Thanks Chris. Your explanations of what the various parts are doing have been helpful to me.
As you pointed out, feedback amplifiers also have output inductors to isolate their outputs from reactive loads. (More commonly found in solid-state amps than in tube amps.) I hadn't thought of that at all.
As I understand it, the build-out resistor is there to isolate the cathode follower's internal (100%) negative feedback loop from reactive (capacitative) loads.
However, that build-out resistor raises the output impedance of the cathode follower by the resistor's value.
Placing the small value inductor in parallel with that build-out resistor bypasses audio frequencies so the output impedance is only reduced by the impedance of the parallel LR circuit (which in this case looks like 4 ohms at 1kHz, 11 ohms at 20kHz).
Now I'm wondering if having a build-out resistor is really necessary. The NFB loop complicates things.
As usual, everything is a compromise between this and that. I'm learning over and over again that the art and science of good circuit design is to take into account the myriad unintended consequences from any part you add, any change in values to those parts, feedback applied, and so on. Add a negative feedback loop and suddenly there are more hidden variables to take into account.
I have a question about individual cathode build-out resistors on each paralleled triode in the cathode follower:
Wouldn't those individual build-out resistors add to the output impedance of the composite CF? They would be in parallel, so that would reduce their composite value, but in a headphone amplifier I'd want to keep the output impedance as low as possible. Am I correct in thinking that a separate 33 ohm cathode stopper resistor for each of three triodes would add 33/3 = 11 ohms to the output impedance of the amplifier?
Were I to build this circuit again, I'd put individual grid stoppers on each triode, and leave space for a couple of chassis-mount 245uF 250V motor run (metalized film) caps I picked up from surplus to be used as the output caps. I might skip the build-out resistor(s) entirely, and perhaps the output inductor with it. Or would that be a mistake?
Thanks again.
As you pointed out, feedback amplifiers also have output inductors to isolate their outputs from reactive loads. (More commonly found in solid-state amps than in tube amps.) I hadn't thought of that at all.
As I understand it, the build-out resistor is there to isolate the cathode follower's internal (100%) negative feedback loop from reactive (capacitative) loads.
However, that build-out resistor raises the output impedance of the cathode follower by the resistor's value.
Placing the small value inductor in parallel with that build-out resistor bypasses audio frequencies so the output impedance is only reduced by the impedance of the parallel LR circuit (which in this case looks like 4 ohms at 1kHz, 11 ohms at 20kHz).
Now I'm wondering if having a build-out resistor is really necessary. The NFB loop complicates things.
As usual, everything is a compromise between this and that. I'm learning over and over again that the art and science of good circuit design is to take into account the myriad unintended consequences from any part you add, any change in values to those parts, feedback applied, and so on. Add a negative feedback loop and suddenly there are more hidden variables to take into account.
I have a question about individual cathode build-out resistors on each paralleled triode in the cathode follower:
Wouldn't those individual build-out resistors add to the output impedance of the composite CF? They would be in parallel, so that would reduce their composite value, but in a headphone amplifier I'd want to keep the output impedance as low as possible. Am I correct in thinking that a separate 33 ohm cathode stopper resistor for each of three triodes would add 33/3 = 11 ohms to the output impedance of the amplifier?
Were I to build this circuit again, I'd put individual grid stoppers on each triode, and leave space for a couple of chassis-mount 245uF 250V motor run (metalized film) caps I picked up from surplus to be used as the output caps. I might skip the build-out resistor(s) entirely, and perhaps the output inductor with it. Or would that be a mistake?
Thanks again.
Yes, cathode resistors are in series with each valve's output Z, which is 1/gm, maybe on the order of 100R for little hot pants triodes like 6DJ8, raising the total to maybe 133R (from each valve). But this is before loop feedback, which reduces the output Z in the same proportion that it reduces gain. But of course you know that.
Don't really see why you'd want to remove parts that contribute to a good design. DIYaudio is a small club, and fashion trends abound, but we don't *have to* follow fashion. Too much of it is either pointless or just silly. I'll start foaming at the mouth if I get into Japanese hifi trends, so must stop here.
All good fortune,
Chris
Don't really see why you'd want to remove parts that contribute to a good design. DIYaudio is a small club, and fashion trends abound, but we don't *have to* follow fashion. Too much of it is either pointless or just silly. I'll start foaming at the mouth if I get into Japanese hifi trends, so must stop here.
All good fortune,
Chris
Last edited: