Towertalk
[Top] [All Lists]

Re: [TowerTalk] Measuring coax loss with MFJ-259B

To: Pete Smith <n4zr@contesting.com>, towertalk@contesting.com
Subject: Re: [TowerTalk] Measuring coax loss with MFJ-259B
From: Jim Lux <jimlux@earthlink.net>
Date: Fri, 24 Nov 2006 08:54:37 -0800
List-post: <mailto:towertalk@contesting.com>
At 07:57 AM 11/24/2006, Pete Smith wrote:
>Last summer, I put in 2 x 190-foot runs of LDF4-50A from the base of 
>my tower to alongside the house.  From there, I run RG-213 about 45 
>feet into the shack.
>
>I decided to measure the loss in these feed-lines, so as to have a 
>point of reference, using the Coax Loss function of the 
>259B.  Measuring at around 30 MHz, I was surprised to read large 
>cyclical variations in the loss reported with relatively small 
>changes in frequency.  For example, one of the feedlines (they have 
>slightly different-length "pigtails" on the tower end) shows as much 
>as 2.0 dB loss at ~30.3 MHz, but at 30.1 the reported loss is 1.2 
>dB, and at 32 MHz it is back up to 2 dB.
>
>Naturally, I was expecting a gradual increase in loss with 
>frequency, but this change seems both abrupt and cyclical in 
>nature.  Anyone have any clues on what is going on?  Defective meter 
>(it otherwise seems to work fine)?  Defect in the flexible coax 
>causing an impedance mismatch?



I would suspect it's that with certain lengths, the impedance the 
meter is seeing varies, and that changes the measurement precision.

The MFJ meter is basically a bridge type reflectometer, and you're 
trying to determine coax loss by measuring how much the cable 
deviates from a perfect reflection.  At different phase angles, the 
precision of the meter will be different.  (it's probably a lot 
better with resistive terminations than with reactive, but I'm just guessing)

You could try an experiment..  Create a load with a known mismatch 
(like a 100 ohm or 25 ohm resistor) and get a 5 meter hunk of  coax 
with a connector on one end .  Then, hook your mismatch at the other 
end with short clipleads.  Measure.  Then, chop off a few inches 
(which will rotate the phase of the mismatch), measure again, and do 
this til you've "spun around the smith chart"  I'll bet the meter 
reads more accurately at some phases than others.

I think W8JI's website has a block diagram of the innards of this 
meter, and one could probably figure out the details.

There have been some articles about accuracy of these meters in QST, 
QEX, and the antenna compendiums, but I don't think they've covered 
this particular situation, but more focussed on whether the SWR 
number is correct.  As amateurs get more sophisticated with 
measurements (facilitated by nifty devices like this, as well as the 
computer based VNAs from N2PK, etc.), I would expect some articles 
(or websites) that cover these sorts of things.

It's sort of an interesting thing, because in the pro world, 
measurements these days are done a different way (using things like 
Agilent PNAs or 6-port analyzers, etc.) and the calibration and 
measurement uncertainty calculations are complex enough that a 
computer program is used to calculate it all, based on the actual 
instrument being used (the uncertainty depends on the values of 
calibration factors).  And, if you go to the pro literature for old 
style reflectometer bridges, it was in the "way before computers" 
era, so a device like a MFJ-259 with diode detectors and a/d 
converters wouldn't have existed, so the assumed error sources and 
probability distributions are different.




_______________________________________________



_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

<Prev in Thread] Current Thread [Next in Thread>