nRF52805 pin matching

RE: nRF52805 pin matching

Rashid Talib — Thu, 08 Apr 2021 07:27:07 GMT

Ok, thanks for your help & best regards

Rashid

RE: nRF52805 pin matching

JONATHAN LL — Wed, 07 Apr 2021 14:43:05 GMT

Usually good TX means results in RX performance also, and radiated testing will result in differences but you should be in the right ballpark now that the conducted measurements are done.

I don't have a logical explanation for the graphs other then that you can see a difference but that's why we measure harmonics and power output, it is the best indicator and easier to see improvements.

Regards,
Jonathan

RE: nRF52805 pin matching

Rashid Talib — Wed, 07 Apr 2021 06:30:25 GMT

Hi Jonathan
Thanks for your reply, yes I did conducted measurements.

I reduced L2 from 3.3nH to 3.0nH and C13 from 1.2pF to 1.0pF and the conducted power increased from -1.56dBm to -0.04dBm!! The total radiated power (TIG measured over 3 planes) increased by 1.2dB (2.402GHz), 2.6dB (2.440GHz) and 2.0dB (2.480GHz). The harmonics are well below -50dBc which is good.

So I'm quiet happy with the result, next I will do receiver sensitivity measurements to make sure Rx mode is not negatively impacted.

Is there anything else I must keep an eye on before going into approval ?

One interesting fact:

Simply looking at the S11 coefficients it doesn't seem obvious that the matching is improved at all but the power measurement clearly shows an improvement (as you said in your reply). Find the comparison of the S11 measurements below. Do you have a logical explanation for this?

Best regards

Rashid

RE: nRF52805 pin matching

JONATHAN LL — Tue, 06 Apr 2021 14:55:36 GMT

Hi,

If you are measuring conducted it will probably differ from radiated testing, so it is only a somewhat accurate whey to tune. When it comes to choosing the component values it is not that simple, it will differ from design to design and HW layout will effect the result.

There is a limit to of effective simulation and calculation is when it comes to the final product. We find it best and most efficient to start with the recommended values from the reference design and then we do one component at the time, increase of decrees in value, or different component type, like form a different manufacturer. Sometimes it even helps to rotate the components(inductors) 180 degrees to filter the harmonics. So we change one parameter at the time while we measure the harmonics and output power, there will (hopefully)be some combination that provides god output power and reduce the harmonics as well.
So we tune the radio by trail and error testing.

feel free to ask for more if anything is unclear.

Regards,
Jonathan