pfc for tube amps

[woodyc]

It turns out nonlinear current draw is not just a curious fact about tube amps. Simple rectified cap-input dc supplies may actually be illegal in the future. According to EN61000-3-2, it appears they were technically outlawed for any consumer 75W+ audio gear in Europe in 2001.

I haven't found a practical retrofit for existing tube amps.

For new builds, the simplest solution might be replacing the HV supply with a pfc boost pre-regulator. These are an odd sort of inductor/switch (~100kHz) arrangement for smoothing out current draw. I think they are "pre" because they are intended to feed a second more conventional switching supply -- but they "boost" enough that they could actually be used stand-alone for tube amp HVDC.

Any thoughts or better/cheaper ideas would be welcome.

[Steve Conner]

A boost PFC is all very well, but it doesn't provide isolation from the line. We would be back to the days of the All American Five. Doubly so because it doesn't provide heater voltage either. What you need is a complete switchmode power supply with a "PFC front end".

I think some of the Crate Blue Voodoo series used a SMPS, and it worked fine. If that European standard was a problem, maybe someone (me? ) would profit by bringing out the SMPS as a module.

But I can't see that it is a problem: plenty of tube amps drawing over 75W have been sold in Europe since 2001. Also, a tube amp power supply may actually meet the spec for harmonic currents. The leakage reactance of the transformer helps filter them out.

In practice it seems tempting for many manufacturers to replace the power tubes with a chipamp and the preamp with a DSP at the same time as they "upgrade" the power supply.

[R.G.]

It turns out nonlinear current draw is not just a curious fact about tube amps. Simple rectified cap-input dc supplies may actually be illegal in the future. According to EN61000-3-2, it appears they were technically outlawed for any consumer 75W+ audio gear in Europe in 2001.

In a fully mature society, everything not forbidden is mandatory.

It appears to be irresistible for governments to make more things illegal. Just recently those include incandescent light bulbs and non-switching wall warts, both on the theory that even if you can afford the price of the energy to run them, you should be forbidden from doing so because it's not good for the diffuse "everyone" for you to be permitted to do so.

It's interesting. I wonder when someone will notice that it is vastly more efficient in terms of energy used per watt of electrical power to use not only solid state power amplifiers, but Class D switching power amplifiers, and to outlaw first the production of new tube amps, then the manufacture of new tubes, then the *possession* of tubes and amplifiers using them.

But actually, PFC modules exist for most switching power supply technologies. I don't know how well they'd work in front of a tube-amp power trannie. I guess I'll go look some day.

[woodyc]

A boost PFC is all very well, but it doesn't provide isolation from the line. We would be back to the days of the All American Five. Doubly so because it doesn't provide heater voltage either. What you need is a complete switchmode power supply with a "PFC front end".

I think some of the Crate Blue Voodoo series used a SMPS, and it worked fine. If that European standard was a problem, maybe someone (me? ) would profit by bringing out the SMPS as a module.

But I can't see that it is a problem: plenty of tube amps drawing over 75W have been sold in Europe since 2001. Also, a tube amp power supply may actually meet the spec for harmonic currents. The leakage reactance of the transformer helps filter them out.

In practice it seems tempting for many manufacturers to replace the power tubes with a chipamp and the preamp with a DSP at the same time as they "upgrade" the power supply.

Steve, I asked you a really dumb question but luckily got back before you, whew. Yeah the isloation is a problem. Maybe an interim solution (while we are waiting for your board) is a transformer with a 1:1 winding that feeds the boost pre-regulator, and a 6.3V winding for the heaters.

There is a boost/buck mode where you get isolation, but I've only seen academic papers on it where they built big unweildy things. Are any of those kind available as evaluation boards? Type (c) in pic below from ST application note AN510/0894:



RG, I don't see this kind of regulation as annoying at all. The problem is that I never knew about it. There are really good reasons to use tubes, and I'd fight for that right -- but line harmonics are stupid, particularly in an audio environment.

[trobbins]

A practical smps would use a pfc front end (which generates an un-isolated HV DC rail with some twice mains ripple voltage on it). Two separate dc/dc stages would then probably follow: one for regulated B+ and another for regulated 6 or 12VDC. These dc/dc stages would each include transformer isolation, and would regulate out any AC mains ripple to negligible levels. The complexity and effort to make such smps has certainly diminished over the last few decades, and the practicality of packaging means that such a supply would probably be at least a few hundred watt rated. The B+ output moves such a product in to a minefield of regulations with respect to how the total system meets safety compliance standards - and so it may not be easy to make a product that can be purchased 'separately' from the amp itself.

[diagrammatiks]

pfc on a good low leakage/low loss power transformer in a linear supply is much better then the pfc of a smps.

[loudthud]

I see the need for a product with a PFC front end and conventional mains output. You can just plug any old amp in and the power company doesn't know the difference. Would the poor power factor of the tube amp mean a higher electric bill?

[Steve Conner]

pfc on a good low leakage/low loss power transformer in a linear supply is much better then the pfc of a smps.

The problem is that the old-fashioned form of passive PFC mainly compensates the displacement power factor. And as woodyc showed in another thread, the low power factor of a linear power supply is caused by harmonics, not displacement.

So, if you used a big enough PFC capacitor to filter out the harmonics, it would cause a leading displacement power factor, and hence the overall PF might not be improved.

Low leakage inductance in a transformer actually makes the harmonic PF worse, because the charging spikes get taller and narrower. More harmonics, less fundamental.

On the other hand, the typical active PFC front end used in SMPSs gives a PF of 0.95 to 0.99 at all times.

If you don't understand the difference between harmonic and displacement power factor, forget I mentioned it. One thing to bear in mind is that for domestic users, the power company only charges for real power, because they can't be bothered with the complication of having the electric meters measure PF. This isn't true for larger industrial consumers though.

The EMC and safety testing cr@p is quite doable, and I have designed things that passed it. But it costs about $5-10k up front. To make it worth my while, I'd want to sell a few hundred power supplies a year at $250 each. If someone bought a batch of 10,000, the price would go down to not much more than a tube amp PT of a similar rating.

There are plenty of SMPS on the market that have never been EMC or safety tested at all, but Conner Labs would never stoop so low as to market such a thing.

[R.G.]

There are simple - but heavy! - solutions to PFC. Putting a ferroresonant "constant voltage" transformer in front keeps the spikes out of the line by supplying them from a *big* resonant circuit which is sine-wave-pumped by the AC line. Think of it as a HUGE power line filter. It still results in some phase distortion because the phase of the current accepted by the resonance may not be either in phase or fixed compared to the AC power line voltage. It is probably no worse than a single phase induction motor, but that's bad enough.

Hey. This is kinda fun. I just remembered some of the less-sane things from my motors classes. You could maybe put an induction motor in parallel with a ferro and adjust the mechanical load so it compensated the phase offset of the ferro. They used to do things like this for power factor correction on the main lines. If you drive an induction motor faster than it wants to go, the mechanical motion tries to move the phase of the AC power line the opposite way from the motor as a motor. That means you can mechanically cause a motor to look like a *capacitor*.

These so-called rotating capacitors used to be seen on power-line poles in the same places where the huge metal-cased PFC capacitors sometimes go today. I saw what had to be rotating capacitors on power poles in south China.