Listening directly to DC Power Supplies

[Dr.H]

Hi everyone, I am thinking of listening to DC power supplies to judge their quality/noise levels.

A few questions:
1. Any tried this (with a DC blocking cap of course)?
2. Is there any merit in this approach?-can it really be used to differentiate between say a bog standard lm317 and a good say SALAS shunt regulator?
3. Links to any other results based on the same approach?

Thanks
Ryan

[Alexontherocks]

can't you accomplish this with an oscilloscope?

[marce]

How would you resolve any low level noise that your ears cannot pick up!

[AndrewT]

Originally Posted by Alexontherocks

can't you accomplish this with an oscilloscope?

No.
I have an AC or DC input that goes down to 2mV/div and I can't see much useful information on the screen. There is no dominant frequency to lock onto.
As I increase the load on the supply I can see the ripple forming and increasing with the higher output current.
It is easy to lock onto that 100Hz + harmonics and then reduce the loading until the ripple voltage disappears into the narrow band of noise I can just about see as not a "flat line".

If I built a very low noise +20dB/+40dB amplifier, I could probably see much more.

[Dr.H]

Thanks, so is it best to boost the "noise component" of the power supply and measure that on a scope, rather than try to "listen" to the quality/noise of the power supply?

What about playing music through a simple FET or op-amp based buffer and using various power supplies set at say +-15V (to the buffer) to obtain an idea about PS quality (again say comparing a simple voltage divider with say zener with say LM317 with say shunt based)?

[SY]

Yes, you can. I've written about this several times in the past- I will play music and test signals through the amp, then listen to what ends up on the power supply rails.

[Dr.H]

Hi SY, can you paste links to your work and findings?

[SY]

They were all here on the forum- a bit of use of the Search function should turn them up.

[steveh49]

Originally Posted by SY

I will play music and test signals through the amp, then listen to what ends up on the power supply rails.

Manufacturers note: more effort needed on user manuals.

[wintermute]

You can also record the PS using your pc sound card (obviously also with a dc blocking cap) and optionally a preamp and then do an fft analysis of it to see the fundamental and harmonics. Note however you are likely to have problems with 50Hz noise pickup.

I've had mixed results using this method and need to revisit.

Tony.