[TowerTalk] Understanding Coax Loss Measurements

[jimlux]

On 11/28/20 3:43 AM, Kirk Kleinschmidt via TowerTalk wrote:
> Hi, gang,
> I'm trying to characterize some coaxial cables and various connectors, and the results from my two measuring methods aren't reconciling well. My understanding of the tools is also less than complete. :)
> Method 1: Transmission mode sweep with MiniVNA.
> Method 2: RF Power meter and TX. Calibrate: Connect TX output to RF power meter, then through the test length of coaxial cable, then to dummy load. Carefully set TX output to to 10 W. Swap test cable and power meter, so TX output now goes to test length of cable, then to power meter, then to dummy load. Record power levels of between 7 and 9.1 W with TX power set to same level as "calibration " run, but with meter now placed after test cable length.
> 
> Testing 100-foot lengths of RG-6, RG-11 and RG-213.
> Problem 1: When sweeping 50-ohm cables from 1-30 MHz, the insertion loss curve/line gently slopes downward, indicating greater loss at higher frequencies (expected). When sweeping 75-ohm cables, the insertion loss line slopes downward a bit with increasing frequency, but the line is a shallow sine wave and not a straight descending line. Averaging the peaks and valleys of the descending sine wave produces results that agree reasonably well with published loss specs.
>  From VNA, loss values for 100 feet @ 30 MHz are: RG-6, 1.6 dB (avg of sine wave peak values); RG-11, 0.9 dB (avg of sine wave peak values); RG-213, 1.1 dB; RG-8X, 1.5 dB.
> With the power meter, loss values are RG-6, 3 dB; RG-11, 1.9 dB; RG-213, 2.8 dB
> As you can see, these methods don't line up well.
>


Did you set the reference impedance (or calibrate) for 75 ohm cable?

You're probably seeing the reflections from the ends of the cable. I'll 
bet you also see wiggles in the S11 magnitude as well.

How far apart are the wiggles in frequency? 100 ft at 75% velocity 
factor is about 40 meter electrical length or around 7 MHz, so every 7 
MHz you'll get peak (where the reflection just happens to line up), etc.
(this is for a S21 measurement.. if you're doing a S11 measurement, with 
the far end unterminated, the "path length" for the measurement is doubled.)