[RFI] EMI/RFI AC line filter design info?

[Jim Brown]

On 9/24/2013 1:23 PM, Cortland Richmond wrote:
> We used to say "UL Approved" though there are now a number of NRTL's 
> whose mark is accepted by the various jurisdictions who have 
> electrical and fire safety regulations. 

Agreed on all counts with your post.  And it's important to understand 
that neither UL nor other NRTLs test for the EFFECTIVENESS of a product. 
Rather, they test ONLY for safety issues.

Another point (or four).  A major reason that electronic trash is 
radiated as a common mode signal on power wiring is improper design and 
construction of circuit "grounds" and their connections to external wiring.

The most common example is a green wire (and/or circuit common) that is 
not properly bonded to the chassis, sort of the power system equivalent 
of The Pin One Problem."  How does this happen?  The green wire is 
bonded to circuit common but not the chassis, perhaps due to paint in 
the way between that bond and the chassis. This is VERY common. In this 
example, the Green Wire is hot with trash, and if you don't choke it, it 
radiates.

This is why good, well designed line filters (Corcom, Delta, etc.) don't 
work on common mode trash.  Ah, you say, the data sheet shows strong 
common mode suppression, and there's a nice graph. But the problem is, 
that the power-systems definition of common mode is voltage between 
neutral and ground, and that's not the correct definition. Yes, those 
filters do what they say they do when installed inside equipment, and 
when that equipment is properly built. But they won't do anything to 
suppress common mode current (or current on the green wire) because the 
green wire goes right through them unfiltered.

Before I figured this out, I crammed big Corcom and/or Delta line 
filters into electrical boxes for use on the output of Honda 2000i 
generators on Field Day and California QSO Party county expeditions. 
They did NOTHING to the noise, which could be fairly strong above about 
5 MHz.  It took 4-5 turns through a big #31 clamp-on (the biggest one 
Fair-Rite makes) or an equivalent number of toroids to suppress the noise.

As to building your own filter from component values -- in addition to 
the very important safety and liability issues you've raised, there's 
also the matter of voltage ratings for capacitors used in these filters. 
By code, they must be rated for 3-6kV voltage spikes that often occur on 
power wiring. There's a discussion of this in my RFI tutorial.  And in 
one of Henry Ott's workshops, I learned that an important element of 
differential mode filtering (that isn't on the schematic) was the 
interaction of the leakage inductance of the choke with the capacitance 
that is between line and neutral!

73, Jim K9YC