Unable to apply MMB

[JacobCheverie]

https://www.pcdmisforum.com/image/gif;base64,R0lGODlhAQABAPABAP///wAAACH5BAEKAAAALAAAAAABAAEAAAICRAEAOw==Y14.5.1-2009 Fig. 4-31 (c) (and 4-30) has a planar datum feature referenced at MMB. PC-DMIS cannot apply MMB modifier to planar datum features (because they are not features of size?) Does anybody have situations like this where they just cannot grant the bonus? Isn't it possible to pass a bad part without referencing B at MMB?

4-31.JPG

[bfire85]

Create Datum B as a width.

[JacobCheverie]

Datum B is not a width, it is a planar surface.
Where would the other side of the width be?

[bfire85]

I think you have to treat it as a width. You would need to run a parallel plane/line on Datum A, then do a width that way. I think that would be the only way for PC-DMIS to get MMB on Datum B.

[joshmxz]

https://www.pcdmisforum.com/image/gif;base64,R0lGODlhAQABAPABAP///wAAACH5BAEKAAAALAAAAAABAAEAAAICRAEAOw== Isn't it possible to pass a bad part without referencing B at MMB?

No. You could reject a good part without referencing B at MMB. Just remember 4.19 within the standard (most people don't know about the simultaneous rule) and those who don't want to deal with it just forgo the MMB bonus you would otherwise obtain.

my $0.02

[JacobCheverie]

My logic is that if the designer establishes the Position tolerance with B at MMB, the holes may be bad as measured from B at MMB. Now if B is not perfect as image (c) shows, referencing B at RMB will allow the simulator to shift from MMB to LMB until maximum contact is made. If the holes were OOT with B at MMB, they may shift into tolerance by referencing B at RMB because the part will rotate about the axis of A while fitting for maximum contact.

[A-machine-insp]

I see what they are doing. The .020 profile of datum b is there the MMB is going to come from. I agree with bfire85 that a width is the best way. Try constructing a plane at the datum B nominal location (that does not move) then create a width between that plane and the actual datum B plane. The nominals for both planes should be the same so the width nominal should be zero. Any boundary modification would come from the resultant width.

This is what I thought of after waking up VERY late this morning and almost getting to work late to take it for what it is. That being said, it makes sense to me.

[JacobCheverie]

The problem is that B at MMB is at 5.1 (in this case). If we measure B at 5.0, it is conforming to it's Profile of a Surface callout, but the part needs to spin about the axis of A until the plane at 5.0 makes contact with a theoretical plane (MMB) at 5.1. I don't see how a width, or centerplane, would accomplish this.

[A-machine-insp]

I don't think it has so much to do with rotation as it does translation. The constructed plane/width would have the same vector as the original taken plane. The difference is going to be where it translates to. That is where the modifier will kick in and effect the dimension.

[Matthew D. Hoedeman]

I don't see how datum "B" can have (M) at all since in none of those example does it do more than control a degree of rotation, not an axis location

IMO (fwtw)

CAN-MAY-MUST

Datum A controls 2 degrees of rotation and 2 degrees of axis location. Datum B only controls a degree of rotation.

[JacobCheverie]

As you are alluding, anything referencing [A|B(M)] is measured in a coordinate system aligned to the True Geometric Counterpart (TGC) of A, centered on TGC of A, and clocked to TGC of B at MMB - a collection of size effects (non-existent) and geometric control effects (in this case Profile of a Surface to A). The reference on B is valid in that B, at it's extreme throughout it's profile allowance, will be in a position that will yield max part material. The rotation about the axis of A is stopped by this plane, at MMB. How do you simulate that in PC-DMIS?

[Matthew D. Hoedeman]

I don't really see a way to do that, and personally, I would question that entire page in the bible. BUT, if the 'high point' of B is the only thing that can be used to stop the rotation (and I don't agree with that) then the low point would be the point to use along with the high point, to create a width feature.

IMO, measure a many-point line on "B" and that is the rotation feature. NOT a high point, and there shouldn't be a (M) on the "B". And the picture is stating to use a point at 5.1 (5+/-0.1) as the datum POINT for the datum FEATURE of "B". Huh? That means that a PERFECT part, absolutely 100% perfect part will have the "B" datum rotated out of square. That's just not right! Your perfect part will show LOTS of error instead of zeros (yes, I know, no such thing as a perfect part).

[JacobCheverie]

Matthew D. Hoedeman do you think that the availability of such a call out would be a way to ensure that the profile always stays on the high side instead of maybe holding a much more restrictive profile callout? I'm not sure. That's a very good point that a dead perfect part would yield error.

I think PC-DMIS is limited in it's evaluation of this callout - especially if trying to use Xact. It may be possible to create a theoretical plane at MMB, measure the actual plane of B, measure the angle between the theoretical plane and the high point plane and rotate the alignment by the same angle?

I don't actually need to do this, I'm just trying to see if PC-DMIS is capable.