The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

AsterG — Sat, 27 May 2023 12:57:28 GMT

I understand. Thank you for information, Kenneth.

With best respect.

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

Kenneth — Fri, 26 May 2023 16:24:05 GMT

Setting up timer compare events to trigger the radio tasks over a ppi channel will have the most accurate timing. All while loops are by design inaccurate, because the event may occur at any time within the loop, which can cause a difference of several 64MHz instruction between exiting the loop for each while loop.

I can't guarantee this will solve your issue, but at least it will make it more accurate.

Kenneth

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

AsterG — Fri, 26 May 2023 14:01:56 GMT

Very thank you for information and your patience. Now I learn the specification more detailed and have last question by this theme before start my long way to find solution of my problem.

If all RADIO events to make use over PPI, for avoid any delays between calling those events, so this will make synchronization much accurate as it possible, So why is needed TIMER?
Delay between events calls on Initiator and on Reflector already will constant from iteration to iteration, because all is used over PPI, so why TIMER is needed in this chain in context my task?

With best respect. Thanks.

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

Kenneth — Thu, 25 May 2023 11:18:23 GMT

Hello,

Sorry, but I don't really have more information, I highly recommend that you update and use the library in nRF Connect SDK.

I think the main problem is that in specific the start tasks here occur asynchrounous (e.g. depending on the inaccurate while() checks you have), I likely suggest that you for instance try to connect start tasks to a timer of for instance fixed 1us resolution, since the on-air data rate is 1Mbps. At least that is where I would have started.

Kenneth

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

AsterG — Thu, 25 May 2023 09:27:15 GMT

Hello. Could you please tell me exactly how TIMER and PPI can be used regarding my situation. I'm not asking you to write code for me, but, at least, direct me to the right thoughts to continue searching solution for my question.

I thought that if the RADIO is not reconnected (not use RADIO->TASKS_DISABLE) when switching from transmitter (TX) to receiver (RX) and vice-versa, then LOs generation remains continuous (for both Initiator and Reflector), then the delta of LOs phases between these Initiator and Reflector is constant and stable over the time.
I tried to use also interrupts for RADIO peripheral (over NVIC_EnableIRQ(RADIO_IRQn)) with same negative result. Please explain me how I can right to use for my experiments TIMER and PPI...?
Is really interrupts (over NVIC_EnableIRQ()) or while{...} (as in example files) solutions introduce such large error into the stable LO generation? Or I 'm wrong in something?

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

Kenneth — Thu, 25 May 2023 07:43:24 GMT

Hello,

Unfortunately we do really only support the distance measurement library you find in nRF Connect SDK.

Our suggestion looking at your code though would be to look into avoid busy waits (such as while(some_event)), and instead try to use TIMER and PPI to get more control over the timing.

Kenneth

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

AsterG — Wed, 24 May 2023 11:51:32 GMT

Hello Kenneth. Thanks for the quick response.

Firstly, I use the project infrastructure based on NRF5 SDK and there is no way to switch to NRF Connect SDK at the current moment.

Secondly, I consider the goal not to create a rangefinder, as such, but something similar to a motion sensor based on its own phase-based algorithm with its own characteristics.

That's why I am interested in the technical possibility of such a phase shift measurement.
Yes, I am familiar with these documents, from which it follows that such a phase shift measurement is possible on these chips.
Having studied the specification, I wrote test code for Initiator and Reflector in which I tried to take into account all the nuances, but, obviously, somewhere I made a mistake in setting up the RADIO peripheral or some else.
Could you please to explain, by my code, what I do wrong?

With best respect. Thanks.

RE: The problem to find the phase shift between the waves of two fixed NRF52833DK boards using CTE data.

Kenneth — Wed, 24 May 2023 10:54:19 GMT

Hello,

Is there any reason why you can't use the ready-made library that calculate the distance for you already?
https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/libraries/others/dm.html#distance-measurement
https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrfxlib/nrf_dm/README.html#nrf-dm

Example:
https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/samples/bluetooth/nrf_dm/README.html

Kenneth