nRF Distance measurement using nrf_dm_high_precision_calc()

[Miguel Roque]

Hello all,

I'm using the sample for distance measurement nrf_dm. I'm using the ncs sdk 2.2.0, with toolchain 0.15.1. Currently the sample runs without any change, but now I'm trying to increase the accuracy of the system by changing the dm calculation function from nrf_dm_calc to nrf_dm_high_precision_calc. However, after flashing the devkit that I'm using (nrf52833dk) the system generates a fatal error:

E: Faulting instruction address (r15/pc): 0x0003cee6
E: >>> ZEPHYR FATAL ERROR 2: Stack overflow on CPU 0

Does anyone has the same problem? What's the reason?

Thank you for your time!

[Øivind]

Hi,

High precision calc is not currently supported in the DM sample, as explained by Simon here:  RE: Distance Measurement library questions (part 2)

There is a pull request on the github for this feature, so you could try to add support using this: https://github.com/nrfconnect/sdk-nrf/pull/7895

[Miguel Roque]

Hi Øivind , thank you for your prompt reply. Yes, I have been able to make it work meanwhile. 

Now I have another question, if I understood correctly, the dm module makes has a initial synchronization phase, and after that an initiator requests a distance measurment to the other device who acts as reflector and answers to it with the necessary data. Is it possible to force a device (for instance "DEVICE A") to always be the initiator and the "DEVICE B" to always be the reflector?

[Øivind]

That should be possible, but I don't think it is officially supported. You will need to do this yourself by splitting the sample into two versions. It may be as simple as one versions where you comment out adv_start(), and another where you comment out scan_start() in bt_sync_init() in main.c. I believe the roles are assigned based on the exchange of SCAN_REQ and SCAN_RESP.

[Miguel Roque]

I tried your suggestion but it doesn't seems to work. Without the scan (just in Device A) there is no measurements at all (both side)

[Øivind]

Ok, removing the adv_work could work perhaps? I will ask the developer for a suggestion, but that is something to try in the meantime.