80kHz-300kHz Conducted emissions are common mode?

[cupoftea]

Hi
The following two Conducted EMC scans of a 60W offline flyback proove that the spikes between 80khz and 220khz are due to common mode noise.
The only difference between scan1 and scan2 is that scan2 has a wurth 7448640415 (10mH) Common mode choke in the AC filter, whereas scan1 has a kemet SS11V-07120 (12mH) common mode choke instead.

Earlier attempts to treat the 80khz to 300khz spikes by increasing diff mode inductance failed badly....again prooving that the problems seen at these frequencies are common mode.

This trash's all the App notes and web articles which profess that anything below 1MHz is a Diff Mode problem, and above 1MHz is common mode problem....would you agree?

wurth 7448640415 (10mH) Common mode choke

https://www.we-online.com/catalog/datasheet/7448640415.pdf

kemet SS11V-07120 (12mH) common mode choke

https://www.mouser.co.uk/datasheet/2/212/1/KEM_LF0023_SS11V-H-CH-1104439.pdf

 

[FvM]

The scans reveal other differences, the peak pattern suggests a change of operation frequency.

CM interferences may or may be not the dominant problem in this specific setup. No reason for generalized conclusions.

 

[cupoftea]

Thanks, the fact that we're even entertaining the point that the big peaks between 80khz-300khz could actually be common mode, appears to show that we are trashing the "nothing below 1MHz is common mode" philosophy.?

 

[cupoftea]

The scans reveal other differences, the peak pattern suggests a change of operation frequency.

Thanks, may i please ask, are you "cast iron" sure that the switching frequency is different between the above 2 scans?
It was supposed to be the same.

I am just wondering if the "pass" scan actually passed because it's loading was reduced....that would have brought about a rise in switching frequency because its COT control.....
....but the loading was not supposed to have been reduced...so i am wondering.

Can we be "cast iron" sure that that big red peak near to the 100khz gradation is equal to the switching frequency?

 

[FvM]

I see about 85 kHz fundamental in first and 102 kHz in second scan, clearly recognizeable by the harmonics.. I conclude that either the configuration or the operation point is considerably different, thus it's a pure guess that the level differences are only causes by the changed CM choke.

--- Updated Mar 23, 2022 ---

My general point is, you surely can make an EMI test setup where CM interferences dominate in the sub MHz range, e.g. if DM is already filtered by suitable X capacitors. The capability of propagating CM interferences depends strongly on the connected load circuit extensions, possible internal Y capacitors between primary and secondary, winding design.

 

[cupoftea]

Thanks, i see what you mean by the "harmonics"...its those regular spikes........increasing the diff mode inductance doesnt get them down...which i think is odd.......always one of them peaks up and fails you, no matter how the diff mode inductance is raised or lowered.....which makes me wonder if they really are harmonics of the switching frequency?...because if they were that, then surely they would always diminish as the diff mode inductance is increased?
I suspect they are some resonance in the mains input filter....involving the strays of the L's and C's in that filter....would you agree?

Scans 1 and 2 in top post were supposed to be at the same power level......but the fact, as you say, that scan2 is at a higher frequency tells me that someone lowered the power level of this scan......resulting in a pass at this lower power level. What do you think......do you think scan2 is at a lower output power level than scan1? (it shouldnt have been, but...)

Its Constant off time so the switching frequency does get higher as the output power reduces.

I looked on the scan data and it doesnt show the power draw of the kit....i am sure it always used to say that.......it seems a poor deal to not append the power draw of the DUT during the EMC scan, would you agree?

 

[mtwieg]

I'm guessing the "trash" app notes are referring to a specific methodology for designing EMC filters. If you didn't follow their methodology, then you shouldn't expect their results. You didn't actually give any links to the "trash" app notes, or show your own design, so we can only guess.