How do I design a pentode constant current source plate load?

[Bitrex]

Hello everyone. Long time lurker, first time poster.

I am trying to design an amplifier with a 6Au6 pentode operating as a constant current source plate load for a medium mu triode. The equation I'm attempting to apply is in the back of Rozenblit's "Beginners Guide to Tube Audio Design":

ip = Epp - ep +mu*E)/rp+(mu+1)*R

where ip is the current through the CCS and triode, Epp is the supply voltage, E is the voltage differential between the triode's plate and the CCS's grid, mu is the gain of the CCS tube, rp is the dynamic plate resistance of the CCS tube, and R is the value of the CCS's cathode resistor.

This equation works fine when I use it for a triode loaded by a constant current source triode, but when I apply it to a pentode as the CCS (estimating at my selected operating point the 6AU6 has an rp of about 1Meg and a mu of 4,000) the equation breaks down and I start getting impossible values for R. I'm wondering if anyone could help point out where I'm going wrong? Or if there's a better way to analyze things when using a pentode as the CCS?

[Yvesm]

Hi Bitrex !

A 6AU6 is, in itself, a very good current source for a wide range of plate voltage.

In the attached graph you can see that to obtain, says 6mA you just need to bias the grid at -2 Volts and fix the screen at +150V, both relative to the cathode.
It's no so easy at it seems at first glance !

This done, the plate current will remain stable for plate voltages ranging from less than 100 to more than 400V relative to the cathode.

Beware, however, that the cathode current is the sum of the plate and the screen current wich is NOT constant.
This is usually a problem when the CCS is used to "source" current in the anode of another tube but cause no problem when used to "sink" current from a cathode.

Yves.

[oldeurope]

Hi Bitrex and Yvesm,

r ≈ s x Rg2 x Rk and i = u / r

r = dynamic resistance of the current source
s = transconductance of the pentode
Rg2 = screen grid feed resistor
Rk = cathode resistor between g1 and cathode of the pentode.

Kind regards,
Darius

[reinhard]

Pentode current sources have a high dynamic resistance, but tend to be noisy.

[reinhard]

For even higher dynamic restistance use a higher value Rk and a voltage divider for G1,

Reinhard

[kenpeter]

Consider a Cascode of Triodes instead of a Pentode? Need not
have any screen forward biased. Same current, source or sink...

Of course, this is also an application where a depletion MOSFET
solves a lot of problems, needs no elevated heater, etc...

I have no idea which solution of these three might be least noisy.

[Miles Prower]

Originally posted by Bitrex
I am trying to design an amplifier with a 6Au6 pentode operating as a constant current source plate load for a medium mu triode. The equation I'm attempting to apply is in the back of Rozenblit's "Beginners Guide to Tube Audio Design":

My first preference would be to solid state that. If you still want to go with a pentode instead, I'd say fuggedabouddatequation. You'll get more accurate results from the plate characteristics.

I have a "white paper" here that describes how to make pentode based CCS's, but I can't find the thing on-line. If interested, e-mail me and I'll send it off as an attachment.

[Bandersnatch]

hey-Hey!!!,
I'd suggest a triode-pentode cascode. Triode on the bottom, pentode on top. I've done this with 12BY7/12HL7 with excellent results. Triode strap the bottom one as triode( or consider something like a triode like the 6C45 or strapped EF184). You'll need to create floating voltage references to allow for largest-possible R-set and to 'space' the pentode above the triode. See the Gary Pimm enhancement-mode MOSFET ccs biasing methods for ideas.
cheers,
Douglas

[Bitrex]

Thanks for all the helpful responses!
I have one further question - if I wanted to take the output of this circuit and couple it to a small power triode, for example, would it be preferable to take the output at the cathode of the pentode, or the plate of the triode? I know that the cathode of the pentode would have a lower output impedance, but it's sitting at such a high voltage that it would prevent direct coupling of the two stages. But would the plate of the triode have a low enough output impedance to drive the miller effect capacitance of a small power triode?

[Bandersnatch]

Take the mu-follower output at the cathode! use a bigger coupling cap.