VNA verification vs user calibration: why you need both
You calibrate your network analyzer with a cal kit every day. So why does it still need to go to a calibration lab? The two processes answer different questions.
Vector network analyzers (VNAs) cause more confusion about calibration than any other instrument we handle. Engineers perform a “calibration” every time they set up a measurement, using a cal kit or an electronic calibration module. Understandably, some wonder why the analyzer also needs to come to a calibration lab.
The answer is that the word calibration is being used for two quite different things.
User calibration: error correction
When you connect short, open, load and through standards (SOLT), or use an electronic calibration module, the VNA measures those known standards and calculates error terms: directivity, source match, reflection tracking, load match, transmission tracking and isolation. These describe the imperfections of the analyzer, its cables and adaptors at the moment of measurement.
The analyzer then mathematically removes these errors from subsequent measurements. This process is often called vector error correction, and it is extremely powerful: it can turn an analyzer with mediocre raw performance into a highly accurate measurement system.
But it relies entirely on one assumption: that the calibration standards are what the analyzer thinks they are. The definitions of each standard are stored in the analyzer as a cal kit file. If a load is damaged, a short is worn or the kit file is wrong, the error correction is wrong, and nothing in the user calibration will tell you.
Lab calibration: performance verification
A calibration at our lab answers different questions:
- Is the analyzer’s hardware performing to specification? We check source frequency accuracy, source power accuracy and linearity, receiver dynamic range, noise floor, trace noise, and the raw (uncorrected) system performance.
- Does the corrected system measure correctly? We perform a user calibration and then measure verification standards, such as precision airlines, mismatched airlines and fixed attenuators, that have been independently characterised and have known, traceable S-parameters. Comparing the VNA’s measurement with the known values shows whether the whole system, analyzer plus correction, gives the right answer.
Don’t forget the cal kit
A mechanical cal kit is itself a set of standards. Its loads, shorts, opens and adaptors wear with every connection, and its definitions need to remain valid. Cal kits and electronic calibration modules should be calibrated or characterised regularly, typically annually, and their connectors inspected with a gauge before every use.
Good practice in daily use
- Gauge your connectors. Pin depth out of specification can damage mating connectors and invalidate measurements.
- Use a torque wrench of the correct value for the connector type.
- Keep the cal kit clean and stored in its case.
- Measure a check standard after each user calibration. A known attenuator or airline, measured and logged daily, will quickly reveal a problem with your kit or cables.
A VNA with a valid lab calibration, a characterised cal kit and good connector discipline is one of the most accurate instruments in any RF lab. Take away any one of those three, and it can produce beautifully precise, completely wrong results.
