Profile worse without datums

[[Mi...]]

Apologies if this has been covered before, but I can't find it with the forum search tool.

I have a part with a wire-cut profile that's supposed to be cut centered to the OD, clocked a particular way. It's called out to the top plane, the OD, and one side flat for clocking, and doesn't pass.

I strongly suspect the form of the curve is good, but shifted off-center from where it should be. But, when I report the profile without datums, it's actually worse than with datums. What is the reason for this?

[[Ma...]]

If datums for profile evaluation are unspecified, they will default to the base alignment.

[[Ma...]]

Mark - that's what i thought.
When i do this it switches to base alignment. I just can not verify it. Older versions can have this behaviour

[[Er...]]

For verification couldn't the Base Alignment features be entered instead of the Datums?

[[Mi...]]

Please sign in to view this quote.

Hmm... so how does one check profile to no datum reference if that's how it's called out on the print? Is that the same as the "form" characteristic?

[[Cl...]]

That's a good question. With No Datum Reference, is the geometry just free floating in space with no constraints?

[[Ma...]]

Just create alignment from curve or surface to apply on itself

[[Ma...]]

Please sign in to view this quote.

You could use the base alignment features but there may still be some differences to account for. These are the differences between the way Calypso calculates datums and alignments.

[[Ma...]]

Please sign in to view this quote.

Without datum reference, I think line profile deviations would be effected only by form and perhaps dimension.

[[Mi...]]

Please sign in to view this quote.

Please sign in to view this quote.

See attached. My base alignment in this case actually uses the same three datums as the profile callout, so it gets the same result. "No datum reference" makes it look worse still. And aligning to the curve (using "align to several curves", all rotations and translations enabled) gets a result somewhere in between.

[[Ja...]]

If you want to see just the form you could delete actuals, copy the curve, recall the points from the original curve into the new copied curve, under evaluation select best fit, under parameters you can select the rotation or translations options. Use the recalled curve in a profile of a line characteristic with no datum reference frame.

Capture.PNG

[[Ja...]]

Please sign in to view this quote.

I would guess it's in how Calypso is best fitting the curve. It's probably fitting them so more of the actual points are closer to nominal and when it does that your form error is more pronounced. And when you evaluate the curve to the DRF you're getting more of the points farther from nominal but you're extreme points are less pronounced.

[[Er...]]

Have you tried CAD Evaluation to see graphically what is going on with the different settings?

[[Aa...]]

Please sign in to view this quote.

Yes, but you can tweak an alignment to work like a DRFand vice versa. If you click the box with the symbol left of the name of the feature, it brings up a dialog with how the feature is evaluated. (In a DRF, the default is Outer tangential and constrained to preceding datum features. In an alignment, the default is Least Squares (LSQ) and not constrained to anything.)

[[Ma...]]

If you try CAD evaluation, then you can see raw data. I guess that you can have some bad data from start/end of scan.

With bestfitting in curve window you have more options against "alignment from curves". But only 2D curves gives you most options ( 3D can only have gauss )