Lidar Lite V3 Accuracy Issue

I tested Lidar Lite V3 with Uno using the Garmin library I2C example. The specs says the reading accuracy should be within ±2.5cm for <5m range and ±10cm for the rest, but my result didn't match at all.

I pointed Lidar Lite towards a wall at different distances, recorded ~1000 samples at ~30Hz rate at each location and checked reading fluctuation:

Location Fluctuation
1m 6cm
3m 9cm
7m 17cm
11m 52cm
18m 101cm

After 20m range, I could barely get any valid reading even though the measurement range should go to 40m written on the specs.

My connection exactly followed the diagram on the specs and I tried both default and maximum range modes.

Could anybody help me with this? Did I receive a faulty unit?

Thanks

Are you using a "70% reflective target" as per the documentation?
The device uses 905nm so there is no guarantee that something that appears reflective/white to you in the visible range is also reflective at that wavelength.

Are the values you see constant offsets or variations about a mean?

I pointed it at regular wall, I assume this should be a very normal use case.

The values are fluctuating around a mean.

Do you see the same behaviour as this #203 Best LIDAR Sensors for Makers (Comparison and Test) - YouTube

Yes, similar. His result seems to have less fluctuations than mine. Also my test beyond at 20m was like what he had for 40m, a lot of 1s and random erroneous values with several correct readings every now and then.

Point it at some different walls, and let us know what happens.

If you fancy a bit of decorating visit your local hardware store and look for 'heat reflecting paint'. These are loaded with various metal oxides that reflect in the near infrared.