Understanding power noise

mark2 · February 15, 2021, 03:09:18 PM

I'm building a silicon tone bender, and each of my 3 power supplies gives it a slightly different, awful sound.

e.g. https://www.youtube.com/watch?v=qlcnwXPGfRc

These problems often go away when I switch over to a PCB (whereas this is on a breadboard), but I'm curious if anyone can identify specifically what is causing that sound. Thoughts?

antonis · February 15, 2021, 03:12:43 PM

High pich interference, maybe..??
(cross talk due to many fllying jumpers..)

FiveseveN · February 15, 2021, 04:37:10 PM

Quote from: mark2 on February 15, 2021, 03:09:18 PM
a slightly different, awful sound.

Different how, like in pitch?
Are all the power supplies of the same design (switched-mode, linear stabilized/filtered)? Big ferrite says SMPS to me and not all are appropriate for audio.
It may be interference indeed, may also be oscillation with all the gain and coupling on a breadboard. Does the noise change if you move leads around? How about moving the whole board around the room/in different orientations?

mark2 · February 15, 2021, 04:56:46 PM

Here are the 3 different sounds: https://www.youtube.com/watch?v=YpK56egHCL8
They're all 9V wall warts. One is coda effects, one a no-name, one came with a pedal (Foxy Tone Box)

The sound doesn't seem to change when I wiggle cables or move the board around. I haven't tried other rooms yet.

antonis · February 15, 2021, 05:25:12 PM

Just to distinguish "in or out circuit" noise source, wrap around breadboard with a grounded aluminum foil..

mark2 · February 15, 2021, 05:31:12 PM

Grounded foil didn't seem to help, though I did have some slight gaps.

Each supply has a different noise which makes me think there's an interaction with the power.

antonis · February 15, 2021, 05:38:50 PM

Quote from: mark2 on February 15, 2021, 05:31:12 PM
Grounded foil didn't seem to help,

Good for you..

Now rearrange your breadboard for as physically as posible short jumpers of only horizontal orientation..

iainpunk · February 16, 2021, 10:17:31 AM

have you put a 100uf cap on the power rails of the BB?

cheers, Iain

mark2 · February 16, 2021, 11:53:18 AM

Yeah I did.

R.G. · February 16, 2021, 12:37:15 PM

You're fighting two different things. First, the Tone Bender has nearly zero power supply noise rejection. Second, the power supplied make some amount of both power and ground current noise, which translates into voltage noise in the pedal and is amplified.

One thing you could try that might work is to wind the power cable from the power supply around something ferromagnetic - a chunk of iron or steel - as many loops as you can get. This makes a common mode choke to help fence the noise into the power supply, not the pedal. Inside the pedal, you could try putting an inductor in series with the power lead and a large-ish capacitor parallel to the power in the pedal.

Bottom line... you have to work on both the source of the noise and the sensitivity to the noise inside the pedal.

As you've found, there are better and worse power supplies for producing noise. Any pedal power supply not specifically designed for low pedal noise is likely to be worse.

PRR · February 16, 2021, 02:56:57 PM

Quote from: R.G. on February 16, 2021, 12:37:15 PM....wind the power cable from the power supply around something....

I noted he already has that. The lump at the end. However the pre-lump dirty part of the cable is laying right on top of the precious audio wiring.

It helps to think of water and sewage lines. You don't want them near each other, because everything leaks some.

R.G. · February 16, 2021, 03:34:49 PM

Yeah, I saw that. Those things are mostly for broadcast band and above RF switching noise (... from painful experience

) so what I meant but didn't say very well was to try completely overpowering the noise with a filter that would suppress even high-audio trnsients.

mark2 · February 16, 2021, 04:30:03 PM

I've got all the other kinks worked out, so unless you strongly recommend an inductor I'll first try a PCB without it since I worry it'll introduce new problems I'm not yet equipped to tackle. In the meantime I'm just using a basic filter a la the Rat.

I'm planning to lay this out and order some PCBs tonight and cross my fingers.

By the way, making this an even riskier build I'm putting a digital IC and a noisy OLED on it. Oh boy

blackieNYC · February 16, 2021, 04:37:16 PM

How is it with a battery?

mark2 · February 16, 2021, 04:48:48 PM

The high pitched and unusual nasty noises are gone when using a battery. It's left with a more typical high noise floor, kind of like an unplugged cable. I'm much more hopeful that'll go away with a PCB.

Not quite as confident about the issues with the wall warts. Many circuits and most boxed pedals are dead quiet with these supplies; Some, like a French Toast I just returned are unusably noisy; and others like this breadboard are insane.

Rob Strand · February 16, 2021, 04:49:51 PM

QuoteHow is it with a battery?

+1 Check that first.

Try putting about 220 ohm across the power rails of the DC Jack. A lot of switch-modes can produce noise with no load, the older transistor pedals aren't enough load.

For a *positive* ground pedal: Add a 100uF across the power rails *of the pedal circuit*, place a 100 ohm resistor between the DC Jack -ve lead and the circuit -ve lead. If that's not enough try adding another 100 ohm resistor between the DC Jack +ve lead and the circuit +ve lead.

mark2 · February 16, 2021, 05:18:45 PM

Quote from: Rob Strand on February 16, 2021, 04:49:51 PM
Try putting about 220 ohm across the power rails of the DC Jack. A lot of switch-modes can produce noise with no load, the older transistor pedals aren't enough load.

Wow, that made a world of difference. A 1k gave it a higher pitched noise, 470 quieted it a bit, then a 220 silenced it.
I didn't have any higher wattage ones so it got pretty hot pretty quick.

Any reason, other than cost, not to just stick a 1/2W 220R into every production pedal to remedy them on junk supplies? Is it to keep their power consumption low?

Rob Strand · February 16, 2021, 05:31:56 PM

QuoteWow, that made a world of difference. A 1k gave it a higher pitched noise, 470 quieted it a bit, then a 220 silenced it.
I didn't have any higher wattage ones so it got pretty hot pretty quick.

Yep, it gets a bit hot.

Good work trying the higher resistors through.

QuoteAny reason, other than cost, not to just stick a 1/2W 220R into every production pedal to remedy them on junk supplies? Is it to keep their power consumption low?

Well the real reason is it's not the pedals problem but in reality it becomes the pedal's problem. It is a bit crazy burning up 40mA for a pedal that might only take 1mA. For a single pedal no big deal but if you had a pedal board the sum of the 40mA power wasters can add up quickly. The power supply only needs a single 40mA load to get it over the line so one suggestion is to have a single dummy load you pass the DC Jack power through.

News:

Understanding power noise