Reliable fit of small cylinders (3.8)

[[Re...]]

Currently we are validating geometric quality of 3D printed parts with narrow tolerances. 

On TPD drawing datum system ABC is applied on small planes/cylinders for long parts. I tried to gather as much mesh on or in these spots so that the fitted planes and cylinders are as reliable as possible. Before the ABC alignment a pre alignment is set where the whole part is best fit. I notice that whenever I rescan the same item, the results of position accuracy tolerancies vary and in some occasions lead to out of specs. 

My question is via which ways I can improve the fitting of small cylinders?

Sometimes the fit results in a slightly different direction and has big impact on all other measurements. 

I read on internet that a possible solution is to make sections where circles can be fit. With these circles I can construct a more reliable cylinder.

 

[[Ma...]]

You can play with angle selection of fitting element.

Also you can use bestfit on certain areas of a mesh to improve CAD<->mesh fitting

[[Re...]]

Hi Martin, 

Thanks for fast response. 

Angle selection of fitting element U mean to play with the Max Angle in fitting element; 

Your second remark: 

I already have a pre alignment with best fit of full part. You say it will help if I create another best fit alignment based on area's on datums or area around it?

This means I have to select MESH data manually with select through surface and make a new best fit alignment? 

This scanning program needs to run via kiosk without any manual steps, is this possible? 

 

 

 

 

 

 

 

[[Ma...]]

Increasing Max. angle can bring more selected mesh.

For bestfit - you can select areas on mesh, which can bring mesh into a correct position.
If it's rotary part, then it will help only with rotation.

Otherwise if you have problems with depth and or angle, then you can select only one mesh-cylinder for bestfit and measure only this one within bestfit alignment.
You can select measured cylinder and in properties you can select to calculate it only in certain alignment.

[[No...]]

Can you elaborate a bit more on what you mean by "small cylinders"? Is it a "short" cylinder (large diameter, short axis) or is it just a cylinder that is so small your scanner resolution doesn't give you enough points to work with?

Short axis cylinders are generally susceptible to a certain amount of tilting when their mesh surface is not perfectly even.

All that was said above CAN help, but depending on how much the surfcae deviates from a perfect cylinder, you may still end up with a tilted result. If all else fails your last resort is to add constraints in the measuring principle. For example, if the cylinder protrudes from a flat plane and you know it is perfectly rectangular, you might set the cylinder's axis orientation to that of the plane's normal.

 

Edited 2 hours ago

[[Re...]]

Cylinder diameter 3.8, axis length is 7. 

I have 334 points on that cylinder for example, that is quiet high I assume.

What kind of setting would you recommend?

I will try out a couple of your tips.

[[No...]]

I think your problem is not so much the parameters of the measuring principle, but the scanning coverage of the part surface. With so many holes in the mesh I'm not surprised you get a bit of spread in your results. How do you prepare the part before scanning? 

344 points should be enough, but it depends on how they are distributed around the cylinder. 

Edited 36 minutes ago