Topband: RX splitter - ground common or not?

[Tom W8JI]

> The only thing I disagree with about Tom's advice is that, with the 
> exception of a VERY large string of #73 beads (at least 100 beads), beads 
> are useless on HF, and even those are not very useful at 2 MHz.

Advice like that must assume some ridiculous amount of common mode impedance 
being driven by extreme voltages. As a result, the advice misleads people 
about the requirements of a CM choke.

Attenuation added by beads, or any series CM impedance, depends on the 
common mode impedance at the insertion point. The beads, or any series 
impedance, are part of a system with the path impedance outside the added 
series isolation.

Let's say we have a common 10-foot long lead inside the shack, between two 
pieces of somewhat well-grounded gear, driven by a 100 mV noise voltage 
source. To keep it simple, let's assume that path impedance is 20 j0 ohms. 
CM current would be 5 Ma. This is reasonable according to measurements made 
here with a nasty SMPS I have.

If we added a single 1 inch long piece of 73 material, we would add 100 ohms 
or so additional impedance to the path. Current would now be .1/120 = 0.83 
mA, instead of .1/20 = 5 mA. This is almost 16 dB attenuation from a single 
bead!!

Now there can be extreme examples where we might need 100's of beads to 
produce a significant change, but those cases are much more effectively 
solved by the addition of a ground or altering of a cable's length. Those 
cases would be specific to very special cases, and not likely at all to 
appear in a receiving cabling system. The problem in life comes in when we 
focus on extreme examples as rules, when they uncommon cases.

Let's look at an uncommon case.

Let's assume we have a half wave of cable with a shield surge impedance of 
hundreds of ohms to earth, and zero external loss resistance at cable ends 
or along the cable. We try to add beads in the exact center. Now we have two 
1/4 wave transmission lines in series at each side of the beads, so the 
beads are at a point where the impedance is infinite. Now no matter how many 
beads we add, the common mode does not change.

This shows why there has to be some rational tempering of extremes we pick 
when we tell people what is necessary to solve a problem.

There is a practical litmus test sense to the notion we might commonly need 
hundreds of beads.

If such was a general case, very few of us would be successful adding beads 
to anything. Most people see very large effects even with small numbers of 
beads, even the WRONG beads, because the series impedance of a system is 
generally pretty low. This would especially be true on lower bands between 
pieces of desk equipment. Common mode impedance HAS to be reasonably low at 
source ends, or common mode current would not be an issue! The violation of 
this would be a very long transmission line with high shield to surroundings 
impedance and low loss, or a source driving the cable with hundreds of 
volts. In that case it would be better to fix or alter the system, rather 
than purchase and install hundreds or thousands of beads.

73 Tom