pentode buffer

[Sheldon]

I need a buffer for a 7pin project. I want to DC couple it to the previous stage. I see it mentioned here and there, but not that many detailed examples. I cobbled this design together, as I have all the parts. Any real advantage of using a CCS with this buffer? If I do, I'll probably sub a more robust part for the bottom transistor (the NPN as voltage reference, I lifted from someone here - Ian, I think). Anything out of wack here?

Sheldon

[Eli Duttman]

Being constant current devices themselves, trying to CCS load a pentode is, at best, highly problematic. "Head butting" occurs. IMO, the way to improve on resistive loading of pentode CFs is the installation of a suitable inductor ($$).

You said this is for a 7 pin project. Think about paralleling the sections of an inexpensive 6J6 twin triode, as the CF, in combination with CCS loading. The cathodes are already tied together.

[Sheldon]

Originally Posted by Eli Duttman

Being constant current devices themselves, trying to CCS load a pentode is, at best, highly problematic. "Head butting" occurs. IMO, the way to improve on resistive loading of pentode CFs is the installation of a suitable inductor ($$).

I know you don't want to use a CCS plate load for a pentode, because the plate is isolated from grid1. Therefore the current through the plate won't change much with voltage changes at the plate, meaning the current source would be trying to drive a high impedance.

I don't see the same issue when using a current sink under a pentode follower. Connected at the cathode, the current sink is working into a low impedance. Isn't this the same as for a FET? You wouldn't put a current source to load the drain, but a sink under a follower is pretty common. Is there an analysis of this written up somewhere?

The twin triode is a nice idea, but I already have the 6AU6. Also, I wanted to do a pentode follower, cause I haven't tried one yet.


Sheldon

[SY]

A 6AU6 is terrific for the CCS (high plate resistance), but its transconductance is too low to make a really good CF.

[Sheldon]

Originally Posted by SY

A 6AU6 is terrific for the CCS (high plate resistance), but its transconductance is too low to make a really good CF.

I should be more specific. I'd be using it to drive a preamp, but not a lot of cable. My current one (DEQX) has 50k input impedance. Would that still be an issue? Or, where would you draw the line for this tube?

Sheldon

[SY]

50k and small capacitance shouldn't be a big deal. But you know me, always trying to get things optimal...

The real advantage to a pentode CF is lowered input capacitance (yes, there's slightly higher gain, but we're talkin' the difference between 0.92 and 0.97). I've gotten interested in trying them myself (see the SY/Pete Millett crossover thread), so will probably have some distortion spectra for the 6AU6 as a CF in the next week or two.

[Sheldon]

Originally Posted by SY

50k and small capacitance shouldn't be a big deal. But you know me, always trying to get things optimal...

The real advantage to a pentode CF is lowered input capacitance (yes, there's slightly higher gain, but we're talkin' the difference between 0.92 and 0.97). I've gotten interested in trying them myself (see the SY/Pete Millett crossover thread), so will probably have some distortion spectra for the 6AU6 as a CF in the next week or two.

Cool. Thanks,

Sheldon

[Sheldon]

Originally Posted by SY

I've gotten interested in trying them myself (see the SY/Pete Millett crossover thread), so will probably have some distortion spectra for the 6AU6 as a CF in the next week or two.

Ever try it?

Sheldon

[kevinkr]

I've been thinking a lot lately about using a zener tied to the cathode to bias the screen grid in a number of pentode based applications - you could use a ccs to drive the screen and zener - I have no idea how linear this would prove to be, but is something I am thinking of trying in a very high voltage error amplifier I am working on for my next amplifier project. Might be interesting and if it works reasonably well it would save the cost of a large high quality screen cap.. Just a thought.

[SY]

Originally Posted by Sheldon

Ever try it?

Sheldon

Yes. Not impressed. The problem is that you really don't get much better performance from it in the context of audio. The capacitances are reduced, which could be significant for an electrometer, but meaningless here. Distortion was low, but source impedance was still 1/gm, so not really any better than a triode. The only other performance advantage (in exchange for a higher parts count) was a slightly higher gain. But if you can hear the difference between a gain of 0.95 and one of 0.98, you're a better man than me.