MOSFET Dissipation (for Power Scaling / VVR)

[uneumann]

Hmmm. Just doing some speculation here.

I'm going to talk about the phase control idea, as it's fairly heretical. ...
A better way might be to put a very capable MOSFET in series between the rectifiers' (+) output and the first filter cap. A little timing and control/delay circuitry and you can turn on the MOSFET at some point later than the peak of the AC mains wave, and the output voltage goes down proportionately, including all the way to zero if you want, and anywhere in between. The MOSFET stays relatively cool, as it's only on for a short time, and is "saturated" during that time, at high current but low voltage across it, not high current and high voltage like the linear dropping approach.

I like this idea... Heretics are often interesting. It seems to me you even preserve the PS sag and it's time constants this way since you're effectively just reducing the AC output of the transformer. With linear power scaling that's not usually the case since the linear device is acting like an inline resistor somewhere - and that changes the local filter time constants. (A nod to the recent sag thread that got me thinking about it.)

[Gregg]

The solution to the heat dissipation has been around for quite a long time and it consists of not dropping the entire voltage but the screen voltage only. It doesn't get more simple than that. There's even a patent which expired some time ago (for more details check out the corresponding thread). I did that several times and I can tell you from experience that the heat sink doesn't even get hot.

FWIW I'd make *all* tube amps available with a 1/2 voltage tap switch, instant power taming and to boot, lowering heat and prolonging life.

It's difficult to invent the wheel these days. Please check the SUNN T reissue schematic where a Club/Arena switch is implemented and which effectively drops the voltage in half from 100W to 25W.

[g1]

Are you saying starving the screens will give the same sound as reduced plate voltage?
I assumed it does not and that is why it is not used.

I like the Fender "The Twin" power reduction scheme. Take off of the lower half of the totem pole and compensate bias. Done.

[octal]

A better way might be to put a very capable MOSFET in series between the rectifiers' (+) output and the first filter cap. A little timing and control/delay circuitry and you can turn on the MOSFET at some point later than the peak of the AC mains wave, and the output voltage goes down proportionately, including all the way to zero if you want, and anywhere in between. The MOSFET stays relatively cool, as it's only on for a short time, and is "saturated" during that time, at high current but low voltage across it, not high current and high voltage like the linear dropping approach

I wonder if you could use an off-the-shelf zero crossing detector chip and then add a delay/lagging circuit to that.

[Gregg]

I haven't done extensive A/B tests to determine how it sounds compared to reduced plate voltage but it is used and there's an expired patent detailing it.
Most recently it was used in Marshall's AFD100 and YJM100 models where it was controlled by a uCU. Using the same driver circuits that don't put HV on pots (or even more simple than those but I haven't seen any so far) it can be done with a 3 gang pot (you can use two stereo pots back to back connected with a hex axis cut from a hex key for example which means you need to get that type of pots). However due to the fact that the gangs' resistances don't match perfectly you can get some deviations from the best operating points at lower voltages which is the most interesting region for apartment/bedroom players. If you're doing it for for non-commercial purposes you can go further getting several stereo pots and select which sections match the closest then disassemble the pots and assemble the matching sections together. 3rd section is for the feedback and is not critical.

Another good idea is to use a 4-6 position 3-4 pole rotary switch and set the screen voltage/bias/power levels you want.
Or just use a 3 position toggle switch for 1, 10 and 100W for example.

US4286492 1979 screen powerscaling.pdf

[R.G.]

I wonder if you could use an off-the-shelf zero crossing detector chip and then add a delay/lagging circuit to that.

Almost certainly yes. Although it might be about as easy to do it with a MOSFET driver chip or even a PIC.

[R.G.]

I think that this would require very little embellishment to work. Mostly the transistor driver output to the thyristor gate would have to pull on the LED of an optocoupled MOSFET driver and the MOSFET would float on the rectified high voltage before the first filter cap. Ought to work.


555-timer-triggers-phase-control-circuit

[trobbins]

Most commercial designs would not aim for a Tj max of much above 100-110C, and certainly never use the datasheet Tj max in the standard thermal equation linking max ambient air temp and heatsink parameters. That of course pushes the heatsink to a larger beast, or makes you think of other means, such as paralleling.

[Gregg]

I think that this would require very little embellishment to work. Mostly the transistor driver output to the thyristor gate would have to pull on the LED of an optocoupled MOSFET driver and the MOSFET would float on the rectified high voltage before the first filter cap. Ought to work.

Do you mean something similar to the circuit in this patent:

US20120242412A1.pdf

[J M Fahey]

It's difficult to invent the wheel these days. Please check ........

My post was not about *inventing* the wheel but about *using* it, everywhere.
In fact in the very next line I mentioned 2 examples of amps already doing that

[Gregg]

It wasn't personal but a more general statement.
Also the Sunn T reissue approach handles bot the plate and bias voltage not affecting the preamp voltage (well, only a couple of volts due to the lower plate current draw).

[R.G.]

Do you mean something similar to the circuit in this patent:

No, but that's a very entertaining patent.

The patent office apparently can't get anyone with any technical smarts or experience to work as examiners. Maybe there's too much money in real technical work. Don't know. But they let any kind of mumbo-jumbo prior art through if it sounds plausible. I wonder that they didn't wonder at the "Pet the DOG" step in the flow charts. Maybe they think it doesn't matter, since a patent is only worth what you can spend on it in legal fees to defend it.

But I digress. No, that's not particularly close except for there being a uC, zero crossing detection and some kind of modulation of something going on.

In practice, phase controllers for peak voltage control (like in a rectifier/filter setup in an amp) all work by delaying the rectifier turn-on later than the peak of the AC wave. The peak of the AC wave gives you the max voltage possible, and delaying it past that point lowers the voltage by putting the turn-on on the dropping slope of the decreasing side of the AC wave. It's non-linear, but not too bad. And the control is easy.

All that's needed is a fixed time delay starting at the zero crossing and being no shorter than the time to the half-wave peak, no longer than the full half-wave, with a pot to control it. Turn the MOSFET (or SCR, or whatever) on after the delay.

This thing qualifies in my mind as invalid for patenting because it's, as the patent rules say "obvious to one skilled in the art", and almost certainly prior art as well.

But then I'm not a patent examiner. Thank God.

[Dave H]

But then I'm not a patent examiner. Thank God.

It was good enough for Einstein

[R.G.]

It was good enough for Einstein

It was. But he was a bigger man than me. He could do it, likely on auto-pilot, while mentally figuring out the universe.

I'd actually think about what I was doing and go stark screaming mad. Or be fired for writing evaluations like "This is so obvious that it screams, you insufferable twit!"

[Leo_Gnardo]

<Dave H: It was good enough for Einstein >

Then Einstein and his friend Leo Szilard got a US patent on a refrigerator @ 1930. Back then they had to make a working model. Never did go into production. Can you imagine if it was a roaring success and their attention was diverted to inventing household items.