Finished two coil choke: 
This is an image of the testing setup. Notice the leads soldered onto the SHIELD of the connectors at both ends of the cable. The test device is a Siglent SVA1015X. 
Now, the work that went into it: The purpose of this experiment is to see if it is possible to measure the RF LOSS of coaxial cable chokes, and to investigate them a little bit. The theory is that winding up a coil of coaxial cable "chokes" off RF currents on the outside of the sheath due to the inductance impedance of the coil of wire. Let's see what we can measure.
Why am I doing this?
When I transmit full power on digital modes on the 30 metre band, the furnace fan cycles on and off. Obviously I have RF on the outside of the coaxial cable and it runs right beside the furnace. So I did some reading and scrounged around for some cable. Here is my attempt. The method is to take a 16 foot length of RG8X I had handy and try a couple of different chokes, and some ferrites I had on hand.
Method:
Is to connect the OUTER connection of the coaxial cable from both ends to the OUTPUT and INPUT connectors of the SVA1015X Siglent VNA. First: Straight connection, no coils yet on the wire. The wire is strung up in my office, lifted up off of the floor. The VNA is zeroed for the current piece of wire. Sweep is from 1mhz to 50 mHz 
PNG13 Step One: Try some little ferrites I have. 10 of them. Upped the frequency to 55 mHz (to include 6metre band) PNG14 and PNG 15 
PNG13 Step One: Try some little ferrites I have. 10 of them. Upped the frequency to 55 mHz (to include 6metre band) PNG14 and PNG 15 
There is an odd GAIN above 12 mHz which cannot be there as a passive device so it must be due to the LC resonance of the coil. This single coil provides no more than 13 dB of attenuation at 10 mHz, and less below, so a marginal attenuation. Something, just not a lot. Step three. Add a second coil of half the size and number of winds in an attempt to provide attenuation for common mode currents PNG 17 
Wow, nothing changed below 15 mhz, but there is a new "NULL" maxing out at 30mhz. let's try separating the coils a little more PNG 19 
A little bit different. Let's throw the ferrites on again png 20 
It would be useful to have the max null at 14.2 rather than at 12 or whatever it is. Let's pull one wind off of the first coil. PNG 21 
Not worth a darn. Perhaps the resonant frequencies are too close between the two coils. now try six turn, 5.5 inches and try to just nudge it up. It was a mess so I went to 5 inches, six winds PNG22 
Looks like I have made a notch filter at 25mhz. Not really what I wanted. Try 7 inch, five winds PNG 23 
Well I made it six winds, and the notch is 10.8 back the coil off to five winds, PNG 24 
Pull off the ferrites, which by the way, were feed side of the coils PNG 25 
A measly 7.5 db of attenuation at 14.00 mhz. Stick the ferrites in the middle No difference. Pull it all apart and stretch out just the wire to confirm the response curve has not shifted too much in the couple of hours this took. It is hanging off of the closet door. Not too bad, and the wire is crossing in the middle so that likely accounts for the variance. PHG 26 
End experiment. Conclusion: The six inch with six winds gives some common-mode attenuation at both 10 and 14 mHz, less at 7, even less at 3.5 and below. Having the second coil really helps at 28 mHz, so I will leave it in. The Ferrites seem to help above 40 mHz, so they will stay and it will assist at 50-55 mhz. PNG-17 above will be a close approximation of the choke I will put up on the roof BEFORE the antenna tuner that drives the antenna, a Cushcraft AP8, on the roof.
