Multi-Connections at same time possible on newest version of Channel Sounding for ranging Initiator/Reflector?

[NickDeLeenheer]

Hi,

I was wondering if the newest sample available in NCS3.0.1 is already capable of holding multi connections. Otherwise, for this little time that is present i will do my own implementation to connect with the boards sequentially...

Also, would it be hard to implement multiple antennas to my nRF54L15 boards? I think that multiple antenna paths would increase accuracy, since the accuracy of the distance estimation is very dependent on the specific shortest path that it found. nRF54L15 only has 1 Antenna path (i Think). And if it has more, is it easy to configure multiple antenna paths?

A last question on how to improve accuracy: Is the ifft better for inside? It uses the same phase samples as the Phase Slope method, but i was wondering how it works and where i can find the implementation, because it seems better for now. I also was wondering why it is now taking longer to get my distances with the same configuration compared to the previous version of this sample?

Kind regards,

Nick De Leenheer

[Elfving]

Hi Nick,

It's possible to run CS on multiple connections at the same time. We don't make much effort to avoid scheduling conflicts however so parameters and kconfigs have to be selected carefully. One thing to note that if you have issues with this, one usual suspect is that you haven't increased BT_CTLR_SDC_CS_COUNT.

Is your end-goal to trilaterate a position? 

Also, would it be hard to implement multiple antennas to my nRF54L15 boards? I think that multiple antenna paths would increase accuracy, since the accuracy of the distance estimation is very dependent on the specific shortest path that it found. nRF54L15 only has 1 Antenna path (i Think). And if it has more, is it easy to configure multiple antenna paths?

Could you expand a bit on what you mean by this, and "antenna paths"? EDIT: I see now that we support multiple antennas for CS much like DF. This was news to me. That is what is being referred to by antenna paths. The NRF54L15DK just has one paths, but you could connect more.

The nRF54L15DK has one antenna on it. You can however connect a different antenna on it as well. I am not so sure if something along the lines of a antenna matrix (like what you use for direction finding) would be able to improve anything here.

A last question on how to improve accuracy: Is the ifft better for inside? It uses the same phase samples as the Phase Slope method, but i was wondering how it works and where i can find the implementation, because it seems better for now. I also was wondering why it is now taking longer to get my distances with the same configuration compared to the previous version of this sample?

Not necessarily. This older webinar goes through the differences.

Regards,

Elfving

[NickDeLeenheer]

Hi,

Thank you for your input on this matter, i will look into parallel connections later.

But for now, i already succeeded in multiple connections being hold SEQUENTIALLY for my older version, i a almost succeeded in this matter using the new sample from Nordic.

I would like to get the problem that i stated in my other thread: "Ranging CS data from multiple connections", to be fixed by this Thursday, since i will be doing measurements with my thesis supervisor.

It's a problem with the fact that we don't get succesfull data steps anymore, so the distance estimates are not getting updated (data_complete) ...

If you could take a look at that thread, or if another engineer can help me with this quickly, i would be very happy.

Kind regards,

Nick De Leenheer

[Elfving]

Great

NickDeLeenheer said:

I would like to get the problem that i stated in my other thread: "Ranging CS data from multiple connections", to be fixed by this Thursday, since i will be doing measurements with my thesis supervisor.

It's a problem with the fact that we don't get succesfull data steps anymore, so the distance estimates are not getting updated (data_complete) ...

If you could take a look at that thread, or if another engineer can help me with this quickly, i would be very happy.

A coworker of mine is assigned to it, and will look into it in time. Though it is hard to say if we'll be able to fix something for you there by Thursday.

Regards,

Elfving

[NickDeLeenheer]

Hi,

I’m currently working on a project that requires managing multiple connections in parallel, although handled sequentially, and I’m running into some difficulties. Could anyone share working code samples or libraries that implement such multi-connection handling? Even a simple proof-of-concept would be extremely helpful.

Ideally, I’d like to see:

• Example code illustrating parallel (but sequentially processed) connections

• Explanations of how the connections are managed (e.g. threading, async, queues)

• Any tips or best practices you’ve learned from your own implementations

If it’s easier, I’d also appreciate the opportunity to schedule a short call to walk through my current approach and get advice on the next steps.

Thank you in advance for your help! I know you’re all busy, so any pointers—even if brief—are greatly appreciated.

Kind regards,

Nick De Leenheer

[Elfving]

Unfortunately the public samples I've linked you to is what we currently have available. I also see that you have another case open on this, we prefer to keep the tickets separate to avoid doing double the work. 

Regards,

Elfving

[NickDeLeenheer]

Hi,

Yes i am sorry, i was a little bit stressed out about getting my code fixed as soon as possible. I appreciate all the help that i get from Nordic engineers. In my first reply to your answer i stated that i would like to have help on my other case (ticket). I wanted to close this ticket already, but i first wanted to also speak to you about my urge to get a working code...

I apologize again for this matter.

In the other thread i wasn't really getting where i wanted to go, which made me stress out a bit.

My objective stays the same: trying to get my code (which already works to get my connections perfect sequentially ONE BY ONE), but now for the new sample using the CS_DE.c for distance estimation calculations for IFFT algorithm. I really think this should be possible, and i came very close to achieving it, but i still need some insights in this. 

I tried out making a struct for every device to make every device have it's own buffer memmory voor latest_local/peer_steps and it's own distance_estimates_buffer. This wasn't really successful. I tried to keep everything the same as in the published sample, and work with 1 buffer, and change the counters, but there is 1 issue that made it NOT work after a while... the Last ranging round was always not fully done (procedure never completed and 1 procedure involves 2 ranging steps (mode 2 and mode 1 for Phases and RTT).

I already know a lot on how to modify the code and what everything does, but i still need someone who can guide me to what is possible or who has experience in this specific sample and/or multiconnections. So, if there is someone, should i close the existing tickets and try a new one? Or should i try a different method?

Thank you very much for your time and patience—I really appreciate the support.

kind regards,

Nick De Leenheer