0
\$\begingroup\$

I have a PCB with a pi-matching network, and I want to calculate the component values for the matching network.

I use LiteVNA and a WE-MCA Multilayer Chip Antenna for this project.

Here is the basic setup of the testing PCB. I use a 10 cm long RG316 coaxial cable:

enter image description here

First, I calibrate the VNA with OPEN, SHORT and LOAD shown in the picture. Measured impedance for OPEN, SHORT and LOAD is now a dot on the Smith chart.

enter image description here

After measuring the antenna impedance with a series 0 Ω resistor, the result is 5 - j62.

enter image description here

Then I add a series inductor. After several attempts, instead of 5 nH under calculation, I found that 7.3 nH should be fine; it will move the point on the Smith chart to somewhere near DP2.

schematic

simulate this circuit – Schematic created using CircuitLab

The measured impedance with a series 7.3 nH inductor:

enter image description here

Here comes the problem: when I add the SRC shunt capacitor, the impedance becomes unpredictable. For example, when I add 1 pF to the SRC shunt, the expected result is:

schematic

simulate this circuit

enter image description here

However, the measured result is:

enter image description here

When I replace 1 pF with 3.9 pF, the impedance is :

enter image description here

The result follows a completely different arc, and I don't know how to match it to 50 Ω. Does anyone know what is going wrong?

I understand components are not ideal, but I think the result is completely out of prediction. I doubt that it really can be explained by the non-ideality of the components.

I tested serval online calculators; basically all show the same result. They should follow the same equations to calculate the impedance. Is there any other model/equation I should consider?

\$\endgroup\$
4
  • \$\begingroup\$ What series and manufacturer are you using for the 7.3nH and 1pF components ? Have you considered if your on-line simulator uses acceptable frequency models for the simulation or are the passive components assumed to be ideal ? \$\endgroup\$
    – citizen
    Commented Nov 22, 2023 at 13:08
  • \$\begingroup\$ I also considered would it be the components issue, so I bought some rf capacitor and inductor: cap datasheet ind datasheet I bought a 2pF rf capacitor to replace the 1pF capacitor, the result is 1.5 - j42, no significant change, also fall to the bottom of the smith chart, so I think its not the components issue. \$\endgroup\$
    – Kelvin
    Commented Nov 23, 2023 at 6:10
  • \$\begingroup\$ but you still haven't given us the Manufacturer and the specific part numbers and or links to the datasheets etc ... \$\endgroup\$
    – citizen
    Commented Nov 23, 2023 at 8:16
  • \$\begingroup\$ Sorry I can't put more links into the post so I puts the datasheet links in the above comment (cap datasheet) (ind datasheet) mouser.hk/datasheet/2/40/600L-2899701.pdf mouser.hk/datasheet/2/281/1/JELF243A_0100-1380931.pdf \$\endgroup\$
    – Kelvin
    Commented Nov 23, 2023 at 9:03

1 Answer 1

0
\$\begingroup\$

I had the same problem, I used Murata's RF capacitors and inductors, and I used si9000 to match the impedance of the transmission line, but the S11 parameters often changed irregularly in the actual test. I have tried nano VNA and lite VNA and this is the case. I think this may be an issue with the accuracy of handheld VNA because of the huge price gap between it and professional devices.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.