Hackers Measure Cable Lengths with Time Domain Reflectometers

[android] has built up a fast edge pulse generator for time domain reflectometry (TDR). TDR is a neat technique which lets you measure cable lengths using electrical signals and can also be used to locate faults within the cable.

TDR works by sending a pulse down the cable. When the pulse reaches the end of the unterminated cable it is reflected back to the source. By monitoring the delay between the original pulse and its reflection you can determine the length of the cable. We’ve seen projects that use TDR before, and it’s often used in telecoms industry to locate faults in long cable runs.

You can try TDR in your lab using only a scope to observe the delay and a function generator to create the pulse. However, the technique works a lot better with pulses that have very fast rise times. So [android] built a fast edge pulse generator based on [w2aew]s design. Then added googly eyes for good measure. His build works great and is a nice demonstration of the technique.

12 thoughts on “Hackers Measure Cable Lengths with Time Domain Reflectometers

  1. We used to call this Pulse Echo Testing (PET). You can see much more than just the other end of the cable. There is an impedance change with every connector or splice along the length of the cable and you can see a small reflection from each of these.

  2. A speed factor of 0.66 is very common. However, that’s only half the truth. Speed of a wave depends mainly on the dielectric material, so here’s the table from https://en.wikipedia.org/wiki/Coaxial_cable#Standards:

    VF is the Velocity Factor; it is determined by the effective \epsilon_r and \mu_r[35]

    VF for solid PE is about 0.66

    VF for foam PE is about 0.78 to 0.88

    VF for air is about 1.00

    VF for solid PTFE is about 0.70

    VF for foam PTFE is about 0.84

    The forum post says “I see a VF of 0.81”, and he’s using RG59. Since RG59 is considered to be relatively cheap cable, I’d guess it’s foam PE (instead of the solid PE that everyone using 0.66 assumes). Other than that, guessing a rise time of ~1.333ns, edge bandwidth is ~1/1.3333ns = 3/4 GHz = 750 MHz. There might be some group velocity dispersion (measuring these things properly is not as simple as the Ham operator’s video suggest), but luckily that’s counteracted by the fact that the amounts of signal that are slower are dampened by up to ~35dB/100m, so that’s 2*5*0.35dB = 3.5DB less amplitude (dB amplitude has a factor of 20 instead of 10 of the power dampening dB) over 5m cable, so if looked at on a linear display, this will dampen these later signal parts by a solid factor of 2.

    Of course, the original video from the article is using a 350MHz scope. Now, this means that even if that analog oscilloscope has, if visual compensating the dampening (which of course will totall influence the length you “think” you’re observing, because the “visual correlation” between TX and RX pulse will reach a maximum too early), a usable bandwidth of 400MHz, that means that your temporal resolution is 1/400MHz = 2.5ns; at the speed of light, and considering the two-way nature of the path, this yields a maximum mathematically achievable accuracy of 37.5cm; scale that number with your VF^-1. So please don’t trust the guy who says he’s building a 40mm TDR device; the bandwidths involved are pretty ugly, and if you do something like that, you better account for wave dispersion.

    1. My best guess is that the other person meant to write 40cm. If they did actually mean 40mm then it is wildly unrealistic and they need to check their numbers again.

    2. My eyes aren’t that goo that I can see the cable but I would be surprised if it were RG-59/U (0.66) or RG-59A/U (0.78) as theses cable types are used for TV antenna’s etc and are thicker – often using ‘F’ type connections.

      More often around an electronics workshop you will find RG-58/U (0.66) which is thinner and most often uses BNC connections just like all the equipment you often find in an electronics workshop.

      Also he uses 5 x 220 Ohm resistors for the output impedance to match the cable. 5 x 222 Ohm comes to 44 Ohms and is closer (with a couple of Ohms from the gat chip) to the 50 Ohms of RG-58/U as opposed to the 75 Ohms of RG-59A/U of RG-59/U.

    1. Try this for a related sensor. This is actually very similar to how modern “guided radar” liquid level sensors work. they fire a pulse down the length of a stainless steel rod and then look at the returns to determine at what level the working dielectric fluid is.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s