Mains quality

[Eeyore]

A friend and colleague of mine is investigating with me methods of
eliminating audible clicks in high-end audio caused by the swiching of
other loads in the vicinity.

Some of it may be radiated from 'loopy' wiring but I reckon a lot comes
up the mains as common-mode noise in the audio band. So An MOV etc won't
and doesn't help.

As a diagnostic, I'v suggested building a diff amp that can check both
common and differential mode noise by examining the mains and notch
filtering the fundamental in DSP. But the waveform is also distorted
(typically 'flat-topped'), so harmonics would need to be filtered too.

Does anyone have any numbers for the harmonic content of a 'typical'
mains supply ? Voltage, not current.

Graham

 

[Tim Williams]

Does anyone have any numbers for the harmonic content of a 'typical'
mains supply ? Voltage, not current.

Graham

5%.

Tim

 

[Eeyore]

5%.

I've never seen it that high myself, only about 2.5% but I was thinking in
terms of the harmonic distribution, as all those of any magnitude will
have to be filtered as well as the fundamental.

Graham

 

[Eeyore]

You'd honestly be surprised by the "loopy" wiring inside the "high-
end" audio equipment and how susceptible it is to picking up
magnetically radiated EMI.

No I wouldn't ! I've seen classic BAD EMC tactics used by supposed 'experts'
who people respect.

The low-end consumer stuff has to put the EI core power transformer a
quarter inch from audio circuitry so the designers knew damn well they
better keep any loops in the audio circuitry super duper small.

You're not kidding.

And besides, they design 2 pieces of consumer equipment every single day,
and they don't sure as heck don't get any bonuses for making the PCB
or enclosure larger than it has to be. High end audio designers use
low-field toroidal transformers kept far from the PCB so they don't
know any better, and this is probably only the 2nd or 3rd audio PCB
they've ever laid out anyway.

One of the more intriguiging examples I came across of EMC was a very quiet
mixing desk with an external power supply. One of my own design but the
'manager' designed the simple PSU with an E-I . Just the routing of the earth
/ ground wire inside the PSU would affect the noise floor. The trick was to
'tack it' to the steel case so as to minimise radiated pickup. You could wave
it about and see the 50 / 150 Hz go up and down. Mind you we were operating at
a very low noise floor, yet such effects may explain other audible differences
where it's not so obvious.

I like to use chassis as earth / ground. That's pretty much the EMC rule.

Graham

 

[John Devereux]

No I wouldn't ! I've seen classic BAD EMC tactics used by supposed 'experts'
who people respect.



You're not kidding.



One of the more intriguiging examples I came across of EMC was a very quiet
mixing desk with an external power supply. One of my own design but the
'manager' designed the simple PSU with an E-I . Just the routing of the earth
/ ground wire inside the PSU would affect the noise floor. The trick was to
'tack it' to the steel case so as to minimise radiated pickup. You could wave
it about and see the 50 / 150 Hz go up and down. Mind you we were operating at
a very low noise floor, yet such effects may explain other audible differences
where it's not so obvious.

I like to use chassis as earth / ground. That's pretty much the EMC rule.

You mean circuit "zero volts" too, right?

 

[Eeyore]

You mean circuit "zero volts" too, right?

LOL.

Don't even get me started !

Graham

 

[GregS]

I myself had "forgotten" through my digital exploits in the 80's and
90's the rules about avoiding low-level noise pickup. I had learned
those rules largely from tube radio and audio equipment made in the
30's/40's/50's that I played with as a kid.

Things like keeping the loops small, and running not just twisted
pairs but routing those twisted pairs in the corners of metal chassis,
were fundamental to layout back then. And those rules applied equally
to the noise/hum generators as they did to the noise/hum detectors.
Perhaps "noise/hum detector" is better known as a "phono input stage".

But through the digital/LSI/VLSI/CAD-PCB years I had untaught myself
those rules...!!!! I relearned them building direct-conversion
receivers... receivers with 80dB or 120dB of gain at audio made me
sensitive to the issues again.

Using the chassis as ground is good but not good enough with phono
input stages. Basically the single ground point and the input jack
become one and the same. DC receivers get built the other way around -
the high current point (headphone jack connection to power supply most
often) becomes the single point ground.

As a rule of building equipment, you ground the chassis for protection,
and shielding for the circuit.I don't ground the circuit usually, except
for a static discharge resistor. My old Hitachi preamp used
100 ohm resistor from the inputs jack common to the chassis ground
which was not really grounded since it had a 2 wire cord. Keeping circuit
impedance high to ground helps keep ground currents low.

greg