Switching power supplies: why some are noisy and some not?

[psst]

After my good experience with the "One Spot", I wanted to use switching power supplies for my rack system (to avoid electromagnetic noise with the normal supplies, which is very noticeable because of the proximity of the adaptors to the rack enclosures).
Since there is no "One Spot" for 15V or 18V, which I need, I bough some "generic" switching power supplies in my usual electronic store, but they are noisy as hell (no electromagnetic noise with them though). I thought they would be poorly filtered, so I built a filtering circuit for them with no luck at all.
Where does that noise come from? Is it fixable? Anybody knows about a good silent switching power supply that outputs 15V, 18V or even more?

[R.G.]

I only have one good data point on estimating the effort to make them quiet.

It took about six months of intensive reworking and design effort to get the 1Spot quiet from where it started. That's with the original design engineers working on it in a fully equipped lab.

[frank_p]

Well, thanks for the question and the answer. 

[psst]

 
Thanks, RG.
I guess it won't be easy at all to find a silent one for more than 9V...
(Unless I use 2 1Spots to get 18V)

[darron]

Thanks, RG.
I guess it won't be easy at all to find a silent one for more than 9V...
(Unless I use 2 1Spots to get 18V)

i don't think that it works that way...

why not try the geofex voltage doubler circuit from your existing 9V

[psst]

RG, it wouldn't be bad to have 15V and 18V "One Spot"s besides the 9V and 12V ones.

Tell the guys at Visual Sound if they like the idea, for sure there are people like me needing a power source like that for sure.   

[nelson]

Inaudible switching frequency, good filtering and close attention to layout - obviously not the whole story but if I were designing one that's where I would start.

[Paul Perry (Frostwave)]

It's true that switching power supplies operate above audio frequencies - but it is also true that they usually radiate a huge amount of RF. Which gets into the equipment & is rectified. You can try the usual RF tricks - bypass caps on the inputs, ferrite chokes - but I advise you to stock up on the (illegal) DC warts while you can.
And get a lifetime supply of incandescent globes while you are at it. Those compact fluros radiate as well.

[PerroGrande]

Switchers are, by their very nature, noisy.  It has always been a matter of great effort to get them to be sufficiently quiet for high-end audio use.  Whether you go linear or switching, there is going to be some form of noise to deal with -- just the inevitable nature of AC to DC conversion.

Apart from the OneSpot, and perhaps a few others that have been engineered with the end-use in mind, "off the shelf" models are going to be a bit of a crap-shoot when it comes to their "cleanliness".  You *might* be able to filter them down to usefulness if you could look at the output on a scope and see the offending frequencies, etc. 

[frank_p]

It's kinda going into a lot of trouble for not much gain, it seems (?)
There is plenty of conventional tranfsormers that do the job without having to deal with that noise.

[psst]

No, I use lots of gain and compression sometimes, and the noise is strong then...

Yeah, linear transformers are clean and don't radiate EM noise if put a little far from the circuits, but if I want to keep them in the rack they are too close to the rack units and noise appear.

[mdh]

Well, isn't that exactly the situation that toroidal power transformers are suited for?

[psst]

Yeah, that's one of the things I have to try. In theory it gives like 1/10 of the EM noise, but I have to try first. Thanks!

[Processaurus]

Why not just put a thick plate of grounded metal in your rack to shield your gear from the EMI?