IWR6843ISK-ODS: Unexpected results from vital signs demo and out of box demo

Hi 

Thank you for your query. Please allow us a couple of days to respond

Regards

Hi,

Based on the picture shown, it seems the device is not able to track the person. What was the distance between you and the sensor? You can try the same experiment at 2m from the sensor

For the out of box demo, Can you use the Industrial visualizer in radar tool box?

You can access the Industrial visualizer at the below path

radar_toolbox_2_00_00_06\tools\visualizers\Applications_Visualizer\Industrial_Visualizer

Binaries for the Out of box demo can be found at the below path

radar_toolbox_2_00_00_06\source\ti\examples\Out_Of_Box_Demo\prebuilt_binaries

You can choose the configuration files from the below link

mmwave_sdk_03_06_00_00-LTS\packages\ti\demo\xwr68xx\mmw\profiles

Regards

Hi,

On the left side of the panel, the statistics shows there is one target. If you are looking for a 3D red bounding box in the plot, I did make that appear for a few minutes but still got a heart rate around 0. The distance between me and the sensor was 0.5 meter.

With respect to the out of box demo in industrial visualizer, actually I've already tried but I don't know what I can get out of it. As you can see from the screenshot below, there are just a few detections in the 3d plot. There is totally blank in "range plot" tab. No range information of the detections at all. I also read the user guide https://dev.ti.com/tirex/explore/node?a=1AslXXD__1.10.00.13&node=A__AHTBr6LTsIaFyYJEZ9kcmQ__radar_toolbox__1AslXXD__LATEST but didn't find anything new there. And the local industrial visualiizer (radar_toolbox_2_00_00_06) is quite buggy. If I send the chirp configuration again, it just stops working, showing either COM error or just stuck there. I need to reboot to get it work again. It is not as smooth as the online visualizer from the visualization & connection perspective. But neither of them can give me the proof of IWR6843-ODS performance so far...

Regards,

/Frank

Regarding demo visualizer, there is another thread below asking a similar question over a year ago. One of your colleagues confirmed it was a known bug that IWR6843_AOP had to be selected in visualizer for IWR6843_ODS target. What I am concerned most is whether the data/plot is right when selecting AOP. But in that thread, your colleague never came back with an answer to that and this known bug is still up there after a year. Based on my observation, the data is not right and there seems to be an approximately 20 cm offset in range...

https://e2e.ti.com/support/sensors-group/sensors/f/sensors-forum/1183221/iwr6843isk-ods-mmwave-demo-visualizer-platform-sdk-version-and-antenna-config

In addition, there is a LOAD CONFIG FROM PC AND SEND function on the plot tab. I've tried to load profile_2d.cfg from mmwave_sdk_03_06_00_00-LTS\packages\ti\demo\xwr68xx\mmw\profiles but it didn't work. First error compllained about invalid bpmCfg command and thus I just removed the bpmCfg line. Afer that, this error went away but no live plots any more. In the console, it shows an exception (see below). It seems this demo visualizer is not very friendly with IWR6843_ODS...

Back to industrial visualizer, it is problematic on another angle. I managed sending this modified profile_2d.cfg but there was no range information of the point cloud plot. If I want ot send it again with another profile cfg, it just stopped working by showing below errors. I've tried connection reset or restarting industrial visualizer and none of them works. The only solution appears to be unplug and then plug in the USB port, which is cumbersome. And the error message is not very helpful. If I want to test with my own profile cfg, I would like to have the error message telling me where the error is .e.g invalid bpmCfg or similar. 

I hope there would be a fix to above issues. Now I am wondering how I can proceed with the performance evaluation with the IWR6843 ODS given the issues with both demo visualizer and industrial visualizer. What would be your suggestions?

Last but not least, there is something unusual (see below) on the radar board IWR6843 ODS when we received it...We want to evaluate the performance as soon as possible because we need to negotiate with Mouser Electronics for a return if it turns out to be a defective unit.

Hi,

1. It is important to run the Out of box demo to make sure the setup is ok. If OOB itself is not running correctly the it is unlikely that Vital signs will run on your setup

2. EVM looks ok. It is a scratch on the ground plane. That should not affect anything

3. You can use the online demo visualizer by selecting xWR68xx_AOP and send the configuration from PC. You can generate the cfg from the sensing estimator at the following link https://dev.ti.com/gallery/view/mmwave/mmWaveSensingEstimator/ver/2.4.0/

4. There is no additional calibration required. Do you see the offset of 20cm for objects which are beyond 2m?

5. ODS board is specifically designed for ceiling mount. Vital signs demo is available for ISK and AOP only.

Regards

I agree that it is more important to get the OOB demo running correctly. AOP has the same antennas layout as ODS.I don't see why there is such a 20 cm offset, even if AOP is selected for the online demo visualizer.

I haven't tested beyond 2 m test as my intended work focus is vital sign detection, which is within 2 meters. I did test for 1.3 meter distance and there also exists an approximately 20 cm offset.

Hi,

I could not make out clearly from your response. But I want to confirm if you were able to run OOB with demo visualizer and Industrial visualizer with the new config that you generated using the sensing estimator. And 20cm offset is the only issue that you are seeing with OOB now

Regards

Hi,

I am able to run OOB with demo visualizer with the default config and another recommended config (profile_3d.cfg).

I confirm this 20 cm offset in range is the only issue I've seen as of now. I haven't digged into other aspects like azimuth and elevation. I feel it is most likely a quality issue as I brought up a corner reflector for testing today on demo visualizer and it shows approximately 15-20 cm offset. I've done similar tests with another radar before (Not TI) and got accurate results...So I would say the radar sensor itself rather than my bench setup is the cause of this offset.

I request to send back this defective unit and get a working one from TI or Mouser.

Best regards,

/Frank