When a power supply fails conducted emissions tests, engineers must look beyond the obvious.
One of my colleagues was baffled by a client’s open frame power supply that was failing conducted emissions tests. It had two or three sections of power line filter between the power line and the bridge rectifier. The filter values should have been adequate to keep the conducted emissions within spec. The filter parts were toroidal inductors and the usual X and Y capacitors. Changing filter component values had little effect. He showed me the spectrum analyzer results and the schematic and it didn’t seem reasonable that the conducted emissions should be that high.
Looking at the supply showed the filter close to the power line input terminals and rectifier. And the main power switch transistor was also nearby. A scope showed us high frequency E- and H-fields near the transistor. Interposing a piece of grounded sheet metal between the transistor and the filter section altered the emissions a lot. (Be sure to wrap the metal in insulation if you do this!)
A second piece of metal underneath the PC board reduced the emissions to below the required limit. The client was advised that either shielding the line filter or relocating the offending switch transistor would make the supply pass conducted emissions.
So it turned out that it was really a near-field radiated emissions problem and the radiated emissions were coupling into the line end of the filter. It was a good reminder that emissions usually exist in both conducted and radiated forms. Having a scope and E- and H-field probes helped locate and solve the problem.
Author Jon Wexler is a freelance engineer who specializes in analog, power and EMC