Is the frequency difference between transmitter and receiver compensated?

[markush]

Hi!

We plan to build an AoA localization system with the nRF52833. To calculate the signal phases per channel, the antennas are connected to the BLE-RX via switches sequentially. Therefore, the CW-carrier part of the BLE signal is sampled sequentially with some microseconds time difference between the channels.

Is the carrier frequency difference between the BLE-transmitter and the BLE-receiver estimated and compensated in the nRF52833? If the frequency would drift away too much, the phase calculation of the IQ-data per channel would be too errorneous.

Thank you very much for answering the question!

[Pete W]

No, you will need to allow for frequency errors on your Tx and Rx. In the ideal case the CTE would appear in your Rx as a 250 kHz tone (at least for a 1M PHY), since it is generated by a run of 1's on the carrier, but if your Tx and Rx have 10 ppm crystals, for example, this could be as low as ~200 kHz or as high as 300 kHz (the limits with 10 ppm at 2.440 GHz). This is why you need to use the leading part of the CTE to estimate the rate of change of phase (i.e the frequency) before you can use it to measure the phase difference between the slots associated with your antennas. So you measure it and then use it to predict where the phase should be in the following slots.

Hope that helps.

Pete

[Kazi Afroza Sultana]

Hello,

Frequency does not matter, the phase difference between the different antennas is taken into account. The phase difference is measured. Based on the one reference antenna, you have to measure the phase relative to that.

You can read this blog post https://blog.nordicsemi.com/getconnected/finding-your-way-with-bluetooth and the white paper https://infocenter.nordicsemi.com/index.jsp?topic=%2Fnwp_036%2FWP%2Fnwp_036%2Fintro.html to get details concept. There are two NCS samples Bluetooth: Direction finding connectionless locator — nRF Connect SDK 2.2.99 documentation (nordicsemi.com), Bluetooth: Direction finding connectionless beacon — nRF Connect SDK 2.2.99 documentation (nordicsemi.com) you can look at. 

Thanks.

Best Regards,

Kazi Afroza Sultana

[Pete W]

If frequency offset is ignored, i.e. you assume the I & Q are exactly 250 kHz (or 500 kHz for 2M PHY), then the measured phase difference will be greater or less than the actual value.