Hole Pattern Datum - Limited Constraint

[somervil09]

Hello All,

This may be a simple question, but I haven't seen a clear illustration that directly addresses my situation. I need to establish a secondary datum using a pair of holes, the datum being the midpoint between the two centers (point P). See the attached sketch for illustration. The rotation of the part should be defined by the dia.12 hole at the center of the part. Essentially I want to define point P as the X,Y origin and fix rotation using the third hole. Calling out the 2-hole pattern as datum B would fully constrain the part without regard to the third hole (correct??), so what would be the best way to establish the DRF as I would like it to be?

 

[chez311]

somervil09,

In regards to explicitly setting your "origin" - this is not typically necessary and can be assumed to be at the midpoint/center of a pattern. If it is desired to specifically call this out it can be done either implicitly per Figure 4-26 or explicitly per Figure 4-28 with basic dimensions.

The rotation of the part should be defined by the dia.12 hole at the center of the part. Essentially I want to define point P as the X,Y origin and fix rotation using the third hole.

Why do you want to set the origin in this manner(see my above)? Are the 2x holes actually important mating features or are you just using them to set an arbitrary origin point? How does this part mate/function and what kind of tolerances are involved in the mating features of the part shown and the features/components that mate with it? Will the features be called out at MMC/MMB or RFS/RMB or some other combination?

By setting your secondary datum as a pattern you have constrained all remaining DOF. You can utilize other methods such as a customized DRF to "establish the DRF as you would like it to be" (constrain specific DOF) however there is a high probability that this will not mimic actual part function/mating conditions and will only exist in theory - I would avoid this if at all possible, utilizing it with extreme caution. The more likely scenario is either (1) your DRF is fine as it is, leaving the 2x hole pattern as secondary and no tertiary datum or (2) a slight adjustment in your datum selection utilizing a standard DRF - depending on part function and tolerances.

 

[jassco]

Hi, somervil09:

I would do as follows:

1. Datum A as primary (no change);
2. Datum B (12mm hole) as secondary;
3. Add a HORIZONTAL dimension between two 5mm holes;
4. Attached datum C to this dimension. Datum C will give you a DMP (derived medium plane) to center point P.

Best regards,

Alex

 

[chez311]

3. Add a HORIZONTAL dimension between two 5mm holes;
4. Attached datum C to this dimension. Datum C will give you a DMP (derived medium plane) to center point P.

Not sure how you imagine this working. For starters you cannot specify a DMP (derived MEDIAN plane) in that manner (dimension between hole axes). Additionally a DMP (imperfect) would in no way relate to a datum plane/center plane (theoretically perfect). You also cannot specify a CENTER plane datum in this manner either as it must be derived from a physical feature/pattern which the dimension between the hole axes is NOT.

You could set the 1x 12 dia hole as secondary and the 2x 5 dia PATTERN of holes as tertiary, however as I said more information is needed as to part function and tolerances.

 

[greenimi]

Hi, somervil09:

I would do as follows:

1. Datum A as primary (no change);
2. Datum B (12mm hole) as secondary;
3. Add a HORIZONTAL dimension between two 5mm holes;
4. Attached datum C to this dimension. Datum C will give you a DMP (derived medium plane) to center point P.

Best regards,

Alex

Do I even want to comment the above "advice"?
I just want to say…. well no comment.
So many, wrong recommendations, illegal stuff (truly, that is the derived median plane definition??….) and poorly worded guidelines.
"Attached datum C to a dimension"??? Seriously?
Stay away from it.

Sorry Alex. You are just flatout wrong.

 

[somervil09]

The 2X holes are important mating features and their location defines accuracy of the greater assembly - this is a highly simplified part that I drew to illustrate the discussion. The location of the center hole is of lower importance, thus the preference for locating from the 2x hole pattern.

The part is a plastic part that has some considerable shrink relative to the tolerances allowed. Skewing to the center hole is a way to align the part with the centerline of symmetry. By aligning the part using the 2x hole pattern, a small error in their positional location created a large offset (due to angular skew) when other features were measured at greater distance from the origin. We've had problems with an existing DRF based solely on the 2x hole pattern, and are looking for ways to resolve the issue.

 

[greenimi]

chez311,
I guess we've posted on exaclty the same time....
The same "message" for Alex.
Hopefully, he/she will get it.

 

[chez311]

The 2X holes are important mating features and their location defines accuracy of the greater assembly - this is a highly simplified part that I drew to illustrate the discussion. The location of the center hole is of lower importance, thus the preference for locating from the 2x hole pattern.

How exactly to these 2x holes mate? With standard fasteners, dowel pins, or otherwise? Also if the center hole is of "lower importance" - does it physically mate with the corresponding part at all?

The part is a plastic part that has some considerable shrink relative to the tolerances allowed. Skewing to the center hole is a way to align the part with the centerline of symmetry. By aligning the part using the 2x hole pattern, a small error in their positional location created a large offset (due to angular skew) when other features were measured at greater distance from the origin.

Could you explain your statement about "considerable shrink relative to the tolerances allowed"? I'm not quite clear on the context of what you mean. Also is this an issue where the print tolerances allow conforming parts with excessive misalignment or are you having trouble meeting the print tolerances in production? Or are you saying that due to the combination of material, tolerances, and datum selection you are not able to measure the part repeatably?

An option could be to hold everything to simultaneous requirements to datum A (hold everything to a simultaneous tolerance zone instead of aligning first to the 2x 5 dia holes) however I'm not sure if that or other solutions playing with your DRF will solve the issue - though it might if somehow fixturing from the current datum features is impacting your ability to repeatably measure the part. It sounds like you have an issue of a less than robust set of datum features which are close together and on the extreme end of the part (both compound to create your excessive angular misalignment) in addition to loose (or looser than desired) manufacturing tolerances. Off the top of my head, two solutions I see are either (1) change the design of the assembly to allow more robust datum features with the current manufacturing tolerances or (2) perform a stack and tighten the hole size/position tolerances to limit misalignment to within acceptable bounds.

greenimi,
Agreed - though I was trying to sugarcoat it a bit....

 

[somervil09]

The problem is two-fold: we are having trouble meeting the print tolerances in production, and the existing DRF (based on the 2x hole pattern) is not fully representative of the part function. As a result, our parts are functional in the assembly, but are failing inspection. I can measure these parts with a DRF as I originally described and see that the parts are within tolerance and should have no issues. The tolerances on individual features are appropriate, but the skew introduced from our current DRF causes some features to consistently fall outside the specified limits. Really the question is more about how best to call out the datums to achieve the desired reference frame. The simultaneous requirement sounds interesting, but not sure it would have the same effect. Could you reference the 3-hole pattern as a single datum for subsequent features, or would all of your positional tolerances have to be made into a simultaneous requirement?

 

[jassco]

Hi, chez311 and greenimi:

I got it. Thanks! I was thinking about Fig. 4-35 (c) on ASME Y14.5-2009 when I posted my comment.

Imagine we create a vertical rectangle zone that expands until they kiss those two small holes. Can't we use width of this zone (feature of size) as datum C?

If we call the small holes (2x) as datum C, then median plane won't pass through center axis of datum B, or will it?

Best regards,

Alex

 

[jassco]

Please see sketch attached.

Does it make any sense? Please see Fig. 4-15 (ASME Y14.5 - 2009)

Best regards,

Alex

 

[jassco]

Hi, greenimi and chez311:

If we use the flat face as datum feature A, 12mm hole as datum feature B, and a pattern of two small holes (5mm) as datum feature C, where are three datum planes? And where is origin?

Best regards,

Alex

 

[chez311]

the existing DRF (based on the 2x hole pattern) is not fully representative of the part function. As a result, our parts are functional in the assembly, but are failing inspection.
[...]
The tolerances on individual features are appropriate, but the skew introduced from our current DRF causes some features to consistently fall outside the specified limits.

Could you more clearly define how these features relate to the part function and assembly as well as some of the size/location tolerances involved? You have stated the 2x hole pattern is the critical assembly locating feature - without further information I would say the current DRF is representative. Also is there an opportunity to open up tolerances if parts that fail inspection are "functional in the assembly"? Or possibly you are referencing these features at RFS/RMB when they could be at MMC/MMB to allow some of these functional parts to pass inspection?

Sounds also like either the tolerances either do not take into account the additional rotation allowed as the features deviate (ie: tolerances are fine to locate the 2x holes to each other but not to locate the features they control with the addition of compounded rotation of features far away from the datum features) and/or are simply too sloppy for a datum feature. Your datum features should be sufficiently more accurate than the features they control - otherwise you run into the issues you seem to be having. Imagine trying to accurately locate a precision reamed hole to a hand drilled hole - maybe not impossible but very difficult. You could try and tighten up the tolerances on your current datum features. Someone could chime in but I think simultaneous requirements would alleviate some of this as it would prevent you from having to locate/orient to inaccurate datum features. Lastly you could choose a more accurate and robust feature on the part (say the center 1x dia 12 hole as secondary and utilize the 2x 5 dia holes for tertiary orientation) as an alternate datum - this is not always recommended but can be a viable option.

Could you reference the 3-hole pattern as a single datum for subsequent features

You can reference it as a composite datum ie: set the center 1x dia 12 hole as datum B, the 2x dia 5 hole at datum C and reference it as [B-C], though I'm not convinced that would solve your issue.

Still also not clear on what you meant by "considerable shrink relative to the tolerances allowed"..

 

[pylfrm]

This may be a simple question, but I haven't seen a clear illustration that directly addresses my situation. I need to establish a secondary datum using a pair of holes, the datum being the midpoint between the two centers (point P).

A similar situation is shown in ASME Y14.5-2009 Fig. 4-46. Unfortunately I think the customized datum reference frame concept is not as well-defined as should be, and I'd be hesitant to use it in this situation.

More details about the interaction with the mating parts would probably lead to better suggestions.


pylfrm

 

[chez311]

jassco,

If we use the flat face as datum feature A, 12mm hole as datum feature B, and a pattern of two small holes (5mm) as datum feature C, where are three datum planes? And where is origin?

Your first plane is on the face coincident with datum feature A, the 12mm hole creates an axis which is at the intersection of the second and third plane. The second and third planes are then oriented by the 2x 5mm pattern of holes. Take a look at Figure 4-8 which shows you that an axis as a secondary datum forms 2x planes which intersect at its axis and 4-6 to show how those 2x planes are oriented by a tertiary datum. It is admittedly somewhat ambiguous as to what exactly this orientation is when you have a pattern of features, but I would say it doesn't really matter in most cases - and if it does it can easily be specified to show the origin and axes orientation desired per Figure 4-28.

As far as your reference to 4-35 showing the use of IRREGULAR FOS as a datum feature, it may be possible but I'm not quite sure. I would not utilize it as a secondary as the two opposing points of contact would probably make for an unstable datum feature but as a tertiary may be okay - one of the more experienced members of this forum would have to weigh in on that. Using datum targets may be a more palatable solution.

 

[chez311]

pylfrm,

Agreed on both points. I initially alluded to a CDRF in my first response but I didn't think thats really a rabbit hole that OP wants to venture down. What do you think about simultaneous requirements? I'm not 100% clear based on the responses but I think part of the issue is probably measuring from and fixturing to sloppy datum features (or while perhaps not "sloppy" just not quite accurate enough) and it seems to me holding the part to simultaneous requirements would alleviate some of that, though I might well be wrong.

Ditto on more details - all I know is that its a plastic part in which "The 2X holes are important mating features and their location defines accuracy of the greater assembly". How exactly this assembly fits together (dowels, fasteners, etc...) and the tolerances involved in the datum features as well as the rest of the part would be very helpful. Hopefully OP can provide.

 

[jassco]

Hi,

Please take a look at attached ugly sketch. I think datum C on this sketch makes a better datum feature for clocking the part. What do you think?

Best regards,

Alex

 

[chez311]

jassco,

I would disagree. As I said I am unsure as to whether your initial proposal was feasible (width dimension) considering the boundary exists inside the part - not to mention what your datum feature simulators would look like, but your second proposal I would pretty confidently say is a no-go. Even if someone said it was somehow legal per the standard I would say it would make a very poor and unreliable datum. Imagine trying to grip exactly opposing points on a cylinder (across its diameter passing through its axis) with two smaller cylinders - it would be almost impossible and you would just about always grab on either side of the axis.

I don't personally think this method of creating an irregular FOS is really a good solution.

 

[greenimi]

Jassco,
So, are you going to control the 5mm holes (two holes) in relationship to A primary, B secondary and C tertiary? If not why tertiary is needed?


All,
Why both holes (Ø5mm) cannot be used as a pattern to establish datum feature B (for example) and the big hole Ø12mm be positiined to A primary and B secondary?
A primary controlled with flatness
B secondary position to A primary
Ø12mm hole position to A primary and B secondary

The outer shape/surface to be profilled all around to A primary and B secondary.

What could be inadeauate inadequate with my above approach? Why CDRF is needed in this case?

I do understand that chez311 wants to expolre and understand CDRF (in general), but not so sure in this case is really needed.

 

[chez311]

greenimi,

Why both holes (Ø5mm) cannot be used as a pattern to establish datum feature B (for example) and the big hole Ø12mm be positiined to A primary and B secondary?

This is what I initially suggested and apparently OP is saying is the current approach, but not sufficient for some reason. OP even claimed it was not representative of the assembly but without any further information than "the 2X holes are important mating features and their location defines accuracy of the greater assembly" I would disagree and say it IS representative. It may have gotten a little buried but its mostly contained in my response from 25 Sep 18 17:48 - I think OP's issues stem from a combination of not having a robust datum feature (2x holes too close together as well as far from the features they control), tolerances on the datum features which are too loose, and specifying at RFS/RMB when MMC/MMB could allow functional parts to pass.

Truthfully unless there is more information about the assembly methods and tolerances involved its difficult to say what the best solution or issue really is.

What could be inadeauate inadequate with my above approach? Why CDRF is needed in this case?

I do understand that chez311 wants to expolre and understand CDRF (in general), but not so sure in this case is really needed.

I don't think anyone said its needed, my proposed solutions did NOT include CDRF - you have noted my interest in it but it is mostly from an academic standpoint, but I am still skeptical about its actual application. I only really mentioned it because it would do what OP wanted and allow constraining of specific DOF but I was careful to say in my first post that it likely will not mimic actual part function/mating conditions and will only exist in theory - and to avoid it if at all possible.

My solutions included:
1) Tighten up the size and location tolerances on the 2x 5mm hole datum features
2) Specify at MMC/MMB if not already and/or loosen tolerances on the other features (NOT THE DATUM FEATURES) on the part (OP said functional parts are being failed at inspection)
3) Utilize simultaneous requirements - I am unsure if this would actually solve anything but it seems viable if the issue stems from measuring from/fixturing to inaccurate datum features
4) Utilize an alternate set of datum features if other options do not work or cannot be used