Hole pattern

[TrailMaker004]

The (2) .119 diameters on the attached dwg are raised bosses. The .226 diameter is a hole. Datum A is the common planar surface these features are normal to. What we are trying to accomplish with the GD&T is first hold the 2 bosses center to center dim to .000 TP if at MMC with bonus tolerance as the boss sizes both on the datum D feature and the opposite reduces (this is for a mating part fit). We then want to establish the location of this 2 boss pattern relative to the .226 hole with a TP tolerance of .004 (i.e. the 2 boss pattern should be centered to datum B within a .004 tol zone). Rotation angle of the three feature pattern is not important at this point.

This should be elementary considering all the GD&T we assign here but we are scratching our heads. Need some feedback on this.
Thanks

 

[MechNorth]

Composite position, not two single segment controls. First level wrt A/B(M)/C. Second level wrt A/B(M). Should take care of it for you. Eliminate the Datum-D as it is redundant.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[powerhound]

Jim is right but the second level has to be a refinement of the first so be sure the .004(M) is in the first level WRT A/B(M)/C and the 0(M) WRT to A/B(M) is in the second level.

Your description of what you want to do is the reverse of how it's described in the standard. What you want to do FIRST is locate the pattern (top level), then you tie the two features together to a smaller tolerance zone (bottom level).

John Acosta, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[pmarc]

There is also another option, without referencing to C.

Upper segment of composite positional FCF - dia. .004(M)|A|B(M)|. Lower segment - .000(M)|A|. Actually, in that case there would no difference whether the FCF was composite or two single segment.

As for datum feature C, it could be controlled by |pos.|.008|A|B(M)|, so that simultaneous requirement would deal with rotational relationship between width C and pattern of holes.

All depends on how tight one wants to control this rotational relationship.

 

[MechNorth]

Pmarc,
If you don't include datum C, then the pattern of 2 holes can rotate around datum-B. C is needed here.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[pmarc]

It can indeed rotate. But is it a problem? If width C is controlled by |pos.|.008|A|B(M)|, it will rotate together with the pattern due to simultaneous requirements rule.

 

[MechNorth]

Fair enough.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[TrailMaker004]

Thanks for clearing that up for us guys.

The one thing I don't understand is (when we will inspect these parts) how the bonus tolerance will be applied for the | .004 m | A requirement. If the one stud is .122 (.001 uhl) and the 2nd is .121 (.002 uhl, total .003) how is that applied (especially in CMM terms) when neither hole is the X-Y origin?

 

[pmarc]

You do not sum the deviations of studs sizes from their MMC.

For stud #1, if its diameter is .122, you have .005 of positional tolerance available (.004 from FCF plus .001 of bonus) for that stud.
For stud #2, if its diameter is .121, you have .006 of positional tolerance available (.004 from FCF plus .002 of bonus) for that stud.

Centers of two as-described cylindrical tolerance zones are spaced basic .390 apart and perfectly perpendicular to datum plane A. If axes of both holes fall within the tolerance zones, as-produced pattern meets print specification.

 

[TrailMaker004]

Thanks, I just realized I made a mistake with my last question.

The area I am concerned about how to apply the tolerance in is the .000 (M) | A | requirement (not the .004 (M) | A | B | requirement). In that case, there is no X-Y origin so in the instances stated (,122 on 1st, .121 on 2nd) above I am not clear how this would go.

 

[pmarc]

I understand that due to some network problems my reply has been posted twice, but I do not get why I keep writing "holes" when the question is about "studs"

So this is how my response should look for dia. .000(M)|A| requirement:

"For stud #1, if its diameter is .122, you have .001 of positional tolerance available (.000 from FCF plus .001 of bonus) for the stud.
For stud #2, if its diameter is .121, you have .002 of positional tolerance available (.000 from FCF plus .002 of bonus) for the stud.

Centers of two as-described cylindrical tolerance zones are spaced basic .390 apart and perfectly perpendicular to datum plane A. If axes of both studs fall within the tolerance zones, as-produced pattern meets print specification."

 

[TrailMaker004]

Thanks. Bear with me if I am a little slow here. In the FCF .000(M) | A | these studs are positioned only with respect to each other (and perpendicular to A) . As such there is no datum that specifies an X-Y origin. I am confused how (in the above example) stud # 1 would have a .001 positional tolerance and stud # 2 would have an .002 positional tolerance. In practice when actually measuring this, I presume one would have to make one or the other the starting point. If so, then would the sum of the bonus tolernace (.001 + .002 = .003) be totaled for a .003 postion tolerance on the 2nd? (with respect to the .390 basic). Or does one measure the 1st to a .001 cyl tol zone with respect to actual location of the 2nd, then the 2nd to a cyl tol zone of .002 to the actual location of the 1st? (seems redundant)

Then the next question that comes up for me that is if either or both have an acceptable out of perpendicular condition, where is the measurement to the other (i.e.: what point along the stud axis) determined from?

 

[MechNorth]

The origin of measurement is determined by a pair of mutually perpendicular datum planes. These planes are centered equidistant between the centers of the basic locations (wrt each other) of the two datum features, and inline between the centers of the two basic locations. The center of origin is based in your datum simulator, not the actual features.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[pmarc]

TrailMaker004,
I would say your dillemma is very common issue amongst CMM operators - you automatically look for fixed X-Y origin for measurements of studs position, which is reasonable practice in general, but does not always have to be a case. On the contrary, in this particular situation - when spacing within the pattern and its perpendicularity to A is checked - if you use one of the holes as a starting point for the other, you may end up with rejecting good part or, even worse, accepting bad product.

Like I mentioned earlier, the tolerance zone for the pattern of dia. 122 and dia. 121 studs is two cylinders, one of .001 diameter, second of .002 diameter, which centers are spaced basic .390 apart and perfectly perpendicular to datum plane A. These two cylinders can float together freely within tolerance zone defined by upper segment of composite positional FCF. This means there is no "global" origin that these cylinders are located from, as well as none of them is a starting point for the other.

Your next question also stems from very common misconception amongst CMM users. There is no single point (or level) along stud axis at which the measurement should be taken. The FCF says that whole axes of studs must be within cylindrical tolerance zones. Unless both axes are perfectly perpendicular to datum plane A, which is very unlikely, there is no chance to describe the axes as single points with x,y coordinates only. One must verify whether lines fall within cylinders of certain size, not points. This in consequence requires probing each stud at least at two levels in order to find two center points that could be used to derive a line approximately reflecting the axis.

If I were to measure the studs and verify this particular composite callout, I would try find the axes as-described above, and then see if they fell within tolerance zone as-described in second paragraph. Since two cylinders of the tolerance zone were free to rotate and translate, it would be possible and probably necessary to use a best-fit routine to adjust actual axes with their tolerance zones. Right after I found a configuration meeting the requirement, I would check whether it satisfied requirement expressed in upper segment of composite FCF as well. If it did not, I would iterate until satisfactory configuration was found or until I was 100% sure that such configuration did not exist.

 

[fsincox]

Technically. Calling out two separate requirements on a feature is not wrong in and of itself; it does not have to be a composite. We believe a composite will be a clearer statement of what, we think, you want. The callout of 2X holes to a general location and then another still on both of the holes including one of them as a datum is what really confuses the issue, here.
Done here as separate statements they can be in any order as far as I know!
The slot extreme radii are not defined as shown either

 

[greenimi]

I am wondering if, in this particular case, a gage could be designed to verify the composite positional requirements (the proposed one). Recently, Tec-Ease had a three part material about how to verify some composite positional requirements with functional gages: (Composite Gage Tip, Part I of III, Composite Gage Tip, Part II of III, Composite Gage Tip, Part II of III)

http://www.tec-ease.com/gdt-tips-view.php?q=269

And I assume the gage will take care of the complex vector datum shift allowed by design, right? Otherwise you have to know to program the CMM (if software is capable) to account for it (you don't want to reject potentially good/functional parts, don't you?)

 

[pmarc]

greenimi,
Sure you could build hard gage for verification of the composite FCF.

Till this point we have been talking about axis interpretation of positional tolerance at MMC, because this is how CMM operators understand such callouts in most cases. However this FCF could also be intepreted according to surface interpretation and the gage would nicely verify it. In fact this is, in my opinion, way better approach - you do not focus on imaginary tolerance zone cylinders and their size dependent on the actual size of the studs, but you check whether studs surfaces violate their virtual condition boundaries or not. The virtual condition boundaries are phisically simulated by the gage.

 

[TrailMaker004]

Since my last post I went down to work with my CMM porgrammer on this. We were able to do this with PC-DMIS in Exact Measure mode with no problem. The software does work from a calculated center point but I am still not clear on how that is determined (by the software) since the an origin derived from the basic dimension would be strictly theoretical and it would only have actual points to work from. We used circles this time around but will change to measured cylinders so as to ensure to full length of the both cylinder axis falls within the tolerance zone.

I absolutely agree, a hard gage can be easily designed/made and would actually simplify the inspection process (and the thinking behind it). We would however need 2 gages: one for the upper segment and one for the lower.