Datum feature / secondary alignment

[[Cl...]]

I'm evaluating a 3d curve for line profile (.003 -A-). In one instance I use the datum feature, and in the
second I use a secondary alignment created with the same feature. Which method is correct/preferred?

[[Ja...]]

They are within .0001, are you in inches or metric? What is the breakdown of the secondary alignment?

[[Cl...]]

Inches. The secondary alignment uses the datum -A- cylinder for planar rotation, and ZX zero.

[[Ja...]]

That's odd, I would expect to see those values reversed. I think it's more common to spell it out to your FCF then use an alignment. Spelling it out allows Calypso to create an internal alignment to your datum structure where the deviation values will make sense (i.e. basic dims). The issue with referencing an alignment in your characteristic would be if the alignment wasn't defined as GD&T standards would define it (Proper treatment of allowable DOF and constraints). Then Calypso just spits out some stuff and hopefully it is OK. If your secondary alignment has Y, Z origin at -A- and so does your base alignment, then I wouldn't expect to see a difference unless your Spatial/Planar rotation are drastically different (dealing with unrepeatable features like small radii or castings).

Putting my trust in Zeiss' algorithms, I would argue that spelling things out to your print would be more reliable than secondary alignments.

We don't report that far out, do you think this could just be a roundoff/precision error? Your difference could be as small as .000080 which is really representative of the accuracy of the typical machine (2 um).

Also, why are your two 3D curve features named differently from one another?

[[Ri...]]

Calypso really likes a fully built DRF.

Most times when you leave something blank, it will pull stuff from the Base Alignment.

I would suggest turning on Best-Fit inside of the Curve, and allowing it to create an alignment based on the Degrees of Freedom you want Best-Fit (the Best-Fit Alignment can be used in your Profile Characteristic). In this case, you will need to ensure that the Curve is measured to an Alignment that uses that Cylinder.

I also suggest this for Profile of a Surface - use a Free-Form and create a Best-Fit Alignment.

[[Me...]]

Feel free to correct me if I'm wrong, but is this a case of the difference between the outer tangential/ datum reference settings to ISO 5459 being enabled or not?

[[Ro...]]

This is one of the things that PCD does much better than Calypso.
In PCD you assign features as your A B & C datums, then your GD&T features are created by building your FCF, including all modifiers & boundaries & special stuffs.
it takes all the guess work out of it because there's only 1 way to do it.
and on top of that the FCF appears on the report.......
Come on calypso, lets get this done!

[[Cl...]]

Please sign in to view this quote.

To the best of my knowledge, I can't use "Best Fit" as this is not a profile to itself. A datum is involved.

[[Ri...]]

That's the beauty of Best-Fit Alignment, you control which degrees of freedom are/aren't restricted.

[[Me...]]

Please sign in to view this quote.

I completely agree. Xactmeasure in PCD works pretty sweet. In addition to the FCF being shown, which gives a warm fuzzy to the customer because it matches the print, it reports simultaneous requirements wonderfully. You create all your FCF's and then just recall all the ones you want SR for. It works a treat as they say across the pond!

[[Cl...]]

To the best of my knowledge, I can't use "Best Fit" as this is not a profile to itself. A datum is involved.
[/quote]

Is that correct, can best fit, inside the curve feature only be used if the profile is to it's self, and no datum is involved?

[[Ri...]]

No - which is why it allows you to pick and choose the axis you want free or locked.

[[Cl...]]

Please sign in to view this quote.

OK, so I have 4 planes to evaluate. Two on each side of the datum -A- cylinder.
The planes are perpendicular to one another. One is horizontal (XY work plane)
and the other is vertical (YZ work plane)

Under "Best-Fit" would I constrain just Z for the horizontal, and just X for the vertical?

[[Br...]]

Please sign in to view this quote.

Can we see the drawing by any chance? Its hard to determine exactly how this should be handled. I feel like we're missing a lot of information. I'll try to use a little deductive reasoning to visualize your application here.

So first of all you say you have plane features to measure here, and you want to do a profile of line on them. If "A" is the only referenced datum in the DRF, and they all have the same DRF, then you have a simultaneous requirement for all of these. In other words, you can't do a best fit of each individual profile and report it that way. What you need to do is create a "Best fit of several curves" alignment that references your "Alignment -A-" secondary alignment. You then will plug in all your curves of those planes. (you could probably just measure them as planes and do a geometry best fit as well.) Typically a profile of line tolerance is going to be a refinement to a Profile of surface tolerance. So this same alignment will have to get plugged into those characteristics as well if they exist.

Please sign in to view this quote.

Why are you using the Datum A cylinder as a planar rotation? If this is a primary datum feature, you should be using spatial rotation on the Y+ or Y- axis if the cylinder's axis origins are X and Z. What is happening here is that the spatial rotation will default to the base alignment. That's one reason why you are getting two different results.

Once you get the curves into the best fit alignment, You'll want to check the translation box for "along Y", and the rotation box for "around Y".

[[De...]]

The best fit should allow them to rotate around the A datum (you said its a cylinder) and also along it if the only profile callout is to A. The way I think about it is like this. You have the Datum A that locks certain degrees of freedom - so the part is restrained in those degrees of freedom. Whatever is left over is what the best fit should be able to 'best fit'. In this case - a single cylinder will lock two translational degrees of freedom and two rotational degrees of freedom. The rest - rotation around the cylinder and translation along the cylinder can be best fit.

If you are using multiple curves and they all have the same call out to A, then they will be simultaneous requirement though.

You can still use best fit around/along A, but in order to do so I think you will want to use 'Alignment from Several Curves' from the menu.

Menu -> Resources -> Utilities -> Alignment from several curves.

If you best fit each of the curves in the curve itself, you will have four separate best fit's which will each have their own rotation in relation to A.

If however, they are in fact just planes, then I would think they should be checked with planes rather than curves and this could still be accomplished using a geometry best fit alignment.

[[Cl...]]

The two surfaces in red are being evaluated. There are two more opposite side and 180 deg from the two visible.
The print states surface profile to -A-, but we get better results with using 3d curves and line profile. The customer
is OK with that.

[[Br...]]

Ok so your "Alignment A" should use the datum feature A cylinder as a spatial rotation(Z+ or Z- axis depending on the space axis of the feature.), X origin and Y origin. Use Outer Tangential, and whatever filtering and outlier elimination is necessary.

Next go to Resources>utilities>Alignment from several curves

Click the select elements button, and select all the 3D curves you are using. In the Alignment Dropdown select your "Alignment A" alignment. Then click the "Evaluation Constraints" button and only check the boxes for "Best Fit", "Along Z" and "Around Z" everything else should be un-selected.

Now use this alignment for all your profile and position characteristics that only reference datum feature A.

[[Cl...]]

Thanks for the help guy's. I'll give the various tips a try.