Can TDR be Used for Diagnostics on a Single Wire?

[CopyOfA]

Can someone help explain why reflectometry is unavailable for diagnostics in a single wire? Or is it unavailable?

I am a little confused as to why one could not use TDR on a single wire to find faults in the wire, such as nicks or corrosion... Is there a way to construct a model whereby this testing could be performed? In other words, if I have a wire connected from Port 1 to Port 2, could I connect an additional matched impedance load in parallel to my wire?

Thanks for any input and advice!

 

[zoki85]

?
Of course, it is available: http://en.wikipedia.org/wiki/Time-domain_reflectometry

 

[CopyOfA]

I thought reflectometry was only for transmission lines. According to the traditional definition, a single wire is not a transmission line, correct?

 

[zoki85]

I thought reflectometry was only for transmission lines. According to the traditional definition, a single wire is not a transmission line, correct?

That doesn't matter, EM waves get reflected in the same way

 

[f95toli]

But it would be MUCH more difficult with a single wire (unless there is a huge capacitance to some return path for the current). It is possible to launch an EM wave along a single wire, but I've never seen anyone do if for the frequencies typically used for TDR. You also need special launchers etc so it is far from trivial.
Just hooking a single wire to the centre conductor of e.g. a BNC connector would just result in huge impedance miss-match at that interface so it would swamp any imperfections in the wire itself.
(you also have "antenna" effects to deal with, for some frequencies the wire will work as an lambda/n antenna).

 

[zoki85]

Normally, other terminal of the signal generator must be grounded (directly the best) or connected to other structure of high electrical capacity. Otherwise, you can't send signals along single wire.

 

[CopyOfA]

What I'm thinking is something like this:

(The BNC connectors are 50Ohm impedance RF cables.) Is something like this going to produce any good results? I understand there will be a large impedance mismatch between the BNC cables and the wire, so is there a way to overcome this?

 

[sophiecentaur]

Can someone help explain why reflectometry is unavailable for diagnostics in a single wire? Or is it unavailable?

I am a little confused as to why one could not use TDR on a single wire to find faults in the wire, such as nicks or corrosion... Is there a way to construct a model whereby this testing could be performed? In other words, if I have a wire connected from Port 1 to Port 2, could I connect an additional matched impedance load in parallel to my wire?

Thanks for any input and advice!

In principle, a TDR can show you any changes on a line. The problem with looking at the return from a line that does not have a well defined 'return path', such as coax or strip line is that there will be severe dispersion of the return pulse. This is because of the multiplicity of path lengths for the returned signal.
If you want to test the unscreened line in your diagram, you might find it best to run it over a ground plane, at a constant height, to reduce the line impedance to something as low as possible ( something arbitrary but constant; perhaps 200Ω?) and then terminate it was well as you can. The trace would then show a step up from 50Ω at the transition point and a (hopefully) horizontal trace at the new impedance value, with blips on it.
However, one reason for using a TDR is when you cannot actually see the conductor you are investigating. I would think that a visual inspection would be far more fruitful than trying to use a TDR to identify lengths with surface corrosion and consequent series resistance.
BTW, For reflectometry, is it not normal to have a termination on the end of the line, rather thanking into another port on the reflectometer? The setup you show would involve an embarrassing reflection where the wire goes back inside the coax. Also, BNC connectors may not be very suitable for situations where time domain reflectometry is used. What frequency is this equipment doing to be operating at?

 

[CopyOfA]

Right, I am not trying to do TDR for a small wire, as shown in the picture. In that case, it would be easier to simply inspect the wire visually :). I am trying to understand how to apply TDR to a single wire for other purposes like long electrical connections or coils.

The problem with looking at the return from a line that does not have a well defined 'return path', such as coax or strip line is that there will be severe dispersion of the return pulse. This is because of the multiplicity of path lengths for the returned signal.

What do you mean by the multiplicity of path lengths for the returned signal? How are the lengths different?

What you are suggesting would be to place another wire in parallel with the wire that is being tested, where the second wire would connect to the ground of the RF impedance cables, correct? Must these wires have the same impedance?

I have no preference for the operational frequency at this point. I would probably tune it for the type of defect via the skin effect. Also, I think you are right about the BNC connectors... I am using RF cables built for VNA modules like the one shown.

 

[sophiecentaur]

I suggested a ground plane, actually - not a second wire. It would give half the impedance of a pair. Imagine the way a coax is connected to a pcb and replace the microstrip with your wire under test.

My comment about the multiplicity of paths is because the wave is no longer constrained inside the coax. Currents will flow along the outside of the coax. There will be a massive discontinuity and equal and opposite currents will flow on the continuing centre conductor and the outside of the screen and waves will propagate in a very unpredictable way. If you want to launch a wave along just the centre conductor, you need to use a horn transition (see Goubau Line links).

It's hard enough to make a good narrow band transition from one line configuration to another. Doing it for a step waveform is more problematical.
If you actually have access to the equipment in the picture then I suggest you try an experiment and see what you get. I would think the trace could be very confusing.

I would suggest that the series loss of a corroded line could be best studied when it's included inside a well terminated length of coax. the time domain is excellent when it's the right tool for the job. :)

 

[sophiecentaur]

. . . . . . Oh yes, and the reflections of a step function from a horn transition would need to be seen to be believed. You could never get any sense at all from that trace. TDR technology needs very specific applications.

 

[jim hardy]

I've tried it in a power plant.

Results are mixed because wires and structures nearby all amount to discontinuities. If you're lucky you can pick out the far end of the wire.Try it yourself. TDR meters have come a long way since 1973.

If your lone wire is pretty much hanging in free space, perhaps far end tied to a tree, you may have some success.

 

[sophiecentaur]

I've tried it in a power plant.

Results are mixed because wires and structures nearby all amount to discontinuities. If you're lucky you can pick out the far end of the wire.Try it yourself. TDR meters have come a long way since 1973.

If your lone wire is pretty much hanging in free space, perhaps far end tied to a tree, you may have some success.

I think that, as usual, you and I are a thinking several octaves apart. :D My reflectometer experience was with picosecond pulses and UHF components on microstrip. For power and telephone cables, TDR will tell you where the line's actually bust, I believe. (Useful in its way of course.)

 

[jim hardy]

Indeed, my TDR experience was signal wiring scores of feet long.
Best I could do was nanoseconds.

Thinking about OPs original question,

A single wire hanging free in space would be uniform along its length so ought to lend itself to TDR

But anything nearby that's conductive would be a discontinuity ?
We could detect manholes that had filled with water...

Your experience with strips sounds interesting.

 

[zoki85]

A single wire hanging free in space would be uniform along its length so ought to lend itself to TDR

But anything nearby that's conductive would be a discontinuity ?

Single long wire "free" in a space atenuates spiky,very short electrical impulse considerably due to EM radiation in space. Coax cables and transmission lines are better for TDR. Yes, sudden coupling to nearby conductive object propagating wave senses as concentated capacity or/end inductivity, and received return signal is generally different than due to broken wire or shorted wire discontiunity. I guess this kind of wire diagnostics and network "imaging" will be so advanced one day we'll be able to reconstruct transmission line surrounding in a great detail.

 

[Mike_In_Plano]

For any wave to propagate, it needs a minimum of two energy storage mechanisms which can trade off from one another. In the case of a spring, the inertia of the metal stores energy as velocity and the compression / expansion stores energy. In the case of sound, again, we have compression and inertia. In free electromagnetic waves, we have e-fields and h-fields. In a transmission line, we have capacitance and inductance (V^2 x C / 2 and I^2 x L / 2).

What are the energy mechanisms for your wire?

 

[zoki85]

What are the energy mechanisms for your wire?

Pretty similar to TEM wave propagation in a coax. with D₁>>D₂. But there's also radiation term which usually can't be neglected and Maxwell eqs should be used for proper treatment. Good read about this you can find in Jackson's Electromagnetism.

 

[Baluncore]

To test a single short wire, thread it into a metal tube using closed cell foam as the separator. It is then a coaxial cable. If you do not use an outer shield you will get a noisy trace due to the wire antenna picking up local RF signals, you will also be transmitting about 20 mW of broadband clicks.

 

[CopyOfA]

Thanks for all the input everyone. I'll try my best to respond to everyone, if possible.

I suggested a ground plane, actually - not a second wire. It would give half the impedance of a pair. Imagine the way a coax is connected to a pcb and replace the microstrip with your wire under test.

Do you mean something like this (with the other end of the DUT terminated into a SMA cable in a similar fashion):

My only experience with TDR has been with surface mount components that were arranged as:

So, I'm curious if there could be an arrangement where a single strand wire could be monitored using TDR. This wire is covered in a dielectric insulation material, as illustrated below (the center is the metallic conductor).

As a mechanical engineer, I'm a little new in the whole electromagnetic wave arena, so forgive me if what I'm asking is super basic.

It's hard enough to make a good narrow band transition from one line configuration to another. Doing it for a step waveform is more problematical. If you actually have access to the equipment in the picture then I suggest you try an experiment and see what you get. I would think the trace could be very confusing.

I do have access to a VNA. What kind of experiments would you suggest? Should I attempt to use a horn transition to launch a wave into the center conductor of a wire?