How to Define a Pattern Datum?

[MrComment]

Ok ever since my last disaster with PCDMIS (v4.1) while trying to use feature sets as a Datum reference feature (btw, never attempt that).

Ive been struggling to figure out how I can still apply any or all of my avaliable bonus tolerance when my Datum is a Pattern of features (i.e. holes) to the dimension features so that when everything gets dumped into Datapage I dont have to explain myself as to why I cant use Bonus on some of my TP features so on and so forth.

Does anyone know of the "correct" or even a quasi-"correct" way of defining a Datum when your Datum is a pattern?

Thanks

[JamesMannes]

Let see a screenshot of you print in question? If it is what I'm thinking it sounds like a "Best Fit" type of alignment that you want to do. More info as needed.

[MrComment]

See attachment.

Datum B (not shown) is the flange surface the holes are in. Kind of limited on the size of file I can post here.

[JamesMannes]

Ok, here goes nothing. What I read out of your print is this. Prior to calling the one hole -A-, it should be measured to another datum structure yes? So, the measured -A- should have the correct nominals input for its' measurement. Alignment translate to -A-. Measure the four holes, making sure that you have the correct nominals input (and correct workplane). Create a Best Fit alignment utilizing the 4 holes. Report the 4 holes incorporating the modifiers into your TP callouts. That should be it. Anyone have comment?

[MrComment]

Datum A is a pattern of 2 holes (note the far left and right holes). (Which are also those noted with the "#" symbol by the design dude.) We can further know that Datum A is a pattern because the tolerance callout above the FCF references 2 places only.

I believe the intent of this callout is due to their assembly process when the intake manifold is assembled to the engine. When only the outer two bolts (which are most likely studs sticking out of the cylinder head) are installed first as the engine rolls down the assembly line. These are then either hand tightened and possibly even torqued before the other 4 bolts are installed by an automated machine. Therefore the inner 4 bolts pattern must be within the specified location tolerance to Datum "A" so that all 4 bolts will start without cross threading in the cylinder head, while the other two holes (or pattern) has already been tightened down and have thus determined the rotation and translation of the entire part on the engine. (a process i have zero input into)

If it wasnt for the tightening sequence of the outer 2 first, the callout of course makes more sense to use as a Datum pattern of all 6 holes while the manifold is free rotate and tranlate in space. Which once all are installed would determine secondary and tertiary datums.

Either way these patterns can also be simulated by using a 2DBF to all 6 holes or just the 2 holes while leveled to Datum B (primary). However just getting the alignment correct alone doesnt help me report the total bonus avaliable to the other FCFs also have material condition modifiers on the Datums.

I could always report out the TP of all of my feature using the legacy style TP dimensioning and of course creating my alignments to my FCFs both manually and ahead of time, but the powers at be want to see the total bonus tolerance within the dimension feature as well. ATM, that is where I'm stuck..

[JamesMannes]

Sorry, I miised that second hole. Interesting, let ne think a little more and maybe by then some else will have piped in. That looks like a tough one. At least as far as using both holes with the modifier for your datum -A- .

[Jan d.]

OK, here's my try...

I always like to put things in (imaginary) hard gauge terms (go/no go). Makes it easier for me to understand. To make the hard gauge, I would construct a flat plate with 2 pins in there at the location of the 2 datum holes. The size of the pins are both 6.4 mm. So imagine a plate with 2 pins sticking out (at the nominal locations for both A holes). Have the gauge engage the part.

At the nominal location where you have to measure the hole, you stick in a pin (6.4mm size) through the plate at the nominal location of that hole. Wiggle and see whether you can push the pin in. If yes, you succeeded, part's OK. If not, the part is out of tolerance.

How to do this in PC-DMIS:
Measure the first hole and declare datum A1 (datdef). Measure the second hole and declare datum A2. Measure the plane and declare datum B. Measure the feature.

Now evaluate the following:
|TP|O 0.5 MMC|B|A1MMC|A2MMC|

If you do that in V4.1, it will give you the full amount of wiggle room (or bonus). I have done this many times before and I feel it works right.

It is my opinion that since both A holes are called out at MMC in the FCF, it does NOT matter which one of those 2 you call secondary or tertiary (try to understand the hard gauge explanation). Therefore you can do it this way, even though it seems weird to call out a pattern as a secondary datum. Analyze it both ways: with A1 as seondary/A2 as tertiary and A2 secondary/A1 tertiary. I think you'll get exactly the same results. Therefore it doesn't matter!

By the way: this does NOT work if you evaluate the feature with the datums at RFS: at RFS it matters greatly which one is secondary or tertiary! But since this is not the case in your problem, this is not an issue.



Jan.

[JamesMannes]

Jan,
Thanks for the input. I was not looking at that simplistically(which is all that was really needed.)

[Jan d.]

Just found it in the book by James Meadows (Geometric Dimensioning and Tolerancing), chapter 6, page 86:

"It is allowable for a pattern of features (such as holes) at MMC to be used as a group to establish a datum. Individual axes exist at the true position of each hole. These are the axes of the virtual condition cylinders of each hole. Once the primary datum plane is established by seating the part on the primary datum feature as it would in the assembly, the secondary and tertiary datum planes of the datum reference frame are established by the datum hole pattern."

This confirms what I tried to argue in my earlier post.


Just FYI.



Jan.

[MrComment]

Hi Jan,

I appreciate the response. I don’t have James Meadows book, but I do have the ASME Y14.5M-1994 instead.

I believe we are talking about the same sort of thing to.

However I respectfully have to disagree with your method for dimensioning either of the two holes as both the Secondary and Tertiary individually. I dont think this will yield the correct result in every case, and will ultimately put a bias on your tertiary hole as well. The attached picture stolen from the Y14.5 (mentioned above) shows that Datum B is constructed from the centroidal axis of all 4 holes in the pattern, and that the centroid of this pattern together defines the tertiary point.

Also what do you do when you have more then 2 holes in your pattern? I further stand by my argument as you cant define more then 3 Datums in your FCF as we know these are (Primary-Level, Secondary-Rotation, Tertiary-Origin).

As far as the hard gaging is concerned though I believe you’re correct in your analysis. A 2DBF (rotation and translation) simulates this sort of alignment in PCDMIS.

[rangerboat72]

Just found it in the book by James Meadows (Geometric Dimensioning and Tolerancing), chapter 6, page 86:

"It is allowable for a pattern of features (such as holes) at MMC to be used as a group to establish a datum. Individual axes exist at the true position of each hole. These are the axes of the virtual condition cylinders of each hole. Once the primary datum plane is established by seating the part on the primary datum feature as it would in the assembly, the secondary and tertiary datum planes of the datum reference frame are established by the datum hole pattern."

This confirms what I tried to argue in my earlier post.


Just FYI.



Jan.

I don't mean to hijack this thread but:

Jan, did you take any GD&T classes with Jim?

[Jan d.]

Jan, did you take any GD&T classes with Jim?

No, just studied the book on issues when I run into them.

[Rob Jensen]

Unfortunately 4.1 does not support Bonus on a secondary datum of multiple features of size (the two holes that define datum A). In this FCF datum B constrains 3 DOF (2 degree's of rotation and 1 degree of translation) and datum A constrains 3 DOF (2 degree's of tranlation and 1 degree of rotation). Since 6 DOF are constrained by the primary and secondary datums there cannot be a Tertiary datum in the FCF. I agree with MrComments interprtation of the callout.

The good news is 4.2 supports this callout. In 4.2 the procedure would be to measure the datum A features as circles or cylinders, construct a feature set from the the 2 holes and then define datum A as the feature set. Next create a FCF dimension of the feature set (PC-DMIS will actually dimension the size and position of each hole in the set). Finally create a FCF of the the holes called out to B and A (mmc) and PC-DMIS will calculate the virtual condition of datum A (calculate bonus based on the actual size of each hole as well as the allowable locating tolerance from the position callout of datum A).

Jan and I discussed this back in 4.0, and I informed him "virtual condition" of datums was not supported so I requested this enhancement which has been added to version 4.2. I tested it 3 weeks ago and it worked excellent. Hopefully 4.2 will be released soon (my guess is by April).

[JamesMannes]

Unfortunately 4.1 does not support Bonus on a secondary datum of multiple features of size (the two holes that define datum A). In this FCF datum B constrains 3 DOF (2 degree's of rotation and 1 degree of translation) and datum A constrains 3 DOF (2 degree's of tranlation and 1 degree of rotation). Since 6 DOF are constrained by the primary and secondary datums there cannot be a Tertiary datum in the FCF. I agree with MrComments interprtation of the callout.

The good news is 4.2 supports this callout. In 4.2 the procedure would be to measure the datum A features as circles or cylinders, construct a feature set from the the 2 holes and then define datum A as the feature set. Next create a FCF dimension of the feature set (PC-DMIS will actually dimension the size and position of each hole in the set). Finally create a FCF of the the holes called out to B and A (mmc) and PC-DMIS will calculate the virtual condition of datum A (calculate bonus based on the actual size of each hole as well as the allowable locating tolerance from the position callout of datum A).

Jan and I discussed this back in 4.0, and I informed him "virtual condition" of datums was not supported so I requested this enhancement which has been added to version 4.2. I tested it 3 weeks ago and it worked excellent. Hopefully 4.2 will be released soon (my guess is by April).

Rob, again you have cleared up a grey area, and I thank you for your timely response.

[cmmguy]

OK, here's my try...

How to do this in PC-DMIS:
Measure the first hole and declare datum A1 (datdef). Measure the second hole and declare datum A2. Measure the plane and declare datum B. Measure the feature.

Now evaluate the following:
|TP|O 0.5 MMC|B|A1MMC|A2MMC|

If you do that in V4.1, it will give you the full amount of wiggle room (or bonus). I have done this many times before and I feel it works right.


Jan.

Let me ask this about that.
If this tolerance scheme is applied to more than one hole, how do you do it. I think that you can't evaluate each hole independently to the datum features without considering the "fit" of the other features being evaluated.

In other words, the datums once oriented to pass the first feature (visualize your hard gage) must be fixed to evaluate the remaining features. Otherwise in the example of your hard gage, how would all of the feature pins fit in at the same time while the part is engages with the datum pins.

OR, am I completely missing something here...