In a previous POP test I found that you could achieve the stated 0.3mm accuracy on small turntable parts. I wanted to better understand the limitation of large area scans. I set up an experiment to scan a large piece of float glass (50mm x 1650mm) that I spray painted flat white using Markers for scan alignment. The plate is very flat and so you can Best-fit the 3D scan data to a plane. You can measure the deviation of the scan data from the plane. In my test of combining 100+ scans I got an error of 15mm over a distance of 1.37M.