Honeycomb measurments.

[[Ro...]]

Hello GOM team,

I am currently working on a correlation study between a Scancobot with Atos Q and a CMM using Calypso. We are measuring a radius section of 194 mm on a honeycomb part.

We have observed differences in the measurements obtained from both pieces of equipment. Specifically, we are measuring three sections at different heights, over an approximate length of 15 mm, with 7 points per section.

On the Scancobot, we use a touch point edge measurement with an edge height of 4 mm. On the CMM, we measure the same points but use a 2.5 mm radius probe.

We have noticed that the point-to-point variation is greater on the Scancobot than on the CMM. The dispersion between points on the CMM is about 0.01–0.03 mm, whereas on the Scancobot it is 0.04–0.07 mm.

Do you have any recommendations for measuring this type of part?

[[TI...]]

Put some green highlighted points(or section cut) you are using to create a radius on that mesh....no one knows where you are measuring the radius at with just a picture of a mesh.

Edited yesterday at 01:25 AM

[[Jo...]]

Hello Rodrigo,

we need way more information.

im just thinking in writing ...

• measurement volume Atos Q
• surface of the part
• are you using any matting agent?
• what exactly are you measuring from that picture?
• how is the mesh looking in this region in detail? how solid is the mesh aquisition on the scaner?
• polygonisation which settings?
• clamping the same way or not at all?
• why compare single points and not results? that will always put the scaner in a very bad position. Use the information available like you would in real world measuring task and dont try to use the scaner like it is a cmm with a switching probe.
• Radius: how much of a full circle?
• point (edge) 4mm vs ballprobe 2,5 mm is a vastly different thing, think about using something else for the scanner.
• Why is your cmm so inaccurate? Could be a hint to a problem in the approach.

Edited 18 hours ago

[[Ro...]]

Hello.

Thanks for youre support, about your questions:

• measurement volume Atos Q

Atos Q, 12M

• surface of the part

Cilindrical surface on the honeycomb

• are you using any matting agent?

Yes, TiO2 it´s aplying with a airbrush

• what exactly are you measuring from that picture?

Rad H, the next are the measuring concept. We use a touch point edge measurement with an edge height of 4 mm. 3 sections and 7 points per section.

• how is the mesh looking in this region in detail? how solid is the mesh aquisition on the scaner?
• polygonisation which settings?

We use standard poligonization

• clamping the same way or not at all?

The part was measure in freestate

• why compare single points and not results? that will always put the scaner in a very bad position. Use the information available like you would in real world measuring task and dont try to use the scaner like it is a cmm with a switching probe.

The problem uf use other method, like a cylinder diameter, is the thickness of the honeycomb and the section or the diameter.

• Radius: how much of a full circle?

it's only a 18/360 of the total diameter

• point (edge) 4mm vs ballprobe 2,5 mm is a vastly different thing, think about using something else for the scanner.

Ok, do you have a propose?

• Why is your cmm so inaccurate? Could be a hint to a problem in the approach.

+/- 0.0027 mm aproximatly

 

 

Thanks for your support.

 

[[Jo...]]

I am still not sure what you want to measure 🙂

Do try to get the highest point on the radius of the walls of the hexagons?

 

The measurement volume is the optic set used on the atosq 12m that determines the size of the box that gets data with one shot. For example MV100, 170, 350 ... 

The reason for asking is, that this is (next to the camera sensor resolution) the second big decider of the min point distance. For small radii or edges you need a close point distance.

From the pictures i would suggest trying out different sensor angles to the part and different settings to get better data. There are alot of artifacts around the edge/radii.

 

From this picture you can tell something is off by the normal vectors of the points (green). They should be all roughly like the vector of secc1.

Also i would reduce the length of the edge of the point edge tool, so you can determine were it touches better, maybe try touch point disc with a very small disc.

The disc could help with a bad mesh.

 

Edited 11 hours ago

[[Ro...]]

Hello Jochen,

Sorry for the confusion.

We are trying to measure the radius at the highest point on the top surface of the honeycomb structure.

The measurement volume used on the ATOS Q is MV170.

Another challenge is that the position of the honeycomb may vary between different parts, and the wall thickness is only 0.1 mm, which requires high-resolution scanning and careful alignment.

The irregular vector orientation in the scan data is due to the plausibility check being disabled. We had to disable it because the system would otherwise reject valid points on the top of honeycomb.

We suspect that during polygonization, the software may be filtering or simplifying some of the scan data, especially in areas with sharp transitions or thin walls. We will review the polygonization settings to preserve more detail.

 

Thankyou very much,

Best regards,