T-Scan Hawk 2 Accuracy Issues

[[aw...]]

I recently purchased a T-Scan Hawk 2, and I've been using it for a couple of months. Overall, it has been a very efficient scanner for larger jobs my structured light scanner is unable to do. However, I only recently started testing its accuracy. For these tests, I'm using ball bars with matte white spheres that I sent out for calibration at Mitutoyo America a couple years ago. Since doing so, I've been using these ball bars to verify the accuracy of my Steinbichler Comet L3D 5M and my manual CMM.

From my testing thus far, the T-Scan Hawk 2 has not seemed capable of meeting its stated accuracy of 0.02 mm + 0.015 mm/m. I hope that I'm just doing something wrong and someone can steer me in the right direction. Here are the results I'm getting currently:

For the distance between the sphere centers, I was initially getting deviations as high as 53 microns over a 279.302 mm distance. Prior to this result, I had only performed a hyperscale calibration, so I then performed a full calibration. After doing so, the distance did start consistently measuring within spec, but it is concerning that I may not be able to rely on the hyperscale calibration to inform me whether the scanner will measure accurately.

The measured results for the 2 sphere diameters are even more concerning. Even after performing the full calibration, sphere diameters are measuring consistently more than 130 microns under their calibrated sizes of 25.405 mm and 19.045 mm.

So now I need to figure out what is going on. Again, I really hope I'm just doing something wrong, but I'm not sure what. I've tried recalibrating, I've tried using the scale bars with satellite mode before collecting the measurements, I've tried different scanning resolutions and different scanning modes, but nothing seems to help. I've reached out to my local application engineers to see if they have any ideas, so hopefully they can respond soon.

Has anyone else had similar issues? Does anyone have any tips for collecting in-spec data with their T-Scan Hawk 2?

I should also mention that I'm performing these measurements in a temperature-controlled location that is staying around 20.1 C.

[[Ch...]]

Hi awy, 
if the values do not match the expected one I assume,too 
that there is either an application topic or a problem with the sensor.
My advice therefore is to get in contact with your service provider, 
to analyse more detail.
For the T-Scan Hawk principally a steady hand and 
aquisition of the spheres in various angles is recommended. 
The Hyperscale itself should be taken from a optimal angle and distance see tech guide:
https://techguide.zeiss.com/en/zeiss-inspect-2023/article/cmd_sys_tscan_hawk2_hyperscale_recalibration.html

[[aw...]]

I did additional testing today and learned some interesting things. I added some Mitutoyo Cera blocks to my scans, and much to my surprise, these are measuring much closer to within spec than the ceramic spheres. Some are even measuring within the 20 micron spec.

So here are my current ideas about what is happening:

• The T-Scan Hawk 2 is inaccurate when measuring all spheres
• The T-Scan Hawk 2 is inaccurate when measuring small spheres
• The T-Scan Hawk is inaccurate when measuring matte white ceramic surfaces

I also have a shiny white 20 mm CMM calibration sphere that is measuring way undersize with the T-Scan Hawk 2, so I'm leaning towards the theory that the T-Scan Hawk 2 is not accurate when measuring all small spheres.

To further corroborate this theory, I noticed the calibration report that shipped with the T-Scan Hawk 2 only reports the diameter of the largest spheres on the artifact - the diameters are not reported for the 2 smaller sizes. 🤔

Test 1 - Default Exposure Mode - 0.2mm Point Spacing.pdf Test 2 - Bright Surface Exposure Mode - 0.1mm Point Spacing.pdf

[[Ja...]]

Hello, 

Im not familiar with the hawk but am with the structured light systems .

My educated guess in both cases is that the system is measuring sub surface rather than anything to do with spheres themselves.

In the case of ceramics this on a micro structure level be seen as semi porous such that light can 'sink' into the surface.

In the case of the cmm calibration sphere it is likely that this will have a laquer on it that the light will penetrate , leading to a very similar effect.

For measurement artefacts these are usually steel and coated with a fine level of coating applied by an airbrush.  

If you have access to such an airbrush and titanium dixoxide solution for example you could try this on both artefacts and see the result difference .

If you dont then your distributor should have.

You can get into the argument of 'how much media am i adding' but when prepared properly this is effectively lost in the noise of the systems as this layer can be approximate 10 microns approx applied  even layer such that the fundamental scale and size is not effected . 

I hope this helps

[[aw...]]

Thanks for ideas, James. One of my local application engineers ended up suggesting the same thing. I didn't want to coat my matte white ceramic spheres, as these are calibrated and they are useful for checking accuracy on my Steinbichler. But I did coat my one shiny CMM calibration sphere, and now the measured results for it are much closer to being within the accuracy specification (~20-40 microns off, still undersized). I expect that the remaining deviation is probably due to the most recent calibration and/or operator experience level, so hopefully I will get better results as time goes on.

This was a good learning experience for me. It has always been my experience up until this point that matte white is the ideal color and texture for 3D scanning, but it seems not all matte white surfaces will work, and it is good to be cognoscente of this possibility.