Like fine-tuning a race car: the role of calibration and maintenance in optimising VNA performance : Page 5 of 6

June 16, 2014 // By Adam Purkiss, Anritsu (UK)
The modern Vector Network Analyser (VNA) is an extraordinary feat of engineering, enabling the precise measurement and analysis of signals across a broad frequency spectrum, in some cases up to 110 GHz.
Importance of co-efficient data

Assuming the use of a verified VNA, the engineer’s first step in implementing an accurate test set-up will be to define an accurate reference plane from which to measure. This is only possible if the co-efficient data, supplied with most calibration kits, are entered into the VNA prior to performing a calibration. Separate co-efficient data sets are supplied for coaxial, waveguide and microstrip calibrations, and their use provides for the most accurate measurement results.

During the calibration process, the analyser mathematically removes the offset data to give a zero or reference point at the connector measurement reference plane. (The measurement reference plane of a connector is regarded as the contact point of the outer conductors.)

If no calibration standards are available, average default definitions are stored in the VNA. These might suffice for some measurements, but inevitably they compromise measurement accuracy, and the use of specific co-efficient data is preferable when high accuracy is called for.

Good practice in test preparation

As described above, it is common to assume that there is a fault in the VNA when test measurements show a marked variance from the expected values. Before assuming such a fault and calling out a service engineer, it is prudent for the user to check that the fault does not lie elsewhere.

  1. Perform the vector error correction calibration a second time. This verifies that the VNA and its calibration cables are working as they should;
  2. Dirty calibration standards affect calibration results, especially at high frequencies. (see Figure 3). They must be cleaned with swabs soaked with de-natured alcohol;
  3. Check that the correct calibration standards were used, and in the correct order.

Figure 3: a dirty calibration standard can invalidate test results.

Design category: 

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.