I had an open wire in a ten pair cable that ran control signals out to
some antenna switches about 500 feet away.

I had the open wire disconnected, so I connected a clip lead to BNC
cable to my MFJ259 and used the distance to fault set at 0.6 velocity
factor.

I clipped the hot to the open wire and the shield to the shield of the
control cable. Going from any impedance dip frequency to the next
closest impedance dip frequency, the MFJ-259 showed 75-feet.

I paced off 75 feet. When I moved an old junk cabinet the painters had
moved, there was a terminal block I forgot about having. There was the
open wire. Problem solved in minutes.

I've also used it to follow lines underground with the aid of a receiver
and a small loop on a stick, and to find lengths of dead power lines
using the same impedance null spreads.  (I actually am the one who
implemented this function in the MFJ-259. Since nulls repeat every 1/2
wave, all you do is shift frequency and see how far apart the closest
spaced null-frequencies are.  That represents the open or short distance
when backed out of the electrical half wave found from F1-F2.) The MFJ
will get a 1000ft line within a few feet, if the correct velocity factor
is used. It is more accurate than pulses because all you do is look for
minimum impedance and the crossing of zero reactance at each.

So here is my question....

How does the NanoVNA TDR work? Is it actually a pulse and listen for
echo, or is it a sweep?  It should work for this also, right?
Is the NanoVNA a true TDR or something else?

73 Tom

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 24/08/2023 01:26, Tom W8JI wrote:
> How does the NanoVNA TDR work? Is it actually a pulse and listen for
> echo, or is it a sweep?  It should work for this also, right?
> Is the NanoVNA a true TDR or something else?
>
> 73 Tom

Tom,

Most NanoVNAs use the frequency domain sweep equivalent of the pulse.

Broadly, the maximum frequency will set the equivalent of the rise-time of the
pulse and hence the distance resolution (higher frequency, shorter pulse, fine
distance resolution), and the lowest frequency will set the maximum length
which can be resolved. That may not be strictly accurate, but you get the idea.

Yes, it should work for your task, but be careful not to overload the NanoVNA
port e.g from static on such a long cable or other signal pickup.

73,
David GM8ARV
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv

On 8/23/23 5:26 PM, Tom W8JI wrote:

> So here is my question....
>
> How does the NanoVNA TDR work? Is it actually a pulse and listen for
> echo, or is it a sweep?  It should work for this also, right?
> Is the NanoVNA a true TDR or something else?
>

Sweep frequency and do inverse FFT.

yes, it works fine for this kind of thing. Since you've got a "long
pulse" the SNR with the sweep technique is much higher than the pulse
technique (where you've only got the energy in the single pulse, unless
you do averaging).

With this technique you can get pretty fancy in the modeling (e.g.
cables where the impedance varies along the cable) too.

I haven't tried it, but I'll bet with a reasonably broadband antenna,
you could tell how far a corner reflector is. I've done that with a lab
VNA set to TDR mode and there's not a huge difference.

And you can measure to very much smaller than the point spacing by
looking at the phase changes.

An idea or question came to my mind. What if you estimate the velocity coefficient wrong, What if you then have access to both ends of the cable. You''l get either too long or too short length for cable? Is it still possible to calculate the fault distance (relative to length perhaps, if not absolute number)

Yes, you should be able to get very close doing that. You will then know
the electrical length and already know the physical length.

On 8/24/2023 5:22 PM, Leif M wrote:
> An idea or question came to my mind. What if you estimate the velocity
> coefficient wrong, What if you then have access to both ends of the
> cable. You''l get either too long or too short length for cable? Is it
> still possible to calculate the fault distance (relative to length
> perhaps, if not absolute number)
>

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com