Don't be fooled by good calibration results--Mistakes in vector network analyzer calibration

Due to the diversity of the devices under test, there are many types of vector network analyzer calibration, and operators are prone to misunderstandings. Sometimes the calibration results look "beautiful", but they are actually wrong values. The following are common misunderstandings.

· Use calibration kit to verify calibration results? ! ! !

The most common mistake that operators make is to use the calibration parts of the calibration network instrument to verify the correctness of the calibration results!!! Usually after calibration, operators will connect the open/short circuit breaker that has just been calibrated to the instrument to observe whether the logarithmic amplitude curve of S11 is near 0dB; or directly connect the two ports to observe whether the logarithmic amplitude curve of S21/S12 is near 0dB.

As soon as the first sentence above is said, many people will realize where the problem lies. I believe that no experiment uses the same thing to verify its correctness. The manufacturer of the calibration parts usually provides a special calibration part so that the user can verify the correctness of the calibration results. If there is no calibration part, it is recommended to use another set of calibration parts or keep an adapter/adapter/attenuator with known characteristics for verification.

Do not tighten the calibration piece by hand

The torque wrench is standard equipment for calibration parts, and the manual usually introduces how to use the torque wrench. The torque wrench is also a "calibrated" product, which can ensure that the calibration part and the instrument/cable contact surface reach "just right" contact. Too tight or too loose contact will affect the contact resistance of the calibration part contact surface, thus making the calibration result worse.

· Want to use “non-intrusive calibration”?

In many cases, the DUT we measure is "non-insertion". That is, the two ends of the DUT cannot be directly connected. Non-insertion is divided into two categories:

1. The connector types are inconsistent (one end is 3.5mm and the other end is 2.4mm);

2. The connectors have the same polarity (both ends are 3.5mm female).

For non-insertion devices, the calibration process requires adapters (electronic calibration components can be configured with connector types, so adapters are not required). For the introduction of adapters, operators usually measure adapters with delay and loss as straight-through without delay and loss (i.e., 0-length straight-through). This method has large errors. At 3 GHz, the transmission tracking amplitude has a jitter of about a few tenths of a dB, while at 40 GHz it can reach 1 dB. This method also causes uncertainty in reflections.

For non-insertion devices, vector network analyzers provide a variety of high-precision calibration methods: unknown through calibration, electronic calibration, etc.

Please use high quality cables

Statistical results show that high-quality cables can achieve twice the result with half the effort in testing, while poor-quality cables can cause unpredictable deviations in calibration results, thus spending a lot of time finding the problem.

A good horse deserves a good saddle! After spending a lot of money on instruments and calibration parts, why not be willing to equip them with high-quality cables?

Some suggestions

Calibration accuracy ranking

From a theoretical point of view, TRL can provide the best calibration quality. In fact, errors introduced by cable bending and mismatch caused by repeated connections of mechanical calibration will make the quality of TRL calibration lower than that of electronic calibration. Considering the errors of manual operation and the consistency of mechanical calibration parts (mechanical calibration parts use the same data), there is no doubt that electronic calibration parts can always provide better quality than mechanical calibration parts in actual use.

How long can a calibration last?

Because the vector network analyzer is calibrated in a specific environment, as time goes by, environmental factors become inconsistent with the conditions at the time of calibration, resulting in invalid calibration results. Environmental factors include temperature, humidity, cable bending, etc., which are difficult to quantify. Therefore, after calibrating once, calibrate again until you feel that the calibration effect does not meet the test requirements.