nrf52 antenna impedance

RE: nrf52 antenna impedance

vea45 — Wed, 18 May 2016 07:34:53 GMT

Well, I don't have the calibration kit, I didn't thought about doing it this way... It's interesting, I will give it a try!

RE: nrf52 antenna impedance

mstrouse — Wed, 18 May 2016 03:05:28 GMT

Out of shear curiosity ... why can't you cal the vna? If you have 2 identical boards, choose 1 board, attach an sma connector at the input to the antenna structure, then modify the board to do an SOL(T) at the input to the antenna [obviously "disconnect" the antenna and use a small 50 resistor (0402/0201) for the load ... precision is going to count here]. Modify the 2nd board with another sma connector in the same location as the first board. Make your measurement(s). Do your calculations. Then go back and apply any additional matching network needed.

RE: nrf52 antenna impedance

RK — Tue, 17 May 2016 23:22:04 GMT

No you can't change C3 and L1, they are an RF choke to remove unwanted harmonics, not really part of the matching network. The pi network or single matching component (single if you are using a trace antenna where you can customize the length) comes at the other end of the feed line.

RE: nrf52 antenna impedance

vea45 — Tue, 17 May 2016 20:56:36 GMT

Yes I have access to an VNA but I can't calibrate it... I read those papers but I want to know if it's possible to change the values fof C3 and L1 so I could tune the pi network... (I added a shunt component to complete the network). Also I used CST to do a simulation of my situation but the measured results and the simulated ones are quite diferent...

RE: nrf52 antenna impedance

mstrouse — Tue, 17 May 2016 17:49:42 GMT

Do you have access to a VNA (vector network analyzer)? Using the VNA, measure the antenna in-situ and tweak the matching components to suit. And except for using a high-end microwave software simulation tool (that can also accurately simulate the silicone), the VNA route is your best bet. [EDIT] See this for information on how to do this - infocenter.nordicsemi.com/.../nwp_017.pdf

Another (sloppy) alternative is ... you could use a spectrum analyzer and measure the frequency / output power (and harmonics!) as you adjust the component values. Ugly, but it could get you close. [EDIT] See this to get an idea on how to do this (example using nRF24L01) - infocenter.nordicsemi.com/.../nwp_013.pdf