Power Amp Grid Leak Resistors

[mwelch55]

I have a couple of questions about the Power Amp shown in this schematic.

1. What is the effect of reducing the Grid Leak resistors from 220K to 100K?  Does it take more output from the PI to drive the power tube?  How does it affect tone?  I have seen this in other schematics on the Internet.
2. Does the 47pF snubbing cap roll off the highs from the PI?  Does increasing the value to 100pF roll off more highs?

[mwelch55]

Sorry for the duplication.  I already asked the first question.

https://el34world.com/Forum/index.php?topic=22477.0

[jjasilli]

1. What is the effect of reducing the Grid Leak resistors from 220K to 100K?  Does it take more output from the PI to drive the power tube? 
Yes.  The signal grid (G1) needs to be connected to - "referenced" - to ground for the tube to maintain bias.  The grid leak resistor is used to lift signal from ground, while maintaining the G1 ground reference.  The larger the value of the grid leak resistor, the more signal voltage reaches G1.  (For fixed bias, it also lifts bias voltage from ground.)

How does it affect tone?
A shunt resistor is a hi pass filter.  Some people think that lower value grid leak resistors sound better.  YMMV.

2. Does the 47pF snubbing cap roll off the highs from the PI? 
Yes.  A shunt capacitor is a lo pass filter.  I.e., it attenuates hi's.

Does increasing the value to 100pF roll off more highs?
No, it rolls off a larger bandwidth of frequencies into the lower range.

The purpose of this shunt cap is to keep the amp from oscillating.  It is not a tone control function.  This is a Gremlin thing.  Given several identical amps, some may oscillate, some not.  If your amp doesn't oscillate you don't need the snubber.  Manufacturers often use them as a prophylactic measure so they don't have to test every amp, or get numerous returns and complaints.

[PRR]

You can do math. What does that tell you?

None of your changes are likely to make trouble.

I suggest that 49 cents of parts will tell you how it "sounds", which is what you want to know. Tack two more 220L across the existing 220k; that's 100K near-enuff. Tack another 47p across the cap.

We "could" Do Math. But what would "69Hz" or "37V" tell you? Blow the parts-money, heat up the iron, and beat it.

[shooter]

while you're collecting parts, hit up your buddies for any speakers they can loan you, you might find "the tone" from the pile

[pdf64]

...2. Does the 47pF snubbing cap roll off the highs from the PI? 
Yes.  A shunt capacitor is a lo pass filter.  I.e., it attenuates hi's.

Does increasing the value to 100pF roll off more highs?
No, it rolls off a larger bandwidth of frequencies into the lower range...

That last response above seems somewhat awkwardly worded

As the value of the plate-to-plate cap in increased, the corner frequency of the low pass filter it creates will reduce.
So the signal bandwidth will likewise reduce, ie more high frequencies will be rolled off.
But yes, I suspect the designer intended that the main purpose of the filter was to increase the margin of stability, rather than make the amp less 'bright'.
The corner frequency of the filter will be set by the 'C' value and the sum of the LTP plate impedances, which, given the series NFB of an LTP, I think will approximate to Rp1//Rg1 +Rp2//Rg2, so 88k//220k + 100k//220k = ~60k + ~69k = ~129k
So corner freq of ~26kHz with 47pF, ~12kHz with 100pF

[jjasilli]

You are supporting the point I made.  Assuming your math is correct (though the general idea is correct): 


*  a 47pF cap freely passes all frequencies above ~26kHz
*  a 100pF cap freely passes all frequencies above ~12kHz


So, the 100pF cap passes all the frequencies freely passed by the 47pF cap, plus more, down to the lower frequency:  the bandwidth of frequencies freely passed by the the 100pF cap is 14kHz larger, by reason of including lower frequencies. 


The frequencies passed by the shunt cap come from both sides of the PI.  Hence they cancel out, because they are out of phase.  The bandwidth cancelled by the 100pF cap is larger, ~12kHz -to- infinity, than that of the 47pF cap, ~26kHz -to- infinity.

[mwelch55]

You are supporting the point I made.  Assuming your math is correct (though the general idea is correct): 


*  a 47pF cap freely passes all frequencies above ~26kHz
*  a 100pF cap freely passes all frequencies above ~12kHz


So, the 100pF cap passes all the frequencies freely passed by the 47pF cap, plus more, down to the lower frequency:  the bandwidth of frequencies freely passed by the the 100pF cap is 14kHz larger, by reason of including lower frequencies. 


The frequencies passed by the shunt cap come from both sides of the PI.  Hence they cancel out, because they are out of phase.  The bandwidth cancelled by the 100pF cap is larger, ~12kHz -to- infinity, than that of the 47pF cap, ~26kHz -to- infinity.

In a previous post, you said that the 47pF snubbing cap is a low-pass filter.

If a 47pF cap freely passes all frequencies above ~26kHz, wouldn't that be a high-pass filter?

[sluckey]

If a 47pF cap freely passes all frequencies above ~26kHz, wouldn't that be a high-pass filter?

Depends on how it is placed in the circuit. If the cap is put in series with the signal path such that the signal must pass through the cap in order to get from PI plates to PA grids, then the cap is absolutely a high pass filter. But that's not the case with this circuit. The cap is connected between the two out of phase signals and will therefore shunt the high frequencies from one side to the other. This cancels/kills the shunted high frequencies while having no effect on the lower frequencies that we can hear. Thus it is a low pass filter or if you rather say, a high freq shunt.

It may be easier to get it if you place that 47pF from plate to ground. Do so for the other side too. Now it's much easier (for me at least) to visualize this as a high freq shunt (or low pass) filter.

[jjasilli]

In a previous post, you said that the 47pF snubbing cap is a low-pass filter.

Yes, what sluckey said.  A cap always passes hi frequency and blocks (attenuates) lows.  So, it can be used as a filter in 2 different ways:


*  a series cap is a hi pass filter.  It blocks lo's in the signal path and passes hi's.
*  a shunt cap is a low pass filter.  Typically it bleeds hi's to ground, while keeping lo's in the signal path.  (Think: treble bleed tone control). 


(BTW, a resistor or an inductor do exactly the opposite of what a cap does.)


A "shunt" is something not in series, thereby diverting current or voltage elsewhere. 


The snubber is a form of shunt cap.  Though it doesn't bleed treble to ground, it causes out-of-phase hi's to divert from their respective signal paths and to meet & greet each other; whereupon they promptly cancel each other out.