Audible noise from SMPS

[aetosa]

I have built a 12Watt SMPS with TNY268 .
Works ok, but I want to use it for a chip amp (which now is connected to a simple unregulated PS. Noiseless ).
The problem is the audible noise from my SMPS (I don't mean the mechanical noise from transformer or some capacitor of the SMPS, but from the voltage ripple I suppose). Not too annoying or too loud, but existing.
I tried a capacitor 2200uF at the output, but no change.
The truth is I have put plain electrolytics in the output and not low ESR, and the inductor in the output (it's a flyback topology) has not the right value.
Can these be the cause of the noise in the amp?
What can I do?

[tvrgeek]

Build a decent linear supply.

Baring that, look at the papers on tantlum-poly caps designed for SMPS filtering, sim the circuit in Spice and see what parts may need tuning. Physical layout, wiring, RF are all suspects. I have no use for SMPS in audio. Not yet.

[luka]

feedback, regulation can create big problems, noise on output

Is your circuit totaly the same?

[aetosa]

Originally Posted by luka

feedback, regulation can create big problems, noise on output

Is your circuit totaly the same?

Same with what?
I built it as designed by power integration's application, except for some parts which I replaced with their equivalents.
Only inductor in the output differ and the caps in the output which are plain electrolytics instead of low ESR.

[KatieandDad]

You don't say what is out of spec with your inductor.

In a flyback design, if the inductor is too small you will get a lot of PSU ripple.

[aetosa]

The inductor is a random one, I made it without any calculations at all.
But the ripple is the same as my SMPSs' in my pc (judging by the ear, I don't have access to an oscilloscope). Which is supposed to have everything well calculated.

Since I mentioned my pc: how it's sound card is noiseless?
So, there is a way of filtering a 'noisy' SMPS (like that of my pc) for use in audio.
I just have to find it out!

For a beggining, I will try to put an inductor of the correct value, but I think that that's not it.

[darkfenriz]

You have chosen a very bad chip for audio supply.
It works with constant pulse width and skips some on the cycles, i.e. depending on the load it can be every 100th or 2nd pulse.
It must have audible noise, especially at light load, because it is how it works, you'd better chosen a PWM one.
This one is better if you need better line rejection and faster load response, the price is audible switching components.

[KatieandDad]

Originally Posted by aetosa

The inductor is a random one, I made it without any calculations at all.
But the ripple is the same as my SMPSs' in my pc (judging by the ear, I don't have access to an oscilloscope). Which is supposed to have everything well calculated.

Since I mentioned my pc: how it's sound card is noiseless?
So, there is a way of filtering a 'noisy' SMPS (like that of my pc) for use in audio.
I just have to find it out!

For a beggining, I will try to put an inductor of the correct value, but I think that that's not it.

A bigger one will do no harm.

Can you find a defunct computer PSU and rob a toroidal core, the biggest you can find. Then wind as many turns as you can onto it.

This one might benefit from a C L C filter.

Not exact but its fun DIYing at nil cost.

[aetosa]

Originally Posted by darkfenriz

You have chosen a very bad chip for audio supply.
It works with constant pulse width and skips some on the cycles, i.e. depending on the load it can be every 100th or 2nd pulse.
It must have audible noise, especially at light load, because it is how it works, you'd better chosen a PWM one.
This one is better if you need better line rejection and faster load response, the price is audible switching components.

I didn't know that each SMPS chip has it's own...eccentricities .
But, as I allready mentioned, my pc's SMPS is also noisy when connected with this chip amp (or any amp). While it's noiseless with my pc's sound card...

Which SMPS chip would you suggest for audio? (SMPS output 12 Watt)

[aetosa]

Originally Posted by KatieandDad

A bigger one will do no harm.

Can you find a defunct computer PSU and rob a toroidal core, the biggest you can find. Then wind as many turns as you can onto it.

This one might benefit from a C L C filter.

Not exact but its fun DIYing at nil cost.

I 'll try that!
In my design the inductor should have an inductance of 3,3 uH (if I remember well it is uH, but I 'm not sure)