DW3000 Direction Finding

Wireless Connectivity Ultra-Wideband
1

I’m selecting a chip for a project that requires support for direction finding. We’re currently looking at the DW3110 and DW3120.

I’m a bit confused about the requirements for this. The DW3120 states that it supports PDoA and TDoA since it has two antenna ports, which makes sense. However, the DW3110 also states that it supports PDoA and TDoA; how is this possible when it only supports a single antenna? Aren’t two antennas needed for that?

I’m also curious if any of the products support 3-dimensional Angle of Arrival (elevation and azimuth). Are there any configurations that would support this (e.g. a DW3120 with one antenna on one port, and a switch between two other antennas on the second port), and if so, are there any firmware examples?

2

Hi Jordan,

Sorry for the confusion but you’re right. DW3110 does not support PDoA and it needs to be corrected.

But, If you need a direction finding application, you can use DW3110 as a tag and DW3120 as an anchor to find out the direction by using PDoA.

You can implement a 3D AoA application by using external switches as you mentioned above. But unfortunately, DW3000 series doesn’t support to do azimuth and elevation PDoA simultanously. That would require 2 different Receiver but DW3000 have only one.

Kind regards,
Emre

3

Thanks @Emre_Ozbas_Qorvo . To clarify, for 3D direction finding, I could do it with the DW3120, but I’d have to do azimuth and elevation subsequently, correct? i.e. do azimuth, switch one of the antennas, do elevation?

Do you know what the minimum distance must be between two antennas to be able to do PDoA?

4

Hi @Jordan ,

Yes, that is right. You need to switch between antennas to do azimuth and elevation.

In theory, using half lambda antenna separation between antennas would provide -90 to 90 degrees coverage but at extreme angles, there will be a phase wrapping. To mitigate that using 0.45 lambda is recommended. So, you can design an L shape or triangular shape antenna array with 0.45 lambda separation.

Kind regards,
Emre

5

Thanks for the details! I’m getting a better idea of this, but one more clarification.

Say we have an anchor in the center of a room, and want to be able to find 3D direction to a tag that is anywhere around it.

Let’s say that our first antenna is at the origin, (0,0,0). We have a second antenna (“+X”) at (1, 0, 0). With these two antennas, if we get a reading of 45 degrees, that direction could actually be 45 degrees, OR 135 degrees, because both (if the tag is “in front of” or “behind” the antennas) would give the same phase difference of arrival. Is this correct?

So to determine a fully-defined azimuth angle, we would need a third antenna, “+Y” at (0, 1, 0). This would also give us two possible elevations, e.g. an elevation reading of 30 degrees could actually be 30 OR 150 degrees.

So to have both fully-defined azimuth and elevation, we would need a fourth antenna, “+Z” at (0, 0, 1).

And this is all assuming that you can get -90 to +90 coverage for each antenna, which is unrealistic because you recommend a spacing of 0.45 Lambda, which will not give complete -90 to +90 coverage, so you’d need additional antennas to compensate.

So to have an anchor in the center of a room, and be able to determine accurate azimuth and elevation for a tag anywhere in 3D space around it, you’d need 5 or more antennas; is this correct?

6

Hi @Jordan ,

There would be a front-back ambiguity if you wanted to use 360-degree azimuth. Half lambda separation would only provide 180-degree coverage. You can also add 4th antenna (3 for azimuth and 1 for elevation) but that would increase the complexity of your antenna array and switch configuration.
In this config you can use 3 pairs in azimuth to cover 120 degrees and the 4th can provide you the elevation.

Kind regards,
Emre