power transformer voltage regulation

[imaradiostar]

I've spent many hours this weekend plotting the voltage and current output of my collection of power transformers from various tube gear I've taken apart over the years. I'm trying to establish reasonable current ranges for each based on the voltage drop at different currents.

I understand that heat is often the limiting factor but I've been surprised at how little some of the transformers heat up even when they've been loaded past what I would call "normal" voltage regulation.

The Fender Twin Reverb and Vibroverb transformers I tested both show about 5-6% regulation from "resting" voltage and current to a load I would consider to be "full load." Is this the standard value for class AB1 guitar amps? Do Marshall or Vox parts perform similarly? I ask because I have none of my own to test.

Is there a traditionally accept percentage regulation I should shoot for? Should this value change with the operating class or intended use of the amplifier?

Should I have already read the answer in RDH4?

Seriously though- I'm looking more for real-world advice, not so much the textbook answers.

jamie

[Chuck H]

This may not qualify as part of the answer your looking for, but...

The voltage drop due to current is in the "black art" category for guitar amps. I would expect a min/max type figure for sound reproduction amps but with guitar amps the phenomenon can effect the sound in positive or negative ways depending on the rest of the circuit. It is possible for a PT to drop many volts under load and still handle it with grace. And some very good amps sound the way they do BECAUSE the power supply is strugling to provide volts. Other amps would be all wrong with too much sag. I don't see it as a black and white percentage for this reason. But... I can certainly understand wanting to narrow the usefulness of an unknown tranny... I've got a few myself.

Chuck

[Steve Conner]

It's a black art, period. Even the transformer maker can't always predict the regulation.

10% drop under load is a reasonable figure for medium sized transformers, but ones designed to feed tube rectifiers can sag much more, if they've deliberately undersized the wire to meet the minimum resistance requirement on the rectifier datasheet. The PT in my Toaster amp gave 550V B+ at idle, but when I changed the power tubes to 6550s they would drag it right down to 400V at full output.

Transformer makers measure regulation with a resistive load hooked to the AC output, and the 10% figure I quoted applies to that. With a rectifier and filter capacitor load, the regulation of the DC output will be worse than the AC value, by some amount that the maker also hates to predict.

[R.G.]

10% drop under load is a reasonable figure for medium sized transformers, but ones designed to feed tube rectifiers can sag much more, if they've deliberately undersized the wire to meet the minimum resistance requirement on the rectifier datasheet.
...
Transformer makers measure regulation with a resistive load hooked to the AC output, and the 10% figure I quoted applies to that. With a rectifier and filter capacitor load, the regulation of the DC output will be worse than the AC value, by some amount that the maker also hates to predict.

What you said, plus some more.

Part of the mystery is that rectifiers and filters cause the current to come out of the transformer in sharp pulses at the peak of the AC power wave. The pulses happen because the filter caps are holding up the DC voltage so the diodes don't let any current into the filter caps until the instantaneous voltage across the diodes is enough positive to let the diodes conduct. Then the current is limited by the wire resistance, transformer leakage inductance, diode resistance, and the value of the capacitors and how far down they've sagged since the last cycle.

100% of the DC power that will go out of the caps for the next half cycle is transferred into the cap during that pulse, so the pulse current is not only nonlinear (voltage difference between the incoming sine and the rising cap voltage across the nonlinear diode) but much higher than the average DC current.

The bigger the filter cap, the bigger the pulse. This is why tube rectifiers specify a maximum filter cap - it's to keep the pulse current below the value that damages the electron emitting surface of the cathode. The bigger the load the bigger the pulse peak, but the wider the pulse, too, so it's nonlinear, too.

Nonlinear nonlinearities on top of nonlinear processes make this an approximator's paradise.

[imaradiostar]

Thanks for the advice. I figured it was something of a dubious science but I thought I'd ask just to be sure. I spend so much mental energy calculating and thinking when I really ought to build a decent prototyping rig so I can build a few test designs and try them instead of needlessly guessing and calculating.

That said...I did generate a LOT of usable info for a lot of neat old power transformers and it's fun to try to match up different designs with the voltage and current a given transformer can generate.

jamie

[loudthud]

Duncan's PSU Designer II will do the math for you if you let it. Once you get the basic circuit configured, right click on the transformer and click edit. If you click on the ... next to voltage you can specify a regulation percentage. If you click on the ... next to ohms you can input data based on open circuit voltages and winding resistance. This is the key to getting an accurate result.

In the options menu, select the line frequency. Now select a delay, I usually use 1 second and the length of the simulation should be set to an integral number of power line cycles, 100mS is good. Now after the simulation runs, look at the numbers in the box at the lower left. Scroll over and in the I(T1) row you will find the RMS current.

[imaradiostar]

Thanks again. I've spent a lot of time with Duncan's ap and while it does get close enough for most stuff it doesn't account for transformer heating, something I've found to be the real limiting factor for some amps.

jamie