Web lists-archives.com

# Re: Time Domain Reflectometer (was Re: internet outages)

• Date: Mon, 24 Dec 2018 15:37:18 -0500
• From: Gene Heskett <gheskett@xxxxxxxxxxx>
• Subject: Re: Time Domain Reflectometer (was Re: internet outages)

```On Monday 24 December 2018 14:10:44 rhkramer@xxxxxxxxx wrote:

> Hmm, intended to send this to the list, sent to John Hasler, only, who
> did respond -- I hope he will copy his reply to the ist (or tell me it
> is ok to do so).
>
> On Monday, December 24, 2018 08:21:57 AM John Hasler wrote:
> > As to TDRs, if you can get by without actually seeing pictures of
> > all the impedence bumps
> >
> >
> > you can get by with a fast counter, a high risetime
> > pulse generator, and a couple of fast comparators.  Maybe \$20 at
> > Digikey.
>
> Can you elaborate a little on how that would work?
I can imagine the first fast comparator would enable the counter,
previously held in reset and the second comparator is then enabled, and
its threshhold adjusted for a stable stop of the counter, giving the
number of input cycles between the start and the stop. Decent calculator
math would then give you the distance to the major impedance disturbance
that caused the echo.

You need a GHz (at least) signal source to count, the higher, the more
accurate. The delay of course is 2 way, out to the fault and back, so
convert that to distance one way with a /2.000 after using the usual
hambooks 984/frequency derivation to get the wavelength in feet IIRC.
Ideally the frequency should be a wavelength short enough to give decent
accuracy because this method will only give you the number of cycles as
an integer. You send the pulse and wait for the echo to come back and
stop the counter. Multiply the wavelength by the counter to get how far
away the fault is, multiply that by the PV of the line being tested and
divide by 2 to get the one way distance.

Clear as mud, right? 2000 feet of 6.125" inch line can be fun due to
losses weakening the return signal. Specially a fast one like a tunnel
diode might make. And no bets at all for a broadcast antenna because the
pulse is not frequency shaped to match the antenna's operating
frequency, so from the fine matcher on into the element array, its a
broadband mess that splatters all over the TDR screen. Seeing that on a
real TDR with trained eyeballs seems to be the only way to tell that,
from a cracked and burned up elbow 50 feet below it at the tower top to
antenna feed connections.  And that 50 feet might be a weeks work for a
tower crew with some antenna styles. And of course it always happens in
bad weather, making the high steel work dangerous because there may be a
6" thick layer of ice on everything 2000 feet up.

That broken elbow might be the starting point, but the fire then moves
down the line towards the power source and may burn up 600 feet of line
before the transmitters VSWR protection circuits can shut it down. Been
there, done that, several times at what was once NETV's KXNE transmitter
on UHF channel 19. Rosemounts rime ice detectors buried on the antenna
structure had an average lifetime of a year. They cost quite a bit, but
their failure mode was always safe mode, so you didn't know a thing
until the main beam power breaker opened in response to the rising VSWR.
By then you were out 10 to 50 grand and several days to get a crew
rounded up to work on it, to get it back on the air. More than once in
that decade we had to let them sit around, or go do another job while
the ice went away. One time litterally tons of it had formed a flag
about 8 feet long on the downwind side of that storm.  Cleaning 600 feet
of that line and rebuilding it took time, rags by the pickup load of
bags, alcohol in several 5 gallon buckets, and teflon parts aren't free
either. I'm glad I'm retired now.

Now, back to the asinine arguments about boobs. Is this our regularly
scheduled programming for this week? :)

Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

```