Well, curiosity got the better of me. Here's how I propose a test for
shielding effectiveness. This is similar to other tests I have done in the
I'm thinking of using 2-feet of cable in the test, suspended between
opposite ends of test fixture I have. The cable will exit through holes at
opposite ends of the fixture. I'll have another single conductor one foot
away parallel to the cable under test as a source. I can excite that cable
as a "current" source by terminating the far end in a shirt to the outer
wall or as a "voltage" source by leaving it open.
This way the cable-under-test (CUT) unshielded ends and attachment points to
the Network Analyzer are not exposed to fields. I can use my Network
Analyzer to plot the attenuation between the CUT and the source.
First I'd establish a reference level to the parallel combination of all
conductors in the CUT with the cable terminated in a load resistance (near
its surge impedance) at the far end and matched through a broadband
transformer. This impedance would be the same a single conductor of the same
outside diameter and insulation type.
This would be the baseline calibrated at 0dB transmission loss into what
would actually be an unshielded cable with fixed spacing to a source.
Without disturbing anything INSIDE the box, I could ground the shield and
measure level into the CUT's inner conductors with the inner conductors
terminated in their differential and common mode impedance. (I want to
terminate the cables so SWR isn't a factor, or at least ripple is
I would also open the shield at the end away from the detector (which is the
way a shield is normally used when interconnecting two separate pieces of
gear with balanced lines) and measure the transmission loss. This really
would only take me a few minutes a cable since the fixture and
interconnections already exist from tests I made for cable manufacturers. I'
m going to make similar measurements in the next month or two anyway, so I
might as well check the audio cables.
I have two requests. If someone sees a flaw, let me know now. If people have
cable samples, please send me some with cable type clearly marked. I'd only
need about 3 feet of cable. But let's see if anyone notices a flaw in my
method. As far as I see, this would let us have a good comparison between
cables as long as diameters are not a significant fraction of the source to
RFI mailing list