[Top] [All Lists]

Re: [Amps] switching power supply for tube amplifier

Subject: Re: [Amps] switching power supply for tube amplifier
From: Manfred Mornhinweg <>
Date: Sun, 16 Sep 2007 14:52:49 +0000
List-post: <>
Hi ye all,

>> 10 pounds, 2KW - It looks promising. I am just wondering, 
>> how come no manufacturers use them?
> 75% efficiency
> 750 watts heat at 3kW load
> 17% regulation
> $700

I knew about its price and the lack of regulation, but I didn't know 
that it has such a low efficiency! That's of course unacceptable for a 
switching power supply. I wouldn't settle for anything less than 90% 
efficiency, and even this is still poor for a high voltage switcher, in 
which the secondary rectifier diode drop is negligible! Low voltage 
switchers easily achieve 92%, and the largest loss is the diode drop in 
the secondary rectifiers, so it stands to reason to expect an efficiency 
near 95% for a high voltage switcher.

Years ago, shortly after my 13.8V 40A switcher was published in QST, I 
was approached by someone wanting me to develop a high voltage switcher 
for a tube type linear amp. While the project was attractive, I didn't 
have much time, and I didn't know (and still don't know) what problems 
one would find with dV/dT issues, which made it hard to estimate the 
time required for development. So I declined. But sometime I might take 
up this project again.

In any case, a switching power supply without voltage regulation is 
pretty stupid. Adding voltage regulation costs almost nothing, compared 
to the cost of the power components which need to be there in any case. 
This is very different from a linear supply, in which the addition of 
regulation means adding lots of parts, which are expensive and generate 
lots of heat and thus lots of loss.

> The main advantage of tubes is a 
> regular transformer can be used.

I don's see that as an advantage. A REALLY advantageous active device 
would be one that can work directly off the line voltage! Insulation 
from the line would then be accomplished in the input and output RF 
transformers, making the whole amplifier MUCH lighter and less expensive.

Such devices might be tubes with large cathodes and small interelement 
spacings, but I fear that the high cathode current would require an 
excessive amount of power put into the filaments. So I have much greater 
hope placed in MOSFETs. There are some available off the bench, and 
quite inexpensive, to produce legal limit power with a single pair, fed 
directly from 300VDC, which can be obtained by simple rectification of a 
220V line or by voltage-doubling a 110V line. This would be very 
attractive, IF these devices would be able to work in linear mode, with 
good stability, no hotspotting, etc. But apparently this isn't the case. 
  In linear mode they need a much lower voltage to remain safe and 
stable, and so we still need a power supply - either a switcher or a 
large, heavy transformer affair.

By the way, based on long experience, I don't share the apprehension 
about fragility of semiconductors. I did professional repair of 
communication equipment between 1982 and 1989, and since then I 
occasionally repair equipment for ham friends. Speaking about 100 Watt 
final ampplifier stages, I have replaced too many failed tubes to even 
try to count them, but only very few transistors. And this is despite 
tha fact that tube output stages in this power class are few compared to 
solid state ones! Clearly in the field the transistors are more 
reliable. And in the 1kW class, I have had to replace about two dozen 
tubes so far, but never even one MOSFET! I admit, of course, that solid 
state high power amps are still rather few, but if they were unreliable, 
I would expect at least one or two to have found their way to my workshop.

The reason for this reliability is of course mostly in the protections. 
Solid state equipment always includes protective circuitry, while tube 
type equipment very often doesn't, or has only very basic protection. 
Then there is the issue of more difficult operation. Many tubes I had to 
replace were 6146A and sweep tubes, with their bulbs molten down, a 
result of improper loading of the radio. Many gassy 3-500Z might have 
failed for the same reason. Of course, a ham should know how to load his 
equipment, but it happens even in the best ham families that the little 
son moved some knobs, dad thinks his amp is still loaded on 20 from the 
last evening, goes on the air without checking the laoding, and the 
tubes are done in before he has a chance to notice.

I have built lots of equipment, mostly solid state but also some with 
tubes. I have never burned an expensive RF power transistor, or even a 
large switching transistor for that matter. Working with expensive and 
supposedly "fragile" devices, one simply has to apply basic safety 
measures: Current-limited power supplies, voltage clamps on the drains 
or collectors, a scope to catch any self-oscillation before it causes 
damage, and doing only quick, short tests until the stability of the 
circuit is guaranteed.
Being a solid state person, I apply some of the same precautions to my 
tube equipment, and as a consequence I have not even made a tube go red, 
much less yellow, by mistake. But I have done it on purpose a few times, 
and it looks mighty beautiful! I admit this is something one can't do 
with transistors! :-)


Amps mailing list

<Prev in Thread] Current Thread [Next in Thread>