Datum setup 3

[SeasonLee]

It seems there are a lot of problems on the attached print, but I will focus on the datum callout. Here is my understanding on the designer’s intent:
The bottom surface is primary datum A
Datum B is a pilot location hole
Datum C is another pilot location hole which is established base on datum B
B-C create the axis orientation
Datum B will be the original point for CMM measuring.

My questions are:
1. Can we allow two datum symbols (Datum A and Datum D) on a coplanar surface?
2. Is it a right way on Datum C callout?
3. Which one is correct on the datum reference frame callout
Primary datum A -> B-C -> Datum B
Primary datum A -> Datum B -> B-C
Primary datum A -> B-C (without tertiary datum)

Thanks for all comments

SeasonLee

 

[weavedreamer]

No offence taken. Juat going to have to review the references cited, and try to ensure that personal preferences are not coloring the interpretations.

The fact that you are reading this thread makes it worthy to be posting within.

 

[NormCrawford]

Kenat - Thanks for clarifying your std signature. And yes, I will review details of rules etc. related to this site. So far, I find it an outstanding forum.


Et al -
I did not realize yesterday that there was a second drawing. So, please keep in mind my previous comments are about the first drawing.

As for the second drawing, I also see it as a stretch. In this case, if functional datum features can't be determined for repeatable inspection, it makes sense to use other means of immobilizing the part. Keeping with how design and inspection may have made decisions on the second drawing, I would probably define datums B & C as a single Datum B with datum target points. Depending on that actual contour of the part, that may in itself stop all 6 degrees of freedom; otherwise known as a saddle datum. But since the flat face opposite datum feature A "appears" to be a mating face to a mating part, I would make that face Datum Feature A to functionally stop the appropriate 3 degrees of freedom and then Datum points for Datum Feature B (the four points of B & C) to stop the other 3.

Norm

 

[JLang17]

In reference to the drawing in the original post...

In Datum C's FCF, why is B at MMC? If datum B is an axis, why does the material condition matter? I see this quite often and could never find an explanation.

 

[ewh]

When B moves from MMC to LMC, the variance can be added to the positional tolerance. I'm sure someone else will give you a better explanation.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[SeasonLee]

Norm

What you said is absolutely true, I don't think datums B-C on 2nd example is repeatable, it makes me spend a lot time to find out the Datum B and C on CMM measuring, its not an easy job to do it precisely.

SeasonLee

 

[NormCrawford]

With regard to Datum B being referenced at MMC, the REASON should be that the fit of the feature of size is a clearance fit. In other words, using MMC or RFS is not really discretionary. It is based on the physical design. Therefore, the part can and will physically be able to move/wiggle/float. And therefore that mobility, otherwise known as datum shift, allows the part to move within the datum feature simulator for datum feature B as datum feature B departs from it's MMC condition so that datum feature C could then fall within it's stated tolerance. (It will be able to assemble with said movement.)
However, beware! You should not get in the habit of incorrectly adding datum shift to the allowable tolerance of a considered feature. Mathematically it makes sense and works for a single feature, but you wouldn't REPORT in or out of a position tolerance value other then what is actually stated in considered feature's FCF. AND, when you have a pattern of features, which not always obvious, mathematically adding datum shift availability "separately" to a stated tolerance of a single feature can lead to accepting bad parts.

Norm Crawford
GDTP-S
Applied Geometrics, Inc.

http://www.GDandT.com

 

[JLang17]

Thank you, that seems to make sense.

 

[PaulJackson]

Referencing drawing #1 and guessing at functional mating features... I would--- identify the datum feature "A" beneath the callout [profile|0.20] "2 surfaces" note. Then I would use only one of the "dia .125 +/- .002"... making it "2X dia .125 +/- .002" and including the position callout beneath [pos|0(M)|A]... then beneath that identify the datum feature pattern as "B".

All other callouts would declare |A|B(M)| as the datum reference.

Furthermore... to reinforce the fact that both alignment holes have equal functional prescedent in stopping translation and rotation in relation to their mating studs ... I would place a vertical dimension line at the midpoint of the two hole pattern "B"... and show basic displacements to all the features "in that view" from the center of the pattern "B"... including one to each feature "B".

With the CMM... "A" stops one translational and two rotational "degrees-of-freedom" about the surfaces and "B" stops one rotational and two translational along the line intersecting the axes of the pattern holes and one at the midpoint of the hole pattern.

An attribute gage... commonly referred to as a "functional gage" would do the same!

In figuring the stack allowances for the pattern @ MMC... I would moderate the liability of tolerance accumulation along the axis intersecting the two hole pattern "B" to +/- .002 {predicting that the pattern will be in interference (compression or tension) with the mating studs most of the time... and I would disregard any rotational stack accumulation for the same reason.

Just a thought... That's how I would do it.
Paul

Great Responses... I would have responded sooner but I was vacationing for a while.

 

[SeasonLee]

Paul

Thanks for your input, I like your different thought on the datum establishing, may I ask:

1. You mentioned will place a vertical dimension line at the midpoint of the two hole pattern "B", is it based on the part is designed symmetrically ? Actually the part is not symmetric.
2. What is the difference between your proposed FCF IAIB(M)I and Fig 4-8 in the standard Y14.5M IAIB(M)IC(M)I ?

Fig 4-8 attached for your quick reference since you are on vacation.

Have a great time on your vacation

SeasonLee

 

[NormCrawford]

Paul is providing yet the cleanest way to re-dimension this part with a simplified way of saying what we have been saying. I tried to stay on the co-datum method for explanation purposes, but if I had to start clean, I would go with IAIBmmcI as explained, with only one minor exception.
On the part detail drawing I would not inlcude the vertical centerline ideally located midway between the two holes. I find that perfectly find for the inspection process (plan), but is not functional to the part (design intent). However, I also know too, that industry is putting more process dimensioning on prints and if it ultimately clarifies team understanding, then I can live it too.



Norm Crawford
GDTP-S
Applied Geometrics, Inc.

http://www.GDandT.com

 

[PaulJackson]

Norm,
Functionally... if both of the .125 +/- .002 holes mate with rigid round features from the mating part with identically designed clearances... then neither hole by itself will stop translation along the axis oriented to the line intersecting their centers... they do it as a pattern.

If the spread between the holes is longer than the basic then the pattern will experience minimum clearance inboard and if shorter... outboard. Either way the tolerance for spread is shared by combining each hole's deviation from its AME size.

If the secondary datum feature was declared as a two hole pattern: (as I suspect that it functionally operates in the assembly)

Then the DRF by design would reside oriented to "A" and the pattern "B" at the midpoint between the holes. That is why I would detail the basics from that origin. It has nothing to do with "process dimensioning"

Then the inspection process (plan) would be incorrect if the measurement DRF was coincident with the axis of one of the pattern's holes rather than the center of the pattern.

SeasonLee,
ASME Y14.5 Figure 4.8 is intended to illustrate how rotational orientations can be accomplished using holes.
It doesn't suggest a method to functionally detail products with two holes. The function does that.

It is interesting to me however, that paragraph 4.4.3(b) states "Orientation is established by the width of hole C".
It would seem to me that the width hole C (oriented to the line between B and C) would need to be identified as datum feature C for that to be true... otherwise the axis of the AME or the Virtual Condition Size of C depending on RFS or MMC would establish that rotation. I think that statement is a mistake.

Paul

 

[NormCrawford]

Paul,
I understand and agree that the two holes as a pattern restrict 2 degress of translation and 1 rotation when referenced as a secondary. I also agree that if the holes are located in two opposite directions within their stated tolerance zone the actual available datum shift is reduced and could be zero depending on the AME.

However, where I think we will have a continued disagreement is locating the pattern to a centerline being the origin of the pattern. What is the tolerance then? The posiiton tolerance you referenced earlier would be the "feature to feature" location tolerance with an orientation control to datum feature A. I know you know that. Using IAIBmmc, the origin of the pattern can be anywhere as long as the simultaneous requirement rule is properly applied.
How would we locate the "origin" of an irregular pattern? The point being that if the centerline uses BASIC dimensions, then why BASIC dimensions be applied to an origin anywhere? What would be the diffence?

"Process dimensioning" was probably not my best choice of words as many will refer to that as manufacturing process. All I was saying is that a measurement plan can and often does dictate a consistant origin to use for repeatability etc. and I certainly agree with that.
If there is a reference or figure in Y14.5 that dimensions a pattern to its own origin, I really want to know about. No kidding. I would really want to check that out.

Norm Crawford
GDTP-S
Applied Geometrics, Inc.

http://www.GDandT.com

 

[ringster]

Back in about 1973 one
of the advantages to using GD and T was to provide '"Universal Interpretation' of a drawing. My how times have changed!

 

[KENAT]

ringster, when we start discussing some of the more complex things that aren't as clear in the standard I wonder the same thing.

Even if it is based on extension of principles in 14.5 if most readers of the drawing, even those with a reasonable working knowledge of 14.5, can't readily understand it how useful is it really?

Is it better to use an approximation/simplification whatever the term may be, that is actually in the standard so long as it stands a reasonable probability of achieving what you functionally need?

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[SeasonLee]

Now, part #1 re-dimensioned :

1. Datum D symbol removed and phantom line added between the two surfaces.

2. The two holes changed as a hole pattern and identify as datum B, all geometric tolerance's DRF changed to |A|B(m)|.

I am not a designer, please let me know if there are any problems on the print.

SeasonLee

 

[KENAT]

1. SeasonLee, if you care about such things take a look at thread1103-237315 where it was determined that for indicating a multi surface datum an extension line was more correct than a phantom line. In reality it probably wont make much difference but FYI.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[ewh]

I feel the same as KENAT on this issue. As it is with so many engineering tools today, there is usually more than one way to accomplish something. We have to keep the KISS principal in mind though, and not loose the forest because of the trees.
As for using a phantom line in this manner, I disagree. An extension line should be used.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[axym]

Unfortunately I only have time for a quick post.

I'll comment on the "continued disagreement" between Norm and Paul about locating the DRF at the "origin" or "center" of a pattern. I seldom disagree with anything Paul says, but in this case I agree with Norm. The basic dimensions don't need to be laid out with any particular origin or centerline. It makes no difference. Any DOF constraints implied from the basic dimensions are just that - implied.

I'll comment Paul's statement:

"Furthermore... to reinforce the fact that both alignment holes have equal functional prescedent in stopping translation and rotation in relation to their mating studs ... I would place a vertical dimension line at the midpoint of the two hole pattern "B"... and show basic displacements to all the features "in that view" from the center of the pattern "B"... including one to each feature "B"."

I agree that both holes have equal functional precedent in stopping translation and rotation in relation to their mating studs. It's a fact. But showing basic dimensions from a centerline to each hole doesn't reinforce that, and shouldn't be taken as that. I suppose that if the drawing user is going to (incorrectly) constrain the part's degrees of freedom based on how the basics are laid out, then laying them out from a centerline will let them get a result that is closer to correct. But if the basics were laid out from one hole, the same drawing user would probably use that hole as a secondary datum feature and the other as a tertiary, which would be wrong. The two holes would still have equal functional precedent regardless of how the basics are laid out and where the origin is - I believe that iss Norm's point and I agree with it.

This is another example of a larger issue, related to another open thread with the compound-angled cylinder datum feature.

I hope to comment more on these tomorrow.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[PaulJackson]

SeasonLee,
Your new sketch is better than the original in that it declares the missing (implied) primary datum features and it controls the original hole "B" for its orientation to "A" (in addition to pattern spread via the 2X pattern position callout). The original drawing was missing that perpendicularity callout of "B" to "A". Even though that control becomes immeasurable as the depth of "B" approaches zero, these building block dimensioning prerequisites i.e. form of the primary, orientation of the secondary, position of the tertiary... are important for design constraint and variation analysis. Without that control a 3D stack analysis would nearly tip the shallow cylinder of "B" over using its tolerance for size.

The only things that I would change... I have already mentioned... Hang the "A" datum identifier from the leader of the 2 surfaces profile control (personal preference... I think that the phantom line option is less emphatic) and reference the basic dimensions from the center of the 2X pattern "B".

Norm,
I agree that with non-symmetrical patterns placing the origin for declaring basic dimensions is a problem without a desirable solution. Technically the origin for describing the basics can be anywhere on the detail... even at a point totally divorced from the datum feature structure… as long as it equivalently relates the basic displacements among all the features being controlled. The problem with specifying basic dimensions from origins and rotations other than those defined by the feature control frame is that when inspections (other than attribute) are performed the inspector must re-figure the basics from the DRF specified in the feature control, perform the inspection, and report the deviation. Those reports will show deviations from each of the re-figured basic displacements rather than the ones specified on the drawing. In order to make the two relate the inspector would have to refigure the measured deviations from the alternate origin solely for the benefit of the person that reads the inspection report… it is a pain in the xxx.

For instance… consider a thru hole in a cast structure that is not square with the DRF that it is controlled from… the designer “for expedience sake or lack of understanding of how inspection is performed” will often “snap” X, Y, & Z coordinates square to the DRF at the intersection of the structure’s surface and the hole… rather than specifying the coordinate axial displacements and associated angular displacements normal to the hole.
The inspector must figure the latter to perform the inspection of the hole’s position to the DRF and then report his measurements and figure the deviation. His measurements will not relate to those specified on the drawing unless he re-figures the X, Y, & Z coordinates of the hole’s axis deviations at the point where the specified surface should be!

SeasonLee’s design is similar. If the secondary is specified as a 2X pattern the inspector must find the center of the pattern, perform the inspections from that reference and then relate the measured data to the origin specified on the drawing. If the two were the same “the center of the pattern” there would be no manipulation required. If the DRF is coincident with one of the “B” pattern holes then the data must be shifted from the center of the pattern… ½ the basic spread so that it matches what is specified on the drawing. Both holes “B” in this case will have identical measured deviations reflecting whether the spread is longer or shorter than the specified basic.

I hope you understand that I am not saying that it is illegal or wrong to specify the origins and orientations for basics with respect to the displacements that must be figured to perform the inspection… it is just INMHO a poor design practice.

I said that “I suspect” from the details of the component the 2X .125 +/- .002 holes probably work as a pattern for location and alignment with the mating part… that my be a false assumption… I don’t know the details of the mating part… but if it is then I would reference the basics from the center of the pattern just as an inspector would have to do to measure the 2X .125 +/- .002 [pos|0 (M)|A] as 7.063/2 displacements to each hole from that reference.

Paul

 

[ringster]

I dont believe that I have seen this application before. It would seem that if we have 2 seperate datum features, we should apply 2 different datum identifiers. That being the case we could refer to the distance between them as width B- X, IF NECESSARY. I thought the idea of centerlines as a datum was PELOSI. I canat recall seeing an example of it in the Standard, although some trainers, include it.

Hard to figure this one.