UBM Tech
UBM Tech

Frequency Domain Reflectometry Locates Elusive Waveguide Faults

-October 01, 1999

Telecommunication systems that incorporate long lengths of cable or waveguide can be very difficult to fault-find if the cable or waveguide is damaged at some point. Electricity and telephone companies commonly use the technique of “time domain reflectometry” to discover why a length of cable no longer works and to determine the position and severity of faults. A time domain reflectometer sends out a very sharp DC pulse, with rapid rise and fall times, and analyses the signal reflected back from the discontinuity.
Read about some
Cool Microwave Anecdotes

This technique, however, is of no use for locating faults in waveguide runs. Waveguides pass signals over only a relatively narrow frequency band due to their high-pass characteristics, which means that waveguides will not propagate DC pulses. Thus, time domain reflectometry, so common and useful at lower frequencies, is no use at all for testing waveguides.

Frequency domain reflectometry, however, is an alternative technique that has successfully evolved to solve the problem of locating the position of faults in waveguides. The method requires the use of a swept signal within the operating frequency band of the waveguide. The set-up launches a signal into the waveguide and a detector situated at the launch end picks up both the transmitted swept signal and the signals reflected back from any faults along the way. The magnitudes of reflected signals relate to impedance discontinuities, and indicate the size of each discontinuity.

As an example, a “Q band” waveguide has mechanical dimensions that propagate microwaves over the range 26 to 40 GHz. This waveguide passes signals at 26 GHz with virtually no attenuation (loss), but will pass no signals below approximately 21 GHz. Above 26 GHz the waveguide’s insertion loss increases steadily, but 40 GHz signals pass with little degradation. Transmission behaviour above 40 GHz is erratic.

Frequency Domain Traps Faults
The hardware set-up you need to perform frequency domain reflectometry requires a synthesised swept frequency source and a scalar network analyser (see Figure 1). Previously, you would have to connect up individual test units to make this measurement, but now these units have been integrated into one instrument in IFR’s 6840 microwave test set. The instrument also embodies a spectrum analyser (see “Microwave System Analyser Is 3-in-1”).

TME9910ET1FIG1.gif (14580 bytes)
Figure 1. A frequency domain reflectometer requires a synthesised swept source, a scalar network analyser, a directional coupler or bridge, and detector.

Frequency domain reflectometry works because faults in the line generate resonances between the reflected and transmitted signals. Over a broad frequency range there will be many resonances, giving rise to many periodic ripples. The frequency spacing between the ripples contains the information you need to locate the position of the fault on the transmission line (co-ax or waveguide) (Ref 1).

The measurement system then subjects the detected ripples to a Fast Fourier Transform (FFT). The FFT outputs pulses that you can view on the scalar analyser to display the location and relative magnitude of the fault. You can use many mathematical and physical techniques to improve the measurements (Ref 2). The most significant technique for waveguide measurements is known as “warped sweep”. This type of sweep uses a non-linear, rather than linear, sweep and compensates for the fact that wavelength does not change linearly with frequency in a dispersive medium such as a waveguide. Using a warped sweep results in a much sharper and clearer image of a discontinuity.

The display shows a graph with distance along the horizontal access and magnitude of the reflection coefficient of the discontinuity on the vertical axis, generally expressed in return loss (dB) — see Figure 2.
10et1fig2.gif (16930 bytes)
Figure 2. A frequency domain reflectometry display shows the location and magnitude of faults in waveguides or coaxial cables.

Steve Gledhill is a test industry consultant now specialising in sales and marketing projects. He previously held various technical management positions with IFR, Stevenage, UK.

1. D L Holloway, “The Comparison Reflectometer”, IEEE Transactions on Microwave Theory and Techniques, Vol. 15, No. 4, April 1967.

2. Robert MacRae and George Hjipieris, “Use Scalar Data to Locate Faults in the Time Domain”, Microwaves and RF, January 1989.

Microwave System Analyser Is 3-in-1

The 6840 series of microwave system analysers embody three instruments within one unit: a synthesised swept source, a spectrum analyser, and a three-input scalar analyser.
  The synthesised source sweeps independently of the tuned frequency of the spectrum analyser to provide a full-range microwave spectrum analyser and tracking generator, up to 24 GHz, with variable offset. The instrument combination greatly simplifies the measurement of frequency translation devices, such as mixers, converters, multipliers and dividers. You can determine frequency converter gain or loss rapidly by simply using the source and the tuned input of the spectrum analyser. Using this tuned input also gives dynamic range in excess of 80 dB.
  Contact IFR, Stevenage, UK. +44-1438-742200,

Cool Microwave Anecdotes

A Winter’s Tale — A major UK telecommunications company experienced an intermittent problem with a microwave link. The link worked well throughout the summer, but on certain days in the winter the system failed. The company carried out tests but the fault remained.
  Eventually, after extensive testing, the company discovered the cause of the fault. At some time, water had entered the waveguide system and partially filled the waveguide at a bend. The water was not deep enough to prevent propagation of the microwave signal under normal circumstances, but on cold winter days the water froze and expanded, and the waveguide stopped propagating! Regrettably, the 6800 with fault location was not available at that time.
  A Nip in the Air — A particularly successful 6800 beta-site trial was carried out at Andrew Corp, Orland Park, near Chicago, USA. The assessors presented a length of Heliax flexible coaxial cable on which to identify discontinuities. The length of Heliax was already defective because there was a fault around its mid-point.
  The 6800 located the fault rapidly enough to arouse the interest of an audience of sceptical engineers. As a further test, one of the engineers left and returned a few seconds later with a large pair of pliers, with which she proceeded to pinch the cable at various points. Instantly, everyone could see the position of the discontinuity move, and the harder she pinched the higher the peak on the display.

Loading comments...

Write a Comment

To comment please Log In