DCAudioDIY.com

Hello, all! I am building an amp using a DIYtube ST-70 driver board and tubes/transformers from a Heathkit AA-100 (got the idea here: http://www.diytube.com/phpBB2/viewtopic.php?t=799). I finally got everything powered up and discovered my B+ is 495 V when I would like it to be 430 V. I decided to attempt to fix this by adding 65V worth of Zener diodes between the center tap of the power xformer's secondary winding and ground (got the idea here: http://www.aikenamps.com/index.php/what-is-back-biasing). Here is my schematic:



Because I'm dumb, I didn't consider how this would affect my negative bias supply. It promptly blew the first filter capacitor (C13), I assume from overvoltage. Here is my plan to correct:

1. Make R12 0 ohms
2. Make R11 22.7K ohms (current is so low I think 1W will be ok)
3. Make C13 and C14 220V or more

Any comments on this plan? Is there anything I'm missing?

First, check to make sure that caps in bias supply are installed in proper orientation. The + terminal goes to ground here since it is a negative supply!

Are you measuring under full load?

Easiest... try a rectifier with a higher voltage drop, which is anything but a GZ34. This is one of the lowest equivalent resistance rectifiers.

You can also happily use a higher inductance choke with higher DCR. That will toss away a few more volts and get you in range.

I'd stick with the GZ34 and change the choke, personally.

If I change the choke won't the resulted be unregulated? As in, as the amp pull more current (when the volume is high, say) the B+ will sag?

I'm measuring under quiescent conditions (input grounded).

A higher drop rectifier is a great suggestion, thanks.

Yes, any series resistance will cause a corresponding voltage swing when current draw is varied...but I wouldn't worry about it.

A higher resistance rectifier will do the same thing, as will the DC resistance of the windings of the power transformer.

A 200 ohm choke will drop 25V @ 120mA or so. I have some small open frame 10H 120mA, 200 and something ohm chokes if you need one.

This is not a "regulated" supply per se. Most amps of this genre run happily on unregulated supplies...in fact, regulated supply is a relatively rare and ambitious enhancement for a PP stereo amp. And if you regulate the B+, you'd better regulate the bias supply also..and it has to be a regulator with flat impedance across the audio band or else it might sound nasty...a slipppery slope.

Filter caps, aka reservoir caps, will provide a bit of juice to counteract this voltage drop action during current peaks.

But as I was asking, did you measure the B+ under operating conditions with the bias set and so on? If the tubes were not drawing current, you don't know what the B+ under load will be.

Yes, Dyna used this value choke and the argument could be made that it enhances regulation, but one time a guy who worked for him told me D. Hafler could give you change for a penny. Inductance ain't free, ya know.

That's what those big "reservoir" caps at the output of the CLC filter are there for. They store enough energy to keep the PSU from sagging.
From that point in the power supply circuit, you do want to minimize resistance.

Ditto on measuring the voltage with some realistic load on the PSU....

I tested with the amp biased and the input grounded (20mA cathode current on all 4 output tubes), however it was probably worse than normal because my B+ was higher than I want.

If I was only talking about dropping 20V or so I would totally understand sticking a resistive load in there. But I need to drop 65V, which at my idle current of ~100mA is 650 ohms. If my current varies by 20% I'll drop 13V extra volts. Is that kindof sag OK? If it would work just as well to put a 650 ohm L or R right after the rectifier I'm all for it.

Rectifier voltage drops can't be purely resistive, can they? From what I can tell they're specified on the datasheets without any current dependency.

I don't intend to "regulate" the B+ really. I just meant if I use the zener method then B+ won't be as dependent on output current.

In general, +/- 10-15% is in the ballpark for tube amp circuits.

Max voltage and max dissipation on the tubes is your upper limit. Otherwise, you can get away with murder, at least until you figure out what is murder and what is not.

Filters and measuring instruments might require precise component values and voltages but a pedestrian vintage style tube amp has a lot of leeway.

I just SIM'd the PSU circuit, and indeed the 5R4 will drop about 50VDC.
I'm curious what your line voltage is, and the AC voltage on the secondary of the transformer. It's supposed to be 370-0-370.

Thanks! Source is 120 Vrms. The voltage across the secondary is 781 VRMS, so I think that means 390-0-390. I figured this kind of discrepancy wasn't a big deal. I've heard line voltage was lower when these transformers were made, like 110V.



Is a 650 ohm resistor after the rectifier "murder?"

You can get by with a much smaller resistor between the rectifier and the first capacitor, like <100 ohms in order to get the voltage drop you need.
Now you have an R-C-L-C filter, and the R-C part doesn't charge completely, so you get a lower final output. It's not a strict V=IR relationship in that part of the circuit. Try the Duncan PSUD program; it's easy-peasy to install and run, and you will see how it works.