Ultra-low Noise Valve Regulated PSU. How to?
 

Wanna build a MM amplifier? What troulbes me now is that most sources indicate that a valve regulated PSU will not perform as well as a solid state regulated PSU, in particular the noise level.

But in history, valve regulated PSU has been extensively used in various applications, including those high-precision instrumentation that demends ultra-low noise level?

So is there any professional article that covers this issue, i.e. how to build ultra-low noise valve regulated PSU, the know-hows, the noteworthy technical parameters, and the enabiling methods, such as tube selection, sheilding, chasis layout, wiring, etc.

Can anybody offer help to provide this kind of references? Thanks in advance.

The HP 400D audio volt meter and the HP435C microwave volt meter have a very well designed tube regulated power supply built in.

Check them out.


Johnny

Janneman has the solution.

In those applications, the signal circuits are designed to have high power supply rejection. Did you know that the original application of the topology that John Broskie markets as Aikido was for DC amplifiers, to accommodate poor heater supply regulation?

Jan's design is excellent and I'm using it in my next power amp, but if you're a tube purist, it won't do- the error amplifiers are solid state. I think Kevin Kennedy has posted his all-tube reg; I'd look at that. Also, since old Tektronix schematics are easy to come by, you might copy some of their old regs. But SS is still the performance champion for low noise, low impedance DC.

Thanks. Tubes have inherent noises that can hardly be overcome, but my goal is very simple: to find a tube regulated PSU as slient as possible for my preamplifier. It seems that the PSU used in HP instrumentation will satisfy this need.

The HP stuff is well designed, mine was designed expressly for the criteria important in low noise line and phono stage applications. A lot of them have been built by me and others..

You can also take power supply schematic from L1-3 or L3-3 Russian vacuum tube tester.

I don't think it beats this:

jan didden

If you want low noise at audio frequencies, why don't you just filter the supply? If you need an accurate DC voltage, make a regulator and put a filter between the regulator and the circuit that needs the low-noise supply.

Most voltage references are pretty noisy: that holds for valve solutions like a neon discharge reference as well as for solid-state versions like a "zener" diode operating in the avalanche region or a band gap reference (noisy because a small VBE difference is multiplied by a large factor). If you want really low noise, usually filtering either the reference or (preferably) the supply voltage is the most practical solution.