Dave VanHorn wrote:

"Robin Cassidy VK3AYZ" wrote in
message om...

Are you using a switch mode power supply to run your radio's?

You can get modulation products produced by switching noise on the supply

rail.

Why do switchers have such a reputation in ham radio?

I realize it's possible to hose up a switcher design pretty thoroughly, (you
can also make a 7805 into about a 2W transmitter on 160M if you're not
careful), but I've never seen the sorts of problems that I keep hearing
associated with them..

BTW, I design switchers, and use them for power supplies in low noise
systems, without shielding or other expensive "voodoo" of any kind.

When designed properly, they are very efficient, and very nearly "silent".

1) Leakage Inductance from the core allows a small magnetic field to be
radiated.

2) By their nature they require a 'small' ripple voltage to exist on the
output as a consequence of the switching principle.

3) The input power, the raw power, is being modulated by the switching
cycle. This generates a transient pulse power on the input lines that
easily couples noise via various susceptibilities into the receivers.

4) The variable switching cycle, or the varying switching duty cycle,
creates broadband noise.

5) All of which are potential sources of trouble in a communications
receiver that wants to find a 0.16 uV/meter signal in the presence of noise.

As a designer of switchers, have you ever had to Qualify a switcher to
MIL-STD-461 and 462?. There is a conducted interference test on the
input power lines to the switcher that is brutal. The ripple on the
output fails the requirement of MIL-STD-462. The radiated emissions have
to be controlled, etc. The switching transients from ALL sources cause
EM Susceptibility in related equipment.

It's not impossible to design switchers for a quiet environment, but
I've spent $millions on getting them quiet enough to meet
MIL-STD-461/462 requirements.

3) The input power, the raw power, is being modulated by the switching
cycle. This generates a transient pulse power on the input lines that
easily couples noise via various susceptibilities into the receivers.

Only if you fail to decouple the input properly.

4) The variable switching cycle, or the varying switching duty cycle,
creates broadband noise.

5) All of which are potential sources of trouble in a communications
receiver that wants to find a 0.16 uV/meter signal in the presence of
noise.

Potential, yes I suppose.
As I said, I know you can make them perform badly, but it's just not that
hard to make them perform well enough that they won't be noticed on your
receiver, assuming you don't make it a practice to connect the power supply
directly to the antenna inputs.

The receiver itself may create more noise than a properly designed switcher.

As a designer of switchers, have you ever had to Qualify a switcher to
MIL-STD-461 and 462?.

That one I haven't. My hamshack dosen't require MIL-STD-461 either. That's
a large complicated battery of tests, that by it's very nature is expensive
to test to, even if you were testing a D-Cell battery.

I have several switchers in use, all commercial designs, and they are barely
detectable on my R-8500 or FT-847. I haven't had to go to any extremes (or
even any measures at all) to quiet them.

My PCs are another matter, they have needed ferrites on the cables, and EMI
absorption material inside the case, but that has proved to be managable.