[Amps] SMPS Question

[Manfred Mornhinweg]

As a "self-appointed specialist" in low noise switching power supplies, I have 
to chime in...

>> You easily turn them off during RX - not what you would normally do with
>> a Linear HT supply.
> 
> For QSK? For contesting?

Yes, Jim, even for QSK you can turn it off during RX! How much time are you 
willing to allow for start up? Is 10 milliseconds acceptable? If it runs at 
18kHz, that gives you 180 full cycles, which is more than plenty for complete 
stabilization. If you need faster startup times, 2ms is still perfectly 
possible. Even 1ms. And that's for that comparatively low 18kHz switching 
frequency! Most switching power supplies these days operate from 40kHz up.

Turning off both the power supply and the tube during RX has the advantage that 
diode hash in the tube is also eliminated.

In a linear amp using a switching supply, indeed I would turn it off during RX, 
both to avoid noise and to improve the efficiency. But it shouldn't be necessary 
for noise reasons alone. I run all my radios from homemade switching supplies, 
and have done so for over 20 years, with no noise at all.

I'm not saying that all switching power supplies are quiet. Unfortunately many 
of them do produce a lot of RF noise. That's because their designers optimized 
them for lowest cost and high efficiency, with electromagnetic compatibility 
being either no consideration at all, or only to the extent of meeting FCC class 
A or class B limits, which are relatively lax. By clever design, sacrificing a 
small amount of efficiency, and adding a few dollars worth of noise-reducing 
components, one can make extremely quiet switchers.

For a low cost HV supply for a tube amplifier, my approach would be this:

- A power factor correction stage, pre-regulating to about 380V on a large 
electrolytic capacitor, using two or three MOSFETs in parallel, or perhaps an 
IGBT, a SiC diode, and any of the current PFC chips;

- A full bridge of IGBTs, controlled by any of the usual PWM controller chips;

- About 8 small ferrite transformers, their primaries in parallel driven by the 
bridge, their secondaries each feeding a SiC diode bridge, filter inductor and 
foil capacitor;

- The ~8 DC outputs connected in series.

This would be complemented with the appropriate feedback circuit, noise filters, 
snubbers, etc. The whole thing would certainly have far more parts than a 
conventional power supply, but it would be lower cost, smaller, more efficient, 
very much lighter, would deliver a regulated output voltage almost immune to 
line and load changes, accept a wide input voltage range, it would be short 
circuit proof, intrinsically limiting flashover current, have soft-start, and it 
would have nearly perfect power factor, allowing the use of plain simple home 
power outlets, relatively small gasoline powered generators, etc.

If a tetrode is used, screen voltage can be obtained from the first of the ~8 
sections.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================