Fresnel Zone - Ranges error?

Hey there!
I have a question and hope someone could actually validates.

I have an MDEK1001 evaluate kit and coded my own firmware and did some calibration on antenna delay.
When i measure distances with LOS scenario i have a low error, less the 10 cm. Then i want to test some scenarios like NLOS and Fresnel Zone occupation.

I made some ranging tests with 3 situations: 1- LOS, 2 - NLOS(with a pvc board on the way) and 3- metallic board almost on NLOS but the tag and anchors have LOS (Near NLOS).
I show here 2 images with the error results:

The test distance was 9110[mm] and the results and the error showing on the histograms. On the caso of pvc the impact is +/-2cm but on the other scenario the results are kinda weird…I wonder if someone had something like that and this is a fresnel zone effect ?

Did anyone get some results like this? Someone find out what they were?