Threads for SS High Voltage Power Regulators?

[dougzilla]

Hi everyone--

I'm trying to design a regulated HV, Low-Z power supply (~450V, 300ma) for the P-P output section of what is basically like a Dynaco Stereo 70 (EL-34s, UltraLinear output transformers). I have already modeled a supply based on NS Application Note/ Linear Brief 47 (1980) in LTSpice, and it works, more-or-less, in simulation. The regulating part of the circuit uses an LM317 with a Darlington transistor pair to stand off the high voltage, and an LM329 precision zener IC (which I modeled separately as a subcircuit) to increase open loop gain in the LM317. Only difficulty is that the speed of regulation adjustment is very slow, such that if I overlay an AC sub-harmonic voltage on the HV off the power transformer, some of it rides through to the output. [I have constantly varying utility-supplied electrical power at my house, so am trying to simulate lousy line conditions.]

I realize I could design an OpAmp-driven pass transistor regulator (such as Joe Curcio designed and sells), but I am trying to configure mine with minimal parts count, such that I can mount the regulator and rectifier in a small circuit board vertically on top of the octal socket which currently holds my GZ34 rectifier tube. I am going to be using a Cree 1200V SiC Schottky rectifier (and I don't want to crank up the FRED vs Schottky diode debate!).

If there are threads in the DIY forums that deal with HV regulated power supplies, I would be grateful for pointers, or any other advice! [Note I have read Wavebourne's post and replies from Sep 2006, but it isn't really furthering the design question, just the pros/cons of SS vs Tubes.]

TIA,
Doug

[sawreyrw]

The power supply rejection of the circuit in LB 47 should be very good. What are you doing different from the circuits shown. If you post a schematic you may get additional feedback.

Rick

[tomchr]

I'm using NSC's circuit outlined in LB-47 as well. Some people refer to this as a Maida regulator after the author of that app note. I'm getting 375 V @ 150 mA easy.

To support higher output current, R3 will need to be decreased as the LM317 runs out of headroom once the voltage across R3 exceeds about 2.5 V. Also, double-check the value of R1 as the base current into the darlington pair will be higher than what was the case in the original regulator. Many BJT's - even darlingtons - don't have very high hfe/beta at low Ic. The component designators refer to fig 1 in the app note.

I doubt you'll be able to fit everything on an octal socket. You'll end up burning a fair amount of power in the pass transistor (Q2). You'll likely need to put a heat sink on it...

~Tom

[jackinnj]

If you have even a remote idea of the amount of energy you need to burn, put a resistor ahead of the Maida regulator and let it do most of the heavy lifting. Don't skimp on the heat sinking of the pass transistor. SY used a Maida on the Impasse preamp -- it is wonderfully quiet.

[star882]

Start out with an active PFC and modify the design as needed.

[dougzilla]

Appreciate the all the replies. I will post the circuit diagram as soon as I can figure out how to make LTSpice do what I want--I am running it in a Virtual PC window on a Mac, so everything is a bit more complicated. Probably not doable until the weekend. Is it OK to attach the .asc LTSpice file? The "printout" leaves something to be desired, resolution-wise.

My first shot at a SS supply was to put a 100 ohm 5W resistor behind the CREE diodes and go with no regulation. It worked fine until the resistor burned up, but I am hooked on the idea of low impedance and regulation now, having read more than is good for me, probably.

Glad to know that Maida's concepts have some validity 30 years down the road!

Cheers,
Doug

[sawreyrw]

If you want to post the LTspice .asc file use myfile.txt; change .asc to .txt. I think myfile.asc.txt will also work.

BTW, heating due to power loss can be significant for this type of circuit, so be aware of this in your design. Someone already mention that you need a good heat sink.

Rick

[SY]

Doug, a great alternative would be Jan Didden's T-regs, which were in Elektor and (I think) on his web site. Jan has done some terrific work over the decades on regulation, and his regs are high performance and bulletproof.

http://www.linearaudio.nl/t-reg-1.htm

[star882]

Originally Posted by sawreyrw

BTW, heating due to power loss can be significant for this type of circuit, so be aware of this in your design. Someone already mention that you need a good heat sink.

450V at 300mA is only 135W. Assuming a modest 85% efficiency (active PFCs easily do low to mid 90s), it would be 24W of heat. That's about as much heat as a Pentium 3.