How much do you know about the real results if you don't calibrate before the test?

author | Liu Lijuan （ Mr. Yibo technology high speed team member ）

Last week we were B The station has put on why the vector network should be calibrated , And the video of calibration practice steps .

Some classmates are asking me ： Then how can I not calibrate ？ See you operate for a long time , It's a waste of time , No school, no school ……

For such a proud and charming person , He can't be convinced simply by theory , First throw out the test results directly and simply .

Sagittal net (VNA) Housekeeping skills are measured S Parameters , Let's first look at the same DUT, After calibration, measure S What is the result of the parameter ？ And if you don't calibrate , Measured S What are the parameters , How different can it be from the real result ？

For the differential routing of a pair of inner layers shown in the above figure , stay 5GHz It's about , The measured insertion loss after calibration is only 1.264dB, The insertion loss measured without calibration is 4.226dB, Much more than the real loss 3dB. This is the extra 3dB It's the instrument +cable Loss of .

For measuring S Calibrate before parameters , The vast majority of students still recognize . Because the instrument cannot be used in the frequency domain 、cable and DUT The effect of , What you see is the three total The effect of . But whether to calibrate the impedance measurement , There will be different voices :“ The impedance results are displayed in chronological order , You can see whether it is calibrated or not DUT The impedance of ”. Don't explain , Or show the comparison results directly .

Let's first look at the direct measurement of impedance without calibration , What is the result . I set a measurement interval , It can be seen that the maximum value in the measurement range is 107.9Ω, The minimum is 101.9Ω.
 
Let's take a look at the impedance test results after calibration ： At this time, the maximum value in the measurement range becomes 102.2Ω, The minimum becomes 97Ω.

The two pictures above , Although it is not calibrated, it can be clearly identified as belonging to DUT Impedance interval of , But the impedance measured directly without calibration is higher than that measured again after calibration 5ohm about . Can you accept this difference ？ For example, the design impedance of the customer is 100ohm, The impedance control requirements for processing in the factory are 100ohm+/-10%, The impedance of the board is 106ohm, Originally, the board was properly satisfied 100ohm+/-10% The requirements of , The result was tested by your holy hand , The impedance soared directly 111ohm, Okay , sentence Fail. Whether you can accept it or not , Anyway, the factory was the first to be wronged , Then the customer was frightened , Prepare for a meeting to discuss whether to proceed PCBA 了 .