Simulate Planar Multiple Datum

[DungT]

How would you do it?


Untitled2.png

[acgarcia]

maybe on a vision system.

[Caemgen]

Do you expect them to be that bad?

If it is normal machining, and you expect them to be "good," then measure both planes (ensure vector and distance nominals per hit are 20mm apart as per that drawing), create plane - offset (make sure the dialogue in the Set Offsets window are correct). Might have to pick three features.

Edit the construction after each plane put:
.hits[1..planeX.numhits]

delete the third feature if you had to enter one.

If you expect the part to be that bad, you should make a simulator and align it with the offset plane. If you can, maybe set the part on the surface plate of the CMM with gage blocks under the steps. You can inspect what you can get to, and then flip the part and use the measured features for the new alignment, like you would if the part was too big and you were sliding it through the machine.

If you can't convince anyone to make you a step block, try taking a bunch of hits in lines, pull the hit closest to to whatever axis, -Y in that drawing, creating points from those high points and make an offset plane from them only.

[DAN_M]

This doesn't make sense to me....

Surface profile is a form AND location tolerance.. the position is only being controlled in one axis. You'd need control in two axes for this DRF to be valid (...right GDT gurus?)

Assuming we're looking down from the Z+ perspective....the offset plane would control everything in Y...but what about X?

[DungT]

Do you expect them to be that bad?

If it is normal machining, and you expect them to be "good," then measure both planes (ensure vector and distance nominals per hit are 20mm apart as per that drawing), create plane - offset (make sure the dialogue in the Set Offsets window are correct). Might have to pick three features.

Edit the construction after each plane put:
.hits[1..planeX.numhits]

delete the third feature if you had to enter one.

If you expect the part to be that bad, you should make a simulator and align it with the offset plane. If you can, maybe set the part on the surface plate of the CMM with gage blocks under the steps. You can inspect what you can get to, and then flip the part and use the measured features for the new alignment, like you would if the part was too big and you were sliding it through the machine.

If you can't convince anyone to make you a step block, try taking a bunch of hits in lines, pull the hit closest to to whatever axis, -Y in that drawing, creating points from those high points and make an offset plane from them only.

Sheet metal, bend and twist different from part to part. I guess step block is my best bet

[DungT]

This doesn't make sense to me....

Surface profile is a form AND location tolerance.. the position is only being controlled in one axis. You'd need control in two axes for this DRF to be valid (...right GDT gurus?)

Assuming we're looking down from the Z+ perspective....the offset plane would control everything in Y...but what about X?

I borrow that sketch from ASME Y14.5M 2009 page 67