Quick question about tubes.

[Kesh]

I'm dipping my toe in tubes, experimentally. Simple common cathode stage, buffered by op-amps. Low low voltage. Maybe DC coupled cathode follower after that.

My question is about the power supply. I'll run the op-amp circuitry on +/-15. All the tube circuitry I've seen runs on a single supply. Can I run the tubes between +/-15 and not involve the ground at all? What are the pitfalls of this, if any? I fairly regularly move between different DC bias in op-amp and transistor circuits with a cap and a bias resistor. Is the same possible here?

I'll be using 12AU7 or ECC88 as these are the best for such low voltages.

The PSU is unregulated, but seriously smoothed, spice tells me it's about 10uV of ripple.

Heater will have separate power.

[amptramp]

If you look at some older oscilloscope gain stages, you will find tubes being operated between a number of different supplies to facilitate DC coupling.  If you have the cathode with series resistor going to -15, return the grid resistor to -15 and use a coupling capacitor on the input.  The plate resistor will go to +15.  Note that the noise on the negative supply adds directly to the input signal, so the - supply at least has to be heavily filtered.  The + supply rejection depends on the circuit topology but is usually nearly direct i.e. a volt change on the supply causes a volt change in output, but some circuitry such as a differential pair will reduce this.

Note that ripple predictions from spice depend on having the filter capacitors characterized for ESR and inductance over the operating temperature range.  It may be cheaper and easier to regulate the power supplies than to obtain the data needed to design a filter that gets the required performance.

[Kesh]

Thanks. I suspected noise would be the issue.

I was hoping that as the cathode is affected almost with the grid from any ripple on -15, that a lot of this noise wouldn't appear at input. Especially with big enough Rk bypass caps.

But anyway, regulating such a small current won't be a problem.

[Ronan]

This is a great read if you haven't already seen it (the linked pdf):
http://www.valvewizard.co.uk/gainstage.html

[Kesh]

Yes, read that, and quite a bit of Merlin's other stuff. He also has a low voltage triode essay which I'm drawing a lot from, and a great thing on DC coupling a CC to a CF.