Cheap circuitry to measure cable length and resistance per unit length.

[axero]

If you connect a cable to a device leaving the other end open, a pulse sent from said device will bounce from the open end of the cable back to the device. This "echo" if you will, also occurs even when the cable is plugged in if the other end is not impedance matched (e.g. see this Wiki article http://en.wikipedia.org/wiki/Impedance_matching#Reflection-less_matching ). By measuring the time it takes before a given pulse bounces back, one can estimate the length of a cable (at least if it isn't broken).

A cable has an inherent resistance (or impedance when talking about alternating currents) which means that there will be a voltage drop. By knowing the resistance , one could gradually increase the voltage so as to compensate for this voltage drop.

So here are my essential questions:

* Is there any cheap circuitry that can conduct this type of measurements?
* Is it a viable solution to use such voltage compensating circuitry for connection of USB devices, power-over-ethernet devices, or say a smartphone to a USB charger?

N.B. Reflections due to impedance mismatched connections is a problem among high-speed data communication applications but not relevant in this discussion.

[Rory]

Watch this video from w2aew. Google search "cheap TDR" and this is the first hit.

[AndyC_772]

TDR tells you roughly the length of the cable, but tells you nothing about its resistance. It's also very difficult to measure short cables accurately this way with a simple, inexpensive circuit.

Every standard I've ever come across simply requires the powered device to be able to accept a lower voltage at its input, to cope with the voltage drop in the cable. If you have a low voltage, high current system in which this would be unacceptable, spending a fraction more on a slightly thicker cable is a better solution than using active electronics to try and measure the cable length and compensate at the supply end for the loss.

[Rory]

I agree with AndyC about standards, dynamic outputs aren't really feasible in this application. Longer runs need larger gauge wire, driving it with higher voltages is not the way to do it. Also, this has already been incorporated into power over Ethernet.

You will also find that cable capacitance rather than ohmic resistance will cause more attenuation over longer runs when it comes to higher frequency signals like USB and gigabit ethernet.

In my experience with TDR usage, cable loss can be measured by comparing the output amplitude with the return pulse amplitude, and if the generator impedance is matched to the line, the calculation is trivial. This method can be done with simple equipment. The problem with using the TDR for this is that the line either has to be hard shorted or completely open for the return losses to be calculated, otherwise part of the signal is absorbed by the load, decreasing the return amplitude.

Another method to determine the ohmic resistance of the line, if the line is unbroken and the same type of conductor for the whole length is to measure the length then multiply it by the ohms per unit length specified by the cable manufacturer or general figures determined by wire gauge and metallic composition. 

If you are going to design this type of thing into the equipment, it probably be less of a problem to incorporate level detection at the far end and a kind of handshaking between ends. This happens with some modem protocols where adaptive equalization is done to increase the effective bandwidth of the line.

[PA4TIM]

TDR tells you roughly the length of the cable, but tells you nothing about its resistance. It's also very difficult to measure short cables accurately this way with a simple, inexpensive circuit.

It depends on the TDR. I measure the lenght and delay of sma connectors. Both my TDR's can measure impedance. I can see very small impedance jumps in connectors.
The simple form to get an idea of length is a pulsgenerator and a fast scope. The faster the better and also for the pulser. You want a very fast Tr and a long periode so you get a step response (that way impedance measurement is more easy)
The other way is a semi-dirac pulse.

[JackOfVA]

Do you really mean resistance per unit length or characteristic impedance? Transmission line impedance is a function of four variables, series resistance, shunt resistance, series inductance and shunt capacitance, although if the two resistance components are negligible, then you wind up with the common formula Zo = sqrt(L/C).

If you have a sample of the cable, and a decent LCR meter, it is possible to measure L (short the far end of the cable), and C (far end of the cable is open) and then compute the impedance.  You also have the C data for a particular length and you can then measure the C of the cable in question and determine its length from the sample cable data.

There are some problems with this approach as L is not constant with frequency and a measurement at, say, 1 KHz, will not yield the same inductance value as at 100 KHz or 1 MHz in many cables. Hence an LCR meter with 100 KHz or greater frequency is desirable. 

The good thing about an LCR meter is that they are relatively inexpensive.

Also, with a TDR, keep in mind that a lot of TDRs were made for unbalanced coaxial cable but the cables you mention, Ethernet, USB, etc. are balanced and will require some additional work to obtain an accurate measurement of either impedance or length with a TDR.

A good TDR will measure with mm resolution and accuracy, and you won't come close to that with an LCR meter. However, accurate length measurements with a TDR require knowledge of the cable's dielectric constant as that determines the velocity of propagation.

It's also possible to measure cable length with a swept frequency measurement. This can be down with a sweep generator, reflectance bridge and a spectrum analyzer, or a spectrum analyzer with tracking generator or a vector network analyzer. Put a short circuit on the far end and look at reflection coefficient versus frequency, repeat with an open on the far end. It can even be done with a cheap signal generator and a receiver. Just depends on the ratio between money available and time available for the measurements.

Also keep in mind the difference between physical and electrical length.

Lastly, of course, a ruler or tape measure is the cheapest way to measure a cable length, but I assume it's in an inaccessible location.

[vk6zgo]

Put a short circuit on the far end,measure the resistance  between the two conductors.
     
If you really want the length,look up the ohms/metre spec for that cable,& calculate it.

In the old days,Telephone Techs knew the ohms/foot or ohms/mile figures for open wire pole routes,so they could measure the distance to many faults using a bridge.

They did this years before TDRs came out of the laboratory.

http://en.wikipedia.org/wiki/Murray_loop_bridge

[AndyC_772]

It depends on the TDR. I measure the lenght and delay of sma connectors. Both my TDR's can measure impedance. I can see very small impedance jumps in connectors.

Just to be crystal clear on this, are you referring to the ability to measure the DC resistance of the cable, or the characteristic impedance?

The height of the reflected pulse may well allow you to measure the impedance of the transmission line, provided you know how it's terminated (or not) at the far end, but I don't see how it tells you anything about the end-to-end resistance of the cable.

[PA4TIM]

This is my Megger that is made for Murray-varley measurements

Andy: I wrote impedance, not resistance. My digital one has a lot of options, my analog one displays in |reflection coefficient|  And yes, I can measure the impedance of the load. If I terminate it with an impedance 30 Ohm for instance. You can also measure R, L and C with it but that is far from easy and involves a lot of math. But I do not know if it can split the impedance in R and jX. I never use it like that, I must look that up in one of my books about TDR

There is a nice book from Tektronix in the concepts series. It is free for download on several locations. They are written in the 60's but contain very usefull information. This volume shows allmost all possibilities for TDR.
HP has some app notes abpout TDR to. Including a nice exaple how they measure the height of a layer of some fluid floating on water.
Today TDR is a function in many VNA's. The books of Hiebel and Dunsmore have a lot of information of the possibility's that TDR today has.

But sorry for this, I get over enthausiast if it is about TDR or VNA's  , the TS wants a cheap way and cheap DIY TDR is only usefull to find things like shorts or opens in transmissionline with a not to great resolution  (for resolution you need a very fast pulse-risetime (like <100ps) and a very fast (or sample)scope (like >1 GHz)

TDR can do more then most people know. I used it to find a defect lamp in a Xmas tree light string