Susceptibility of low level audio signals to mains-borne interference (2023 Update)

Tabby gets Catty wrote: ==================== > > > > > > > Low Z make long cable runs possible.

** Then take f****ng notice of it !!!!.

** Then the OP has a good, practical way.

** You did no such thing. See above.

...... Phil

So, that's problem #1: you need an antenna to pick up the interference, that matches somehow the accidental pickup. Easier to use differential circuitry.

Lock? The 'tone' being local current variations in an AC distribution system? But, the AC part is repetitive currents ONLY when the loads don't change. Every oven that switches its elements PWM style will modulate that current.

Acoustic noise reduction, yes. Electromagnetic, though, you basically just try to remove coupling (like, with coaxial wiring).

======================== .

** Low Z allows long cable runs. Co-ax cable rejects mag hum fields better than twisted pairs. Metal boxes reject all E field hums. Differential inputs do none of the above.

...... Phil

No. Phase lock to the hum and its harmonics in the single signal. Then subtract them out. Might get interesting.

Yes, that's problem #1: 'the hum and its harmonics' are modulated, which makes your amplitude unknown, so what do you subtract? Phase locking gives you a phase, it does NOT give you an amplitude; it does not give you the signal.

An N-path commutating capacitor filter could notch out the primary and harmonics. It can be a very high Q filter even with awful components.

I once tried to salvage a live recording corrupted by mains and its harmonics. I used a comb filter, which got rid of the interference but the effect on the music quality was fairly unpleasant.

John

John Walliker wrote: ================

** How the heck did that happen ? Cabling issues? ** Works like a flea rake - right ?

** Is that a "masterful understatement " ?

..... Phil

What was the Q of the fundamental's notch?

Tom Gardner wrote: =============== >

** Fun Fact:

If you take a 1/3 octave ( or 31 band) graphic equaliser and adjust the sliders so every adjacent one so they are either 3dB up or 3dB down from the zero level, the effect is damn near inaudible.

FYI most such units have a bypass switch so you can easily try this while listening to speech or music.

Ruins the "flat response myth" completely .

...... Phil

Yes. The stage manager decided to move the Revox recorder shortly before the performance which required unplugging a couple of XLR microphone connectors. He did this by pulling on the cables until the connectors came out. I managed to fix the broken wires in time, but he had also repositioned the microphone cables next to the stage lighting circuits and the thyristor dimmers made a very nasty low-level noise that I didn't notice in time to fix.

That would be a fair comment. Filtering the fundamental wasn't enough. Thyristor dimmers do have a very nasty audio spectrum.

I really don't remember now. This was around 1978. I did have access to a well equipped acoustics lab, so it would have been an expensive professional box. The Q can't have been ridiculously high, or the harmonics would have missed the notches at higher frequencies due to tape recorder wow and flutter. I would have tried to do the best that the equipment was capable of.

John

It was about then that I used "my" filter to reduce noise at 8kHz. It had a Q of ~4000 with 10% capacitors :)

Obviously that wouldn't work with wow and flutter, and it would have been difficult for me to track any frequency variations in my application.

John Walliker wrote: ===============

** That must be quite a while ago ! Last time ( cira 1979) I had dealings with a worn out A77 at live show it was no fun at all.

But the incident that closely parallels your experience involved a show called "'Elvis Lives". One rather overweight Elvis impersonator, two pretty gal singers and a six piece band. I was roped in to be the live sound engineer - though a humble electronics tech normally. A half decent PA was set up in a old cinema type venue, relying on one Phase Linear 700 amp.

8 channel mixer, mostly AKG mics and decent Cerwin Vega speakers.

On the opening night - all went beautifully, the gals sang like angels, the band was a-hopping and the fat guy was not half bad. Then the FOH sound stopped DEAD ! Elvis chucked his mic away and stormed off stage. WTF had happened ??

Soon found out.

Like in your case, the *stage manager* ( 62 year old 1/4 wit ) decided to move the amp rack to a place he liked better. While leaving the 15 inch floor fan placed right behind it where it was.

Thermal cut offs being what they are .....

.... Phil

It was either 1977 or 1978. I had a few tries at recovering the recording over the next couple of years when I had access to potentially suitable equipment.

Clearly you'd have to measure the amplitude and phase of the hum at each frequency, and synthesize the cancel signal. If the hum profile changes over time, the cancelers will track but lag a bit and let through snippets of hum. I suspect that would have low audibility, but golden ears would have to test that.

The nice thing about an FPGA is that you only have to do the thinking once, and let the compiler make 300 copies.

That would be the advantage of subtractive cancellation; it wouldn't change the signal path at all.

Don't n-path cap filters generate a lot of distortion?