Capacitor in parallel with cathode follower cathode resistor??

[leroy]

Hi

I've seen inside a few Friedman amps and noticed a capacitor in parallel with the 100k cathode follower cathode resistor (Pin 8 to ground).

I'm not sure what it's function is. I'm assuming it's some kind of low pass filter but as the output impedance of the follower is only a few K the cutoff would be in the 100's of Khz range??

I've looked at frequency charts after simulating the circuit it in LT Spice but it seems to make no appreciable difference. The values I've seen are between 500p and 4.7n.

Any ideas of what it's actual function might be??

Thanks
Leroy
 

[tubeswell]

Really need to see a schematic to comment.


You won't get much action out of a fully bypassed cathode load resistor on a CF stage.



Are you sure your'e not confusing a CF with something like the tremolo roach driver stage in classic Fender BF/SF circuits (which employ tremolo using a LDR/neon lamp roach that is driven by an inverting stage with a 10M plate resistor and a 100k cathode resistor that is bypassed)?

[jjasilli]

Ditto.  It does not seem possible that the verbal description of the circuit is correct.  Because the output of the CF is from the cathode, above the cathode resistor, a cap in parallel with that resistor would seem to bleed all signal to ground.

[shooter]

might serve the same purpose that it does in a gain-stage, stabilize a squealer 

[sluckey]

That small cap will decrease a bit of high freq signal. It won't totally kill the signal. You would need a much bigger cap to totally kill the signal.

[jjasilli]

Good point; I missed the statement of cap size. But credit again to tubeswell: a picture (schematic) is worth a thousand words.

[92Volts]

I think this would decrease input impedance (raise input capacitance) bringing it closer to a normal gainstage. But again, only for very high frequencies.

Aside from squealing/oscillation maybe this stops it from picking up radio signals?

In a one-off build it would probably fix those problems, but in a production amp there would be more logical (and cheaper) fixes like grid stopper resistors to do the same thing. Or fixing whatever layout/power supply problem causes it.

[jjasilli]

This is above my pay grade.  I don't know how to analyze the effect of this cap.  In a normal gain stage, a cathode bypass cap increases the gain of the frequencies it affects, by reducing the inherent negative feedback effect inside the tube.  But that's when output signal is sourced from the plate.  In a CF, signal is sourced form the cathode, so hi's will bleed to ground through that cap.  Whether or not hi's are also boosted inside the tube, or whether that even matters in a CF circuit, I have no idea. 

[shooter]

Aside from squealing/oscillation maybe this stops it from picking up radio signals?

RF, whatever the source, typically causes the HF oscillations, in my old life I believe we called them snubbed caps.  I would suspect a 5cent cap is a cheaper fix than paying an engineer to actually fix it

[PRR]

Welcome Leroy.

Details matter.

If it was a "good" cathode follower with small signals, the effect is above maybe 33KHz (for 5nFd) , as you say.

But at the 5F6a tonestack driver we got very large signals. When the CF cuts-off, the discharge path for the cap is 100K, and sags along a 300Hz line.

Fender biased his CF near grid-current. Some other guys bias on the edge of grid current. As the CF saturates it becomes a diode (much power comes from the plate before it). Its cathode impedance may be 20K, not 1K.

Where it transitions from "good", to "kinked", to "bent", will depend a lot on tube variations.

So it is a mess. At the current price of few-nFd caps, experimentation may be the least-insane path.

[leroy]

Thanks for the responses all.
I wasn't sure about posting a schematic as it's a current production amp. I've cut out the section i'm referring to.



Thanks
Leroy

[tubeswell]

Okay - that's a partial bypass for shunting HF to ground.

[jojokeo]

Okay - that's a partial bypass for shunting HF to ground.

Yep, and schem was just what I was thinking all along. I'd expect tubenit to do something like this utilizing more "smoothing caps" ideas.

To me it's extra parts that may not be needed if he didn't bypass the 820r resistor. Many times when that's done it increases gain but also gives a lot of high end harshness along with it and the gain increase isn't all that much or needed. That's why many builds don't bypass that resistor.

It's one of those different strokes for different folks kind of things. But I usually will put this on a switch. Sometimes w/ certain guitars or type of music where lower gain settings are used it enhances things but if dialing things up and then switching this in it gets to be too much and not very musical, and causes ear fatigue pretty quickly.

[joesatch]

old thread but was helpful. This is the BE50. I removed the caps off the 100k cathode follower resistor and it's as if someone took the pillow off the speaker. i put a 300pf instead of the switchable 1000pf / 2.2n

[jojokeo]

old thread but was helpful. This is the BE50. I removed the caps off the 100k cathode follower resistor and it's as if someone took the pillow off the speaker. i put a 300pf instead of the switchable 1000pf / 2.2n

Now that you've removed the high end losses, experiment lifting the 47uF completely and then try subbing a 10uF then a 1uF. You'll see a bit of a difference with each. The 1uF will keep the high end gain and maybe tighten up the low end (not boosting the lower frequencies).

[tmknight]

...or consider this