I’m having trouble getting my scan measurements to match the actual dimensions of my models. Despite trying a lot of different approaches, my scanned dimensions are consistently 0.3–0.5mm off from ground truth.
What I’ve tried:
Recalibrating with the calibration board (yes, warmed up to operating temp)
Verifying with the included CMM-certified carbon fiber ball plate
Single-line, crossed, and parallel scan patterns
Marker tracking and global marker tracking
Fusion point distance from 0.1–0.5mm (results vary significantly across this range)
Scanning with and without the turntable moving
Moving to mesh (this actually worsened accuracy)
My closest results are still ~0.3mm off, achieved with single-line scanning, global marker tracking, and no turntable. Repeatability is good (scanning the same way twice gives consistent results within ~0.1mm) but absolute accuracy is the issue.
Scanner: Revopoint MetroY Ultra
Model size: approximately 30×30×40mm
0.3mm may sound minor, but for my use case it’s beyond acceptable tolerance
Since my scans are dimensionally consistent with each other (even if offset from ground truth), would uniformly scaling the mesh up or down to correct one known dimension translate accurately to the rest of the model? Or does the error vary across the part in a way that would make a single scale factor unreliable? This seems to be what the CMM certified carbon fiber ball plate helps with, though it does not seem to help me here.
The scans look visually excellent. The issue is purely dimensional.
Are there any approaches I haven’t tried? Any help would be appreciated!
Thanks for listing out the things you’ve tried. But I’m still a little confused.
When you scan the ball plate what are your results? Are you .3mm off on that or is your error pretty low? Or is it the other object you are scanning that is off?
How many frames are you collecting? Can you share a picture of your setup? Are you scanning a shiny part in regular mode and getting reflections?
Laser scanning and turntable dont really go together, you want to be moving the scanner to collect data in different areas instead of fixed. Automode is really what the turntable is for.
If your accuracy test using the balls plate is off between 0.3 -0.05 mm and you did not dropped it before , please contact the customer@revopoint3d.com for technical assistance .
You should not have that huge error as that is not normal , even 0.0065 is already too big error.
Thank you both. The ball plate is only 0.012 off. I’m convinced it is user error of some sort. I am coming from a MetroY (not ultra) and was experiencing the same issue on that scanner as well.
Nickluvin makes a good point about reflections. The parts are silicone molds and to have some shine. Automode doesn’t allow me to select shiny metallic object type, so that could be the issue there. Unfortunately the white spray does not adhere to the silicone molds so I may need to find another option there.
Handheld I was getting around 2500-4000 frames. Automode much less. Around 100 if I recall correctly. I’ll experiment more when I’m back in the shop and send some photos of the setup. I appreciate your help!
Auto mode should have no more than 360 frames per one rotation and the same for Full Field, laser mode need at least 8000 or more frames .
If you over scan small object using Full Field, you going to make things worse , but for Laser mode , how more frames how better , lately I am doing at least 20 K in laser mode.
Silicone shine is not big problem , silicone as material yes because it is absorbing and the measurements can get off as the light is measured much deeper as it should , 3D sprays scan add also to the accuracy as they usually are not for high accuracy scanning , if you can find some water soluble paint that you can use on the silicone it would be better , or a powder . Many auto scanning service uses water paint on cars that are hard to scan due to its surface and it is easy to clean out . Do some research online .