[TowerTalk] How many connectors?

[jimlux]

On 10/30/18 7:21 AM, Wayne Kline wrote:
> My old gray matter is not  100% recall on an article  in QST or CQ regarding connector loss.  The tester uses a spectrum analyzer and a TDR  with other lab equptement.
> 
> He had  up to 20  barrels ,T. and Pl259’s  with short lengths of coax  .
> 
> And the findings were that @ HF and insertion loss was minimal and the BIGEST issue was lossy CABLE
> 
> 


This all comes from folks reading data sheets and misinterpreting the 
values found there.


A typical data sheet might say "connector loss <0.03 dB"

The 0.03 isn't really a "worst case", it's a "the smallest non-zero loss 
we can measure".

In reality, the connector loss is much less, but making 0.01 dB loss 
measurements is more art than science.

A paper by Pakay and Torok, "Analysis of insertion loss repeatability of 
coax connectors" (in "The Radio and Electronic Engineer, v47, #7, pp 
315-319, July 1977) gives a table at the end of the loss for various 
connectors.  I'll give the loss at 100 MHz to save time:
SMA 0.007
N (stainless) 0.004
UHF 0.028 dB

So that 0.03 for UHF connectors (at UHF frequencies) might not be all 
that far off.


The uncertainty in loss (across multiple mates) expressed as standard 
deviation is
SMA 0.0003
N 0.0002
UHF 0.0002

(all at 100 MHz)


I note that they cite an earlier 1970 paper by Bergfried and Fischer, 
and comment that B&F got bigger uncertainties, but most likely because 
their measurement equipment wasn't as good.

You could dig up an earlier paper from Nov 1961 by Sanderson: "A new 
high-precision method for the measurement of the VSWR of coaxial 
connectors"   MTT-9, p 524  for a description of a technique using a 
slotted line, and a variety of precision equipment.  He doesn't give any 
sample data, though, just that he can drive the uncertainty of the 
measurement down below 1.0001 to 1.0003 "so a VSWR of 1.001 is well 
above the noise level"

Reading his methods and analysis, this was a fairly painstaking and 
tedious measurement (esp since he did it from DC to 9 Gc)  - he worked 
at General Radio, and they had a new connector they were pushing then.


I suspect that the average ham isn't going to worry about whether the 
match in the connectors is good to VSWR<1.001.

Precision metrology is an interesting, but frustrating, area...

When you're looking at microBels (0.00001 dB), little things like the 
temperature of the connector (much less the cables) have an effect that 
is much larger than what you are measuring.

I am always much impressed by this kind of work from the 50s, 60s, and 70s.

Jim, W6RMK