how is this noise on isolated power supply module

[po210]

I´m using this isolated power supply module

**broken link removed**

and I see the peaks shown in the photo in the output.(scale 10mV, 2 microsecond period)
they are due to the internal constitutuon of the module, but how can I get rid of them? A diode?

 

[DerStrom8]

First off, I recommend giving your threads a decent title. If you just write "how", most people won't even look at it. You need to give your threads descriptive titles. If you don't, don't expect anyone to answer.

 

[MikeMl]

Switchers are crap, especially if you are trying to run low-level analog, data acquisition or audio. Long-live analog powersupplies (hint, they contain iron).

 

[ronsimpson]

I´m using this isolated power supply module and I see the peaks shown in the photo in the output.(scale 10mV, 2 microsecond period) they are due to the internal constitutuon of the module, but how can I get rid of them? A diode?

No diode!
1) Add capacitors. 0.1uf, use short leads.
2) Add LC filter. Small coil and 0.1uf cap. The resonant frequency needs to be below the switching frequency. Fr<= 100khz.
3) Add two coils, one on each lead. (+,-) then the cap.

 

[crutschow]

That's a very cheap DC-DC converter and you can't expect low noise. Why do you need to reduce the noise? What is it powering?

 

[po210]

I´m using it for small scientific experiments.The isolation module drives an ampmeter.the ampmeter measures current in the output path,but introduces the same voltage peaks in the output voltage.So I´d like to remove them before feeding the ampmeter

 

[crutschow]

Along with what ronsimpson suggested, you may also need to consider putting it in a shielded metal box and use feed-through filters on all input and output lines including the input common and output common.

 

[ronsimpson]

You might try a "common mode" transformer filter. Google it. I don't know if the noise if common or differential noise. (both)

 

[po210]

Hello again.
I´ve tryed all with no sucess.I made a T filter with two coils and one capacitor to ground,with resonant frec below 100Khz.I´ve tryed several regulators (LM317,7805) and same peaks are there. And some small transformers both in ground an Vcc lines taken from and old TV.Does somebody kown a link with similar solutions to this problem just to have an idea of real solutions?

 

[ronv]

It may be measurement error. Leave the scope ground where it is now then scope that ground point. You may still see the noise.

 

[po210]

Not sure what you mean.I´m just a hobbyist.
I´ve ha a two channel scope.If I measure the output isolated voltage using only a channel probe I see the peaks.Nothing changes if I connect the other probe ground to same isolated ground. If I connect it to the "real" (non-isolated) groung input,big dumped oscillations appear,but i thinks this is expected because the somehow two grounds levels are present.

 

[RCinFLA]

20 mV of ripple or more is not uncommon for a switcher. The ripple is coming from the peak current pumping into and out of the filter cap on the output. It is likely an electrolytic cap and its series resistance causes the ripple.

If you want lower noise output follow the switcher by a low drop out linear regulator or add the additional passive filtering already suggested previously.

 

[po210]

I´ve already tryed all that, but with a 7805 and LM317,but there is little difference.
Why should it be a low drop-out regulator?

 

[RCinFLA]

I´ve already tryed all that, but with a 7805 and LM317,but there is little difference.
Why should it be a low drop-out regulator?

You are doing something wrong. Likely not giving enough voltage overhead for the linear regulator to operate properly. You should be able to get 40 db to 60 db of power supply noise reduction. The two regulators you mention are not low dropout regulators. They require at least 2.5v to 3.5 v higher input voltage then regulated output voltage.

Low dropout regulators requires the least amount of input to output voltage differential to maintain regulation and reasonable input noise rejection. Depending on current requirements, a low drop output regulator only requires about 1 to 2 vdc higher input voltage.

In general, linear regulator lose their input noise rejection when the voltage drop across the regulator is not great enough. Noise rejection will degrade before output regulation is lost, as the input to output voltage differential decreases.

 

[ronv]

Not sure what you mean.I´m just a hobbyist.
I´ve ha a two channel scope.If I measure the output isolated voltage using only a channel probe I see the peaks.Nothing changes if I connect the other probe ground to same isolated ground. If I connect it to the "real" (non-isolated) groung input,big dumped oscillations appear,but i thinks this is expected because the somehow two grounds levels are present.

Leave the A channel ground where it is and using the A channel probe, probe the ground point where the A channel ground is. It should be perfectly clean. If it is not you are picking up the switcher noise with your scope set up