Radiated emission

I am not sure if I understand what you are measuring.

I think you are measuring the output ripple, am I right?

I do not understand when you say "But when oscilloscope’s probes connected with a cable to output capacitor of PSU we see that a no ignorable noise being occurred with 36V" What you mean is that you are not connecting the scope probe to the output capacitors but instead there is a cable between the output capacitors and the scope probe. So your scope probe is measuring the ripple on one side of the cable while the other side of the cable is connected to the output caps, am I right?

If I am right, then I need to ask “why do you have the cable between" what is its purpose?

Whenever you measure output ripple you should follow the set up as indicated in in the section 11.7 of the attached link.

http://www.powerint.com/sites/default/files/PDFFiles/der260.pdf

If you do not follow the instructions as indicated in the document, you will see so much noise in your oscilloscope

Hello,

Firstly, thanks for your respond, i will explain deeper what i've done.

It is not about ripple, it is about a noise. It takes about 200ns and up to 10V as you can see from pictures. And it occours about every 1us depends on load current. This frequency increases with the load current. So i think it is the switching noise. It is both conducted and radiated because osciloscop shows same noise signal even i do no connect the probes. I'am connecting oscilloscop probe to a free cable and put it close to my SMPS then i see same noise with lower amplitude (2-3V). So the thing i want to learn is that how can i prevent from that noise?

Because it is getting over all of other circuits on my device. I have a DC/DC converter (LM25011) after that SMPS ouput and a choke circuit after it to realize a dataline over power line and that noise pierces all that layers and goes up to my dataline(up to 10V noise) so i have to prevent that noise. It occurs when my SMPS is alone or with resistive/LC loads or my other circuits.

I have some other SMPS circuits that i use such datalines made with Topswitch and Tiny swicth, none of them have similar noise. 

Best Regards,

All the switching power supplies radiate noise to the environment because they are switching high current and high voltage waveforms.

You need to find out what is the source of your noise. One very common source of noise is the core of the transformer. You can connect the core by a bus wire to the primary return or the primary high voltage DC bus. This will eliminate this source of noise. The PCB layout is a critical component when tuning radiated or conducted EMI. You want to have all the noisy traces as short as possible, all the noisy node's areas as small as possible, and  the loops than are switching high voltage/high current waveforms should be as small as possible. One example of such loop is the high voltage Cap (+ terminal), primary transformer, Topswitch and high voltage cap (- terminal).  This loops switches high voltages. To minimize the radiated noise, the mentioned components should e kept as close as possible in order to have a small loop. The larger the loop the more noise is radiated to the environment.

Another example of a noisy loop is the transformer’s output winding, the output rectifier and the output capacitor. This loop needs to be kept very small

Hope my recommendations can help you to fix your problem