ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

previous: Bruce q{There would be no signal across t... -- 12/26/1999 7:17 PM

Re: cathode bias cap value

12/26/1999 8:46 PM
Ken Gilbert	Re: cathode bias cap value Sorry Bruce... T'was late in the day and my mind was fading fast. The gin didn't help. In any event, if the output stage is push pull, and running in class a, then it's essentially a well balanced differential amp. As one tube turns on, the other one turns off just as much. The cathodes (which are joined together) sit at pretty much the same DC potential throughout the waveform. Therefore the same voltage is developed across the shared cathode resistance. If the voltage across a fixed resistance does not change, then the circuit could be called "constant current." And due to the fact that there is fixed DC potential across that shared cathode resistance, then shorting out the ac transients (which is one way to look at the function of the bypass cap) is not needed, since there AREN'T any. Adding and removing the cap will not stabilize the cathode voltage--it's already stable. As a result, it won't increase the gain either, since there is no cathode degeneration going on--the cathode's voltage is fixed, and is not decreasing the gain of the stage. All of this gets neatly thrown out the window as soon as one tube starts cutting off. Then, there is not a smooth transistion of plate current from one tube to the other--the nice complimentary handoff of current is no longer so complimentary. As a result, you get a fluctuation of current being pulled through the shared cathode resistor. That fluctuation of current leads to a fluctuation in effective bias voltage which is developed across that resistor. Degeneration starts to take place as one tube cuts off, and the other turns on really hard, since the "on" tube cannot steal any more current from the "off" tube. Once it starts to do this, the cathode will be "pulled" up in voltage, causing the well known degeneration and subsequent loss of gain. At THIS point, adding a bypass cap would obviously reduce the problem, since the tube would be able to pull current out of the cap instead of relying on the other tube. The cap also tends to store the charge across the cathode resistor, keeping the "bias voltage" much more stable. Like any other stage, you can tailor the LF response of this diff-amp by adjusting the value of C. Lower C, faster time constant, higher break frequency, less LF. There was a very interesting paper that I printed out from the golden days that outlined the differences in THD and TID of differential output stages WRT bypass caps. I'll see if I can get my paws on it. I hope that explanation is a little less hard to understand! Ken

12/26/1999 8:46 PM

Ken Gilbert

Re: cathode bias cap value

Sorry Bruce... T'was late in the day and my mind was fading fast. The gin didn't help.

In any event, if the output stage is push pull, and running in class a, then it's essentially a well balanced differential amp. As one tube turns on, the other one turns off just as much. The cathodes (which are joined together) sit at pretty much the same DC potential throughout the waveform. Therefore the same voltage is developed across the shared cathode resistance.

If the voltage across a fixed resistance does not change, then the circuit could be called "constant current." And due to the fact that there is fixed DC potential across that shared cathode resistance, then shorting out the ac transients (which is one way to look at the function of the bypass cap) is not needed, since there AREN'T any. Adding and removing the cap will not stabilize the cathode voltage--it's already stable.

As a result, it won't increase the gain either, since there is no cathode degeneration going on--the cathode's voltage is fixed, and is not decreasing the gain of the stage.

All of this gets neatly thrown out the window as soon as one tube starts cutting off. Then, there is not a smooth transistion of plate current from one tube to the other--the nice complimentary handoff of current is no longer so complimentary. As a result, you get a fluctuation of current being pulled through the shared cathode resistor. That fluctuation of current leads to a fluctuation in effective bias voltage which is developed across that resistor. Degeneration starts to take place as one tube cuts off, and the other turns on really hard, since the "on" tube cannot steal any more current from the "off" tube. Once it starts to do this, the cathode will be "pulled" up in voltage, causing the well known degeneration and subsequent loss of gain.

At THIS point, adding a bypass cap would obviously reduce the problem, since the tube would be able to pull current out of the cap instead of relying on the other tube. The cap also tends to store the charge across the cathode resistor, keeping the "bias voltage" much more stable.

Like any other stage, you can tailor the LF response of this diff-amp by adjusting the value of C. Lower C, faster time constant, higher break frequency, less LF.

There was a very interesting paper that I printed out from the golden days that outlined the differences in THD and TID of differential output stages WRT bypass caps. I'll see if I can get my paws on it.

I hope that explanation is a little less hard to understand!

Ken

Replies:

Ricky Is there some sort of "rule of thum... -- 12/26/1999 9:38 PM