Design an undergroud cable fault locator unit

There are a number of basic ways to do TDR however the reality of doing them in practice is difficult (you need to know how to design avalanche circuits for one thing).

however in principle what you do is send a fast rising edge down the cable as it has not reached the termination or load it will have one voltage on the line. When it hits the load or termination some or none of it will be reflected depending on the impeadence of the load. This reflects back up the line and is seen as a step transition on the voltage on the line at the generator. Again if the generator is not correctly matched a step wave front will travel back down the line and bounce back etc. Thus eventualy the exponentialy reducing steps will appear to give a constant voltage.

provided you only have a single fault on the line then you can use the "step" method.

However in practice life is never simple and each and every connection and bend in the cable will reflect some sort of signal. Thus some TDR systems use a high energy very narrow width pulse (think 500V 100pS wide) pulse or series of pulses, others use JPL ranging or Gold code systems and cross corelate each return signal out ussing the equivalent of Direct Sequence Spread Spectrum (dsss) systems (see code division multiple access to see how it works with multiple cell transmitters which is effectivly what each fault is).

No specification for time resolution or e.g. complexity of the connected cable system has been said yet.

In practice, a lab standard fast pulser (e.g. 1 ns rise time) would be sufficient for many tests. I remember a TDR instrument dedicated for fault localization in TV antenna distribution systems, that been manufactured in Germany in the 70th. Of course, the voltage had to be suffcient low not to damage connected TV receivers, so it had to be restricted to a low voltage step. But it gave a visual representation of a network with multiple regular discontinuities and a failure. I agree, that you won't be able to localize multiple faults in some cases, but I don't think, that this is a common problem.

A manufacturer offering specialized TDR systems, will want to set apart from competition and start to offer advanced features. But that doesn't mean, that you can't work with a basic instrument.

100 ps pulses would "melt" after a few 10 meters due to cable dispersion.

Today i researched frequency domain reflectometry techniques. There are some different techniques. i.e frequency-modulated continous wave (FMCW) systems, SWR reflectometry systems, phase detection FDR (PDFDR) systems...
SWR systems analyze superposition of incident and reflected waves. And they locate faults with number of ripples
Phase detection system multiplies the two signals and obtain signals with frequencies at sum and difference of incident and reflected signal frequencies. They locate faults with phase differences

But i couldn't find or get it that which one is better at microwave frequencies for cable fault location. Which system do network analyzers use?

A VNA (in 1-port mode) measures magnitude and phase of forward and reverse wave and calculates S11 respectively complex impedance versus frequency. By applying a fourier transformation, the result can be converted to a TDR-like time domain representation (within the transformation's time/frequency limits for uniqueness).

I didn't exactly understand your explanation of "PDFDR", partcularly the frequency mixing point. Can you give a link?

Frequency-domain reflectometery for on-board testing of aging aircraft wiring

FvM, you can reach to article about pdfdr.

I am only want to learn frequency domain reflectometry techniques for location cable faults at microwaves frequencies. Which fdr techniques include my subject?
Which type of cables are used for microwaves? Only coaxial?

FvM, you can reach to article about pdfdr.

Click to expand...

Thank you. From a short review, the said PDFDR looks like a simplified version of the technique used in VNAs, only measuring the real part of S11. Apparently it's sufficient for the intended purpose.

Which type of cables are used for microwaves? Only coaxial?

Click to expand...

Effectively yes. But reflectometry can be used also for non microwave cables, as in the PDFDR paper. The term microwave is somewhat vague, it would be clearer to specify a frequency. In the usual understanding, microwave covers a frequency range from 300 MHz to 300 GHz.

Thank you for your answer FvM.
In general time domain reflectometry method uses the time delay of reflection from fault points to estimate the location.
Frequency domain reflectometry method uses phase delay for it. Is it correct?
You wrote that vna's can calculate s11 from magnitudes and phases of incident and reflected signals. How does it do this? As i know, magnitude doesn't change at reflection. What is the formula for this calculation?

I need some tdr fault location experimental note which were realized in labs. How can i find it? On the internet there are some tdr experiment procedures but most of them about soil moisturing or bridge scour

underground cable fault locator, while time domain reflectometry may tell you how far from one cable end the fault is (if you know the characteristics of the buried cable - and they are constant), it will not guide you to the cable fault (i.e. the spot on the ground just above the fault). For this you need a "cable thumper" used in the power industry - whereby a large cap is charged and applied to the available cable end (having shorted or isolated the other available end if you can) approx once a second - the likely cable path is then walked along until one can hear the thump near the spot where the cable is open or short - or alternately a small loop antenna is used to pick up the radiated impulse from the cable fault. This is a very quick method used to find buried power cable faults.
Regards, Orson Cart.

I have a question about TDR? What is the relationship with incident pulse width or rise time between characterization bandwidth of the transmission line (coaxial cable) ?
i.e. I want to test my cable until 500 MHz or 1GHz frequencies, how should i calibrate the incident pulse width and rise time at pulse generator? Are there any formula?

As a rule of thumb, 1 ns risetime corresponds to 350 MHz bandwidth. Pulse width is rather related to energy content at lower frequencies.

Thank you for your helpfully answer FvM!
As i understand that if i want to test a coaxial cable in order to find faults between 0-350 MHz frequencies, i should have a pulse with 1 ns rise time. i.e for 1 GHz bandwidth 350ps rise time. But is the pulse with 350 ps rise time practical or workable in testing long cables for getting reflections? Frequency dependent cable attenuation is a important problem for long distances. Or if we adjust the pulse width considering cable's length, we can get the reflections?

In some application notes from companies like Anritsu which uses FDR, is said that traditional TDR method can only locate DC faults and not RF faults. But companies which uses TDR for example MOHR says that Tdr method is better than FDR for testing cables at rf/microwave frequencies.
Which one is true? If i want to locate faults in a coaxial cable at rf/microwave frequencies, which method is more sufficient?