Is the nRF52840-CKAA port impedance for the various transmit power settings (or at least for 8dBm) available?
Alternatively, is the expected port impedance, when viewed from the antenna side into the end of the recommended matching network, roughly 50 ohms (see "Circuit configuration no. 1 for CKAA WLCSP" in nRF52840_PS_v1.7)?
Hi, Johnny.
The port impedance isn't included in the PS, as we rather recommend using the matching network found in the reference circuitry and then performing a tuning of the radio's matching network to secure an optimal output power for the carrier frequency and at the same time keeping the power of the harmonic frequencies below the threshold required for certification.
Regarding the antenna, we recommend tuning that towards an impedance of 50 ohms.
Best regards,
Mathias
Hi Mathias,
I'm still investigating whether I didn't do anything wrong in my procedure (bad soldering, etc.), but as far as I know, I tuned my antenna to 50 ohms (by trimming my antenna and adding additional matching components after the official matching circuit, thus the 3.9nH, 1pF and 4.7nH as recommended in https://docs.nordicsemi.com/bundle/ps_nrf52840/page/ref_circuitry.html#ariaid-title10 ). I did, however, notice that the spacing between those matching components in my design is larger than the reference layout provided by Nordic, so I might need to spend some additional effort to characterize the deviations in my design.
At least you confirmed that the targeted antenna impedance after the matching circuit is 50 ohms.
Thanks for replying.
Have you tried tuning the radio's matching network? The reference circuitry is a good starting point for the matching network but usually, some additional tweaking is necessary to achieve optimum performance, as there will be some differences from design to design.
johnny89 said:I did, however, notice that the spacing between those matching components in my design is larger than the reference layout provided by Nordic, so I might need to spend some additional effort to characterize the deviations in my design.
Having larger distances between the components is also something that may introduce additional losses.
How is the performance of your device compared to what you were expecting?
You can upload your schematic and layout here for me to look at the design, but then I should make the ticket private first.
Can I accurately characterize the impedance of the "output port" (with this, I mean the port looking back into the matching network from the antenna side, with the matching network still connected to the nRF RF port) by putting the nRF into RX-only mode and getting the S11?
I intend to use the above S11 to tweak the matching network until I see a near 50-ohm impedance from the output port. This will hopefully provide an optimal match for my antenna, which I tweaked (in isolation) to provide a 50-ohm load.
I ask because I'm not sure if the output port impedance would look different depending on the nRF's configured transmit power (+4dBm, +8dBm, etc.). Thus, would I rather need to somehow characterize the output port with the appropriate transmit power configured so that I can optimize my matching network?
If the output port impedance does change with transmit power, I assume the only "easy" option I have is to set the nRF to transmit a CW tone with my intended transmit power (8dBm) and measure it with something that has a known 50-ohm load (like a spectrum analyzer and some attenuators) to confirm that I get 8dBm out. If not, I experimentally tweak the matching circuit until I do.
Can I accurately characterize the impedance of the "output port" (with this, I mean the port looking back into the matching network from the antenna side, with the matching network still connected to the nRF RF port) by putting the nRF into RX-only mode and getting the S11?
I intend to use the above S11 to tweak the matching network until I see a near 50-ohm impedance from the output port. This will hopefully provide an optimal match for my antenna, which I tweaked (in isolation) to provide a 50-ohm load.
I ask because I'm not sure if the output port impedance would look different depending on the nRF's configured transmit power (+4dBm, +8dBm, etc.). Thus, would I rather need to somehow characterize the output port with the appropriate transmit power configured so that I can optimize my matching network?
If the output port impedance does change with transmit power, I assume the only "easy" option I have is to set the nRF to transmit a CW tone with my intended transmit power (8dBm) and measure it with something that has a known 50-ohm load (like a spectrum analyzer and some attenuators) to confirm that I get 8dBm out. If not, I experimentally tweak the matching circuit until I do.
johnny89 said:If the output port impedance does change with transmit power, I assume the only "easy" option I have is to set the nRF to transmit a CW tone with my intended transmit power (8dBm) and measure it with something that has a known 50-ohm load (like a spectrum analyzer and some attenuators) to confirm that I get 8dBm out. If not, I experimentally tweak the matching circuit until I do.
Yes, that is the way to do it. Look at how the output power of the fundamental frequency and the harmonic frequency change depending on the matching network.
mathiaso said:You can upload your schematic and layout here for me to look at the design, but then I should make the ticket private first.
This is still an option if you want to. We also offer a tuning service, free of charge.
Yes, that is the way to do it. Look at how the output power of the fundamental frequency and the harmonic frequency change depending on the matching network.
I tweaked the network until I got > 7dBm fundamental power out and the harmonics where I'm comfortable with them. Tweaking the matching circuit to get an optimal output impedance of 50 ohms (via an S11) would not provide any good insight into harmonic performance, so looking at the output power of the fundamental and harmonics seems to be the best approach, thank you for confirming it.
What helped guide me was reverse-engineering a theoretical RF port impedance value from the proposed matching network, assuming it tweaks it exactly to 50-ohm real impedance - this landed at approximately 84+j16.5. I used this value in my simulations, and it reflected my practical observations quite well.
