I have been performing some EMC testing on a product (which I didnt design)
Its been a right barrel of laughs and it was quite challenging decreasing the radiated emissions below the 50dBuV/m limit
I had an issue with a 100MHz radiated emission which was decreased a lot with a proper earth and screen termination
The other problem area was around 33MHz and I only managed to just scrape it, took the emissions from 57dbuv to approximately 49dBuV/m which is just too close for comfort
After some investigation we determined the radiation is transmitted by the input power supply cable
I am designing an add on PCB fitted directly to the input to handle surge and my thinking is why I have the chance is there anything I can do to filter these troublesome 33MHz radiated emissions
Filters are normally designed for conducted emissions and theres little I can find to help me, the radiated emissions (to my knowledge) are a function of common mode current probably drawn from the SMPS that is fitted directly to the input cable that is radiating
So my question
If you had a radiated emissions issue at approximately 30MHz what would/could you do to filter them to reduce them
Edit
I have laid out a simple PCB, its basically a standard EMI filter but with surge and transient protection with some RC snubbers included to have a play around with
Is an LC snubber a bad idea?, my thoughts are I can include an L that is larger than the parasitic inductance from the supply so I can (in my mind) tune my components to shunt whatever frequency I am interested in as the parasitics I have no control over become negligible but am I just including a resonance that will make things worse?
Is it even possible to filter signals that are radiating? nobody seems to talk about this when it comes to EMC its always shield a cable or earth things better with ground planes etc no mention of filtering which makes me think its just another case of I havent got a clue what I am doing!
Best Answer
If you have radiated emissions from the power cord at 30 MHz, you probably also have a conducted emissions problem. I would address both now, not wait until you flunk the conducted emissions test later.
The simple answer is a off the shelf line filter. These can be quite effective, and come with specs as to how much they attenuate common mode and differential mode signals at various frequencies. 30 MHz is "high" in this context, so relatively easy to filter out.
Off the shelf line filters are simple and robust, but can also be large and expensive relative to the components inside. If you don't feel comfortable with filtering emissions and dealing with line power, just get a line filter.
If you want to do this yourself, I'd start with a balun and a capacitor to ground on each of the two line side terminals of the balun. That will mostly knock down common mode signals, but will also attenuate differential signal somewhat at this frequency. Keep the requirements for the capacitors in mind. You want capacitors specifically rated for power line use.