Tube shunt regulator for phono amp

[velvetsunrise]

Hi, i am considering building a phono amp with the following topology.
2 Parallel ECC88s sections, cascoded with a further ECC88, Passive RIAA eq, and a further parallel ECC88 common cathode output stage, or possibly replace the lower part of the input cascode with a Jfet.

I was planning to use a Tube CCS fed shunt regulator, of the simple type, with a single triode, with zener reference in the cathode and error voltage applied to the grid.

I understand a Mosfet would give higher performance in this configeration due to its higher transconductance.

I would however, prefer to use a tube in this position, would a cascoded triode with its higher gain be an improvement over a single triode?

Or, does it need a high gain opamp to control the shunt element? in view of the poor power supply rejection of the input stage cascode.

[astouffer]

Or, does it need a high gain opamp to control the shunt element? in view of the poor power supply rejection of the input stage cascode.

A small signal pentode would work fine. A 6AU6, 6AK5, 6BA6 are all suitable for error amps. The latest AudioXpress has an article about tube regulators and using a triode to isolate the voltage reference.

Mods, what do you think of a sticky thread about tube regulators? Its becoming a very common topic now.

[Zen Mod]

I used plain 0A2 as shunt reg previously .......
even for first stage , with ECC83

ages ago ......

[Miles Prower]

Steve Bench has a series of articles about hollow state voltage regulators. Shunt Regulators

[velvetsunrise]

Thanks for you comments, this is a sketch off the regulator i was considering using for my MC phono preamp. I have used the constant current source section before, feeding gas regulators in a line amp, ripple was very low.

I have read somewhere that loading a pentode with a current source is not a good idea, this is what worried me about the proposed regulator.

If i were to remove the capacitor from the upper triode grid, as this voltage is from the output of the regulator rather than the raw supply, would this reduce the output impedance of the regulator as a whole, or would it be worse due to the reduced open loop voltage gain?

[kevinkr]

One potentially large problem with the cascode in this configuration is its high rp, which will be in the same territory as a pentode of comparable overall transconductance (or worse) - makes a great error amplifier for a series pass topology but is not so good on its own.

I have forgotten to include that cap in some of my error amplifiers from time to time and in one case it reduced the ac open loop gain by about 30dB or so - quite obvious something was wrong, but it took a little while for me to figure out what was wrong.

I have done this with medium mu triodes (single, NOT cascode) and found the performance a bit mediocre.. Very high transconductance triodes might however do a creditable job.. I'm thinking triode connected D3A (mu ~77) or the 5842 (mu ~43) both have rp under 2K or so at 20mA and could make really good shunt regulators I suspect. Another possible candidate is the 5687, but the mu is on the low side for this application - still it is much cheaper than the other two listed and will probably work well. Drop about half the voltage across the pass element and the rest across the zener in order to avoid wasting too much loop gain. (Don't drop more than 150 - 200V across the tube, and observe max cathode current and plate dissipation ratings.)

There are a lot of odd and non-intuitive interactions because of the moderate open loop gain, so to get the best performance some resistor value trimming is inevitable. Put a cap across rfb to make the closed loop ac gain unity - this should result in the lowest output impedance.

Use 5W or better rated zeners, they tend to be a bit quieter due to lower current density in the die, but use within the recommended zener current range and bypass with a couple of caps. (High voltage zener dynamic resistance may not be as low as you would like.)