I’m having an accuracy issue with my Metro Y Ultra CMM scanner and was hoping someone here might have some suggestions.
Whenever I scan an object and compare it to the original CAD model or the physical reference part, the scan is consistently off by about 0.005 inches.
Most of my testing has been done using metal gauge blocks, but I’ve also scanned several non-metallic objects and observed essentially the same amount of error. Because of this, I don’t think the issue is related to the material being scanned.
Here’s what I’ve already tried:
Recalibrated the scanner more than four times.
Performed the calibration in a temperature-controlled room that matched the original factory calibration temperature.
Verified that the calibration completed successfully each time.
I also performed the ball-point measurement accuracy test, and I sometimes receive an error of approximately 0.02 mm, which is much higher than I expected.
For my intended application, I need scan accuracy of ±0.001 inches or better. A consistent error of 0.005 inches makes the scanner unusable for the measurements I’m trying to perform.
Has anyone experienced a similar issue or have suggestions on what else I should check? Could this be caused by a hardware problem, calibration issue, software settings, alignment method, or something else?
Any advice or troubleshooting recommendations would be greatly appreciated. Thank you!
I have an ultra too - not really used yet apart from a test scan to verify if the accuracy is actually within claimed specs and if I want to keep it:
Test object was also a gauge block, but ceramic to avoid needing spray. Result was the scan results having quite a bit of surface noise. I then imported the mesh into GOM and fitted two surfaces at 5 sigma on the business ends of the jo block scan. The measured distance result was within 2/100mm (~0.0008") of the nominal value on a 60mm or 40mm block (don’t quite remember which). The noise though (orange peel surface) was easily 8/100mm peak-to-peak.
So my point is, how you measure exactly makes quite a difference.
I did not get the ball plate - no real use for me as it is only a scanner-is-good test object for rubber stamping meas protocols and doesn’t influence the results - so I can’t speak to that.
Also the scan did not take place under controlled ambient temp - cera blocks at roughly room temp vs basically a handheld toy scanner with a plastic body and active electronics blown with a fan inside and claimed 15um accuracy makes the whole temp controlled room thing a bit of a moot point in my opinion.
There can be a lot of variables at play. What temp are you scanning at (temp of the scanner, not room temp)? what temp was it calibrated at? If the scanner is too far out of the range you calibrated it at then there can be deviations. How is the surface of the block? Is it super polished? That can cause a lot of noise in the scans. How clean is your scan? How many frames are you capturing? What mode are you scanning in? Turntable is going to be the most accurate mode.
The ball plate is more of a sanity check. Is the scanner working as it should, does it need to be calibrated. Basically says “hey the scanner is capturing good data”. You said you “sometimes” get a .02mm error. What are your other values? I sometimes get very poor gas mileage in my truck, that’s not indicative of my gas mileage all the time.
Can you share more about what you are trying to measure and need .001" measurements? They call it the CMM, but you are not going to match equipment costing 30K. The results you get at this pricepoint are pretty darn good.
My suggestion is to do a bunch of scans and see if the scan is consistently too small or consistently too large. If you compare the scans and the scans are all accurate to themselves then it’s not an accuracy issue. Then the issue lies in calibration or workflow somewhere. I personally had a tough time getting it calibrated correctly and for a while everything was coming out slightly too small.
Hi.
Like @Nickluvin said, perform the calibration at (about) same ambient temperature at which you will be scanning later. Also, let the scanner run for about 10 minutes or so before you do the calibration so it reaches the “working temperature” as this can impact the accuracy. The same before you do the actual scan.