Frequency tracked measurements

Kingdom Pty Ltd
Saturday, 11 June, 2005



Frequency selective devices such as amplifiers, filters, directional couplers, attenuators and the like are characterised by their performance versus frequency.

Measuring these characteristics during development or for inspection is performed by various methods, most are frequency swept. Sweeping the frequency provides a convenient comparison between the specified and measured characteristic.

The high end of swept frequency measurements are performed by vector network analysers. These systems provide phase and amplitude versus frequency of device characteristics, such as S-parameters.

The cost of these analysers, especially at the high end of the microwave frequency band, is very high and can run into the hundreds of thousands of dollars.

These are used mainly by research and development companies, for developing frequency selective devices.

When the required characteristic is amplitude versus frequency only, simpler and cheaper methods are available, with tracking signal generators being the most common.

Tracking generators are signal sources with their frequency controlled by a frequency sweeping receiver or spectrum analyser. The typical block diagram of such a system is given in Figure 1.

The measurement of the amplitude versus frequency characteristics of the device are typically relative values such as attenuation. The measurement involves a two-step procedure:

  • The tracking generator is connected directly to the receiver or spectrum analyser and a base line or reference level is obtained by sweeping through the frequency band of interest. This base/reference line is stored in the receiver/spectrum analyser's memory.
  • The amplitude settings of the tracking generator are left unchanged from step 1 and the device is inserted as in Figure 1. The measurement is now repeated. The result is stored in the receiver/spectrum analyser's memory.

The ratio or difference in dB between the results of step 2 and 1 are the desired characteristic, amplitude versus frequency of the device.

The accuracy of this type of measurement is governed by the accuracy of the amplitude measurement of the receiver or spectrum analyser, which is typically ±2 dB, but more so by the accuracy of the frequency match between the tracking signal generator and that of the receiver or analyser.

In the case of the setup of Figure 1, the frequency match accuracy is governed by the accuracy of the frequency of the internal oscillator in the tracking generator, f1.

This accuracy requirement is especially important when measuring narrow band devices such as filters and especially those with steep slopes, such as bandpass or bandstop filters. Inaccuracy in the frequency match (RF in versus the tuned frequency of the receiver or analyser) will result in erroneous results in the measurement of the steep slope characteristics.

The accuracy of typical spectrum analysers is ±2% and that of quality tracking generators, about the same, so filters with slopes of 70 dB/decade of frequency, such a frequency accuracy may result in unacceptable errors, which may cause a good filter to be rejected, as in Figure 2.

In this example, an inaccuracy (Df) of ±2% in frequency, or a total of 4% (for the sum of the inaccuracies of the two measurements, calibration and test), will result in an amplitude error of [image]:

When this inaccuracy is added to the inherent inaccuracy of the amplitude measurement of the receiver or analyser of ±2 dB, this error may be unacceptable, especially when the tolerance of the specifications of the device is tighter than the measurement error.

The capability of the Dynamic Sciences International (DSI) frequency tracking technique in DSI receivers solves this problem by using a frequency tracking technique that is insensitive to the accuracy of the tracking signal generator and relies only on the frequency accuracy of the receiver.

The setup for a swept frequency tracking measurement using the DSI method is given in Figure 3.

The frequency of the receiver is stepped according to the requirements of the measurement. In the case of frequency tracking measurements, the receiver dictates the frequency of the signal generator, using a GPIB data link to set the required frequency.

The signal generator then generates the required frequency with the accuracy of the signal generator, which need not be very accurate. It will be between fo-Df and fo+Df.

The receiver then step-sweeps across the desired frequency, covering the expected location of the centre frequency of the signal generator and provides a reading of the amplitude with the accuracy of the receiver, as depicted in Figure 4.

The resultant plot over the desired frequency range is that of a 'picket fence' rather than a linear curve as obtained by other methods of swept measurements.

This display can be programmed to appear linear by mathematical interpolation, as in Figure 5.

With the DSI technique, swept frequency tracking measurements can be made with a high degree of frequency accuracy, without the need to use an expensive tracking signal generator. Any reasonably stable signal generator with a GPIB control will suffice.

Any quality EMC laboratory and RF product development establishment requires sensitive measurement equipment in a wide frequency range, for testing and quality control of its products and purchased components for its products.

The receiver provides both the measurement capability required for such applications and the frequency tracking capability in the same receiver, as a cost-effective solution.

Related Articles

Optimising digital RF communication systems

Alignment techniques can dramatically improve (or degrade) performance of the radio equipment you...

Measurements on the go

Here are the basics of four measurements you can perform on the go — RTSA, noise figure,...

Interference elimination on the air-bands

Tracking down an intermittent source of interference has safeguarded operations at one of...


  • All content Copyright © 2019 Westwick-Farrow Pty Ltd