Fundamental Measurement Techniques - Network Analysis (Part Two)
Part Six of our occasional series around ‘fundamental measurement techniques’ continues to look at the basics of network analysis and some of the basic measurements considerations needed when using a network analyser (Part 2).
Network analyser channel capabilities might include sweep type, frequency, power, IF bandwidth, number of points, trigger state, averaging and type of calibration window. From the signal trace we would be looking to identify key parameters and have the ability to vary format and scale. In most cases advanced markers and trace math are benefits.
General measurements involve Electrical delay, Phase offset, smoothing, and limit testing
More specific measurements of interest could include:
- Group Delay where ripple indicates phase distortion, average delay indicates electrical length of DUT and the aperture of measurement is very important
- Directivity, which is a measure of how well a coupler can separate signals moving in opposite directions
In all of these and others there is a key need for calibration otherwise all measurement integrity could be lost. Most simply put, calibration is needed because it is impossible to make perfect hardware or if possible it would be prohibitively extremely difficult and expensive to make.
To attain good accuracy look for vector-error-corrected calibration to remove systematic errors. Systematic errors are generally the largest contributor to measurement uncertainty.
Vector-error correction is a process for characterising systematic error terms by measuring known electrical standards such as mechanical or electronic open, short, load, and through.
Now, electronic calibration reduces calibration time, makes calibrations easy to perform, minimises wear on cables and standards and helps eliminate operator errors. Also highly repeatable temperature-compensated terminations provide excellent accuracy.
Other errors to be aware of include:
- Random errors which can vary with time in random fashion (unpredictable). Main contributors to random errors include instrument noise, switching and connector repeatability.
- Drift errors due to system performance changing after a calibration has been done and primarily caused by temperature variation.
- Crosstalk or signal leakage between test ports during transmission and can be a problem with high-isolation devices and high-dynamic range devices
In summary, network analysis is imperative to gaining a true understanding of a devices performance and characteristics but can become a complicated measurement. If in doubt get in touch and we’ll be only too glad to help.