David wrote:

I agree that there's plenty of audio-band crap on the
mains, but not from SMPSs, which generally work at
frequencies well above the audio band. AF crap is either
harmonic distortion of the mains waveform, or comes from
commutator motors and the like.

Just because SMPS "work at frequencies above the audio band"
doesn't mean they don't also produce audio-frequency noise.

A SMPS with a fixed switching frequency can easily employ a
mains input filter that targets that frequency, although it
isn't so easy to deal with all of its harmonics.

It's likely that its PFC boost converter, if it has one, has
a different problem. It tries to ensure that current drawn
by the supply is in synchronous proportion to the mains
input voltage. This is problematic in the face of the large
following capacitance which is trying to drag the current
out of phase. It's errors appear in terms of phase and
amplitude, so harmonic distortion is a likely outcome. A
large number of mains harmonics fall within the audio band.

Also to take into consideration is how the SMPS is
regulated, and what it does when under or over-loaded. It
could be that it operates in bursts, and the frequency of
bursts could be within the audio spectrum.

AF crap shouldn't be a problem with a linear PSU. The
transformer will remove common-mode noise, whilst the
filtering on the rectified output (which you need anyway
to remove mains-frequency ripple) will eliminate the rest.

Sounds fair enough. Why is it a problem for Ian Bell,
though, I wonder?

It's a problem for my AVO valve tester, which only rectifies
the grid voltage, and does no smoothing at all. There's a
simple European standard for mains harmonic distortion from
consumer equipment. It allows a fair amount, IIRC. They
don't appear to have taken my AVO into consideration.

Ian