Slot Perpendicularity and Position Methodology

[Ben1989]

I hope everyone is well.

I have a print that has caused me a tiny bit of confusion and thought I'd gain some valuable information from the PC-DMIS community.

I have a print that has a Datum A on a flat surface with a flatness call out and the parallel face with a callout. All good so far.

We then have Datum B as the first slot. How would Datum B be established from this? A cylinder measurement each end of the slot and a line created and rotated to?

Datum C is then the third slot. How would Datum C be established? Same method as I mentioned above except with a point between the cylinders? But it also asks for a position to B. How is this possible? It can float anywhere in the y-axis unless we have a temporary alignment of a rotation to the Datum B slot and an origin to a central point?

What are people's interpretations? The print isn't the real print but very similar. pdmis forum.jpg

Thanks in advance.

[gibsonridge]

as an initial alignment i would suggest measuring B & C as slots just to establish your most equalized rotation.
your manufacturing process is going to be a big factor here.
punched slot (taper from top to bottom) vs. machined slot (more consistent surface) .
these could influence your results in whichever way you decide to measure.

[Ben1989]

Thank you gibsonridge. It's a composite component where the slots and profile will be machined.

With that said, what would you advise?

Also, what is the advice for measuring perpendicularity on a slot? A mid plane from two planes on the inner slot walls?

[gibsonridge]

not sure if pcd will do it (never tried it before) might try measuring the same slot at 3 depths construct a line thru them. then evaluate perp.

ill have to brake out the test block and see what happens

[gibsonridge]

got it to work measure slot 3 times each at a different depth. construct a point of each slot , then
construct a line thru those 3 points. eval perp. to plane

[Ben1989]

Great! Thank you!

[Ben1989]

What are people's approach with measuring profile on a part so thick? A profile measurement @ 5mm height or a couple of profile measurements at 3.33mm & 6.66mm?

[bfire85]

What are people's approach with measuring profile on a part so thick? A profile measurement @ 5mm height or a couple of profile measurements at 3.33mm & 6.66mm?

If it is a profile of a line, then measure at one level. Profile of a surface is multiple levels.

I think two levels would be OK. Three would be best. Is this a machined or stamped part?

[Ben1989]

Hi bfire85. It is a machined part. Thanks

[vpt.se]

Is this really the way the leader lines are connected to the slots on the original drawing? Normally, the lines connecting the tolerance and the slot feature also indicate in what direction the position/perpendicularity applies to (if non-cylindrical tolerance). Your image does not do this - intentional?

[Ben1989]

Hi vpt.se - yes this is exactly how the leaders are on the slots. The only thing that's removed from the original drawing is things non-applicable to the slots (such as profile, part marking etc)

[vpt.se]

Then you don't know which direction the tolerance should be applied to or if they simply forgot the diameter sign from the tolerance. Doublecheck with the deisgner.

[Ben1989]

So just for clarification, level to Datum A, measure slot B and slot C, origin on slot c and rotate to slot B? I'm confused how Slot B is asking for a true position when it has a perpendicularity callout in addition?

[Schlag]

FCF A-B actually constrain all the D.O.F. C will never be used for anything other than rotation, and that will be a LINE from B-C. If a DATUM can constrain a DOF it has to.

[Ben1989]

For the slot perpendicularity I've measured points in the slots (regular vector points) at two levels, constructed scans, constructed points from those scans and created a line to measure perpendularity.

Did that with both slots and have results of 0.001 & 0.000. It just seems suspiciously small.

[neil.challinor]

It would be better to measure the sides of the slot as planes, construct a 3D width from those planes and then dimension perpendicularity of the width.

[Ben1989]

But what about the other direction?

[neil.challinor]

For the other direction you would need to measure cylinders, set the active workplane so that you are viewing the part as per the RH view in your screenshot and construct a 2D width

[Matthew D. Hoedeman]

measure a plane (A), measure a slot (B), measure a slot (C), construct a line from B to C
Level A, rotate to constructed line, origin X & Y to (B), origin Z to (A).
Position of (C) is really nothing more than a 2-D distance between the (B) and (C) features.
(B) is XY origin for the 3rd slot, position of 3rd slot to the current alignment.
Perp must take into account both X&Y at the same time, can't do 0.05 in "X" and 0.05 in "Y" due to the fact that if it is off 0.045 in X and 0.045 in Y it will NOT be within 0.05 perp.
Pcdmis doesn't have a "slot" cylinder of any kind, which would be ideal.
But, how to do it correctly with what Pcdmis has.... that's a good question.
You could measure a slot at the bottom and a slot at the top, construct a point at the center of each slot, then a 3D line from point-to-point, and do the perp of that to (A).

[Ben1989]

measure a plane (A), measure a slot (B), measure a slot (C), construct a line from B to C
Level A, rotate to constructed line, origin X & Y to (B), origin Z to (A).
Position of (C) is really nothing more than a 2-D distance between the (B) and (C) features.
(B) is XY origin for the 3rd slot, position of 3rd slot to the current alignment.
Perp must take into account both X&Y at the same time, can't do 0.05 in "X" and 0.05 in "Y" due to the fact that if it is off 0.045 in X and 0.045 in Y it will NOT be within 0.05 perp.
Pcdmis doesn't have a "slot" cylinder of any kind, which would be ideal.
But, how to do it correctly with what Pcdmis has.... that's a good question.
You could measure a slot at the bottom and a slot at the top, construct a point at the center of each slot, then a 3D line from point-to-point, and do the perp of that to (A).

Thanks Matthew. Isn't that how I did it however but with constructed slots and not auto slots? The reason a took vector points was to allow me to use tip changes due to shanking risks. Unless there's a way to tip change half way through an auto feature? It's 2009 being used

[Matthew D. Hoedeman]

flat piece of stock, 10mm thick, 4.5 to 8.0mm slot (minimum size). I could run that all day long with a 3x20 probe and never worry about shanking.

[Ben1989]

I have a 2x20. I'll test the slot rads at different angles to test for differences. Hopefully none appear and that'll save a lot of time constructing. Thanks for your help