupling cap charging, and do not address the reduction of plate and screen voltages under full load conditions, which requires that the bias voltage be reduced to maintain conduction through the transition from one valve to the other.”

Any good examples where large grid stoppers have been used to address blocking distortion in the original design (rather than as a fix or mod later on)?

I’m happy enough to accept small grid stoppers are the way to go for typical guitar output tubes and that some, such as el84 may benefit from up to 10k.

Is it a reasonable design approach to consider that there are opportunities for tone shaping and the managing of distortion characteristics other than through the use of large grid stoppers in the first instance?