[TowerTalk] The hunt for a feedline fault - update

[David Gilbert]

Yes, that was my thought also.  A TDR is pretty much a frequency sweep 
converted to time domain (and therefore to length assuming you know the 
velocity factor of the line)) via an inverse FFT.  In fact, I was 
originally going to take the measurements at different frequencies and 
either try to have my math major son throw some math at it,  or I was 
going to load the data into EXCEL and try to do an inverse FFT on it 
that way (not sure how many data points Excel needs, though ...).

But then it occurred to me that I should first see what TLW would tell 
me, and that turned out to be a usable approach.

And yes, I was worried that even the very low power of my FA-VA5 vector 
impedance analyzer might be enough to alter the fault characteristics 
and make the measurements useless.  At a couple of frequencies when I 
was taking measurements from the tower end of the feedline (the long 
side of the fault) there was some jitter in the readings and I thought 
"oh crap", but luckily I had done enough damage that nothing 
substantially changed.  ;)

73,
Dave   AB7E


On 11/8/2019 6:48 PM, jimlux wrote:
> On 11/8/19 4:00 PM, David Gilbert wrote:
>>
>>
>> By the way, I was wrong when I said earlier that TLW was telling me I 
>> had a dead short.  The graph only made it look like that, and a more 
>> careful look at the axis dimensions said that the dips for each 
>> frequency were always about 28 ohms resistive at the low power of my 
>> analyzer ... which makes more sense than a short for a fault caused 
>> by a high power RF arc over.  I find that even more impressive (the 
>> data, not my error), since it means that the location of even 
>> non-absolute anomalies (neither open nor short) could potentially be 
>> found by this method.
>
> It's essentially like doing the FFT of a frequency sweep to do time 
> domain reflectometry.. And it's very, very sensitive.  The one thing 
> that will cause troubles (with either approach sweeps or TDR) is some 
> sort of repetitive partial fault - because of ambiguities.
>
>
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk at contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk