ASME Y14.5-2009 4.3g usage stirs the pot 5

[ptruitt]

I expect to back off an approach to tolerancing that I am taking which is understandably unpopular with a valued supplier. But here's what keeps going thru my head: When it comes to rather small, uniform injection molded or die-cast parts, tolerances are more based on inch/inch or mm/mm than individually toleranced features of size. So all of the features in each mold cavity will be well controlled with respect to each other. Then there will be a 'pattern shift', so to speak, with those features in the other mold cavities (or cores). I came up with a sample part to consider. (Please ignore wall thickness issues and other good practices, which I have ignored, for the most part.) Does anyone see improper usage of the 2009 Standard? What do people like/dislike about this approach?

 

[drawoh]

ptruitt,

I can see what you are doing, and it makes sense to me. What is your vendor's objection?

JHG

 

[ptruitt]

They have not seen this approach before and they are very busy. Good-news/bad-news! Perhaps, if they had the time to go over the 2009 Standard, they would be more likely to embrace it. I assume that they will think that they will need to buy the latest CMM, software, and seminars/training. I am thinking that there is really no more reason for CMMs and software, (maybe less) but no way to get around the seminars/training. I am flexible and happy to do what it takes to get to production. Basically, I just want to be sure that my understanding of the 2009 Standard is harmonious with the rest of the gang.

 

[Belanger]

The difficulty I can see is how to simulate the datum. Remember, we can call anything we want a datum feature, but somehow it still needs to be simulated -- since you say that they are using a CMM, the computer has to derive the datum from the contact points on the physical part. Are you expecting them to scan the entire surface that is meant to be "A"?

I think it's one of those things where it is legal per the standard, but from a real-world manufacturing viewpoint I too might be left scratching my head.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[pmarc]

Have you looked to ASME Y14.8 "Casting, Forgings and Molded Parts" standard? There are some dimensioning practices described in it which might be helpful.

 

[ptruitt]

Hmmmmm.... I admit I don't have a good answer, Belanger. In reality, the products I design are not too expensive and not directly safety-related. Regardless of how I dimension them, the first articles will get an extensive inspection and then production may get functional gages. My aim/hope(?) is to have very process-capable designs, qualified suppliers, and a program that keeps reducing inspection for the supplier and the OEM, if possible. When you compare the use of profile tolerancing to plus/minus dimensioning, without geometric controls, I would stand my ground. The supplier we use never complains about drawings that have no GD&T on them. I won't be arguing that point with them, but it is interesting to think about. One thing I don't want to do is add datum target features that have no other purpose. I can see why that has been done in the past, but I think it is probably moving technology in the wrong direction at this point. It seems to me that there should be a way to exploit profile tolerancing when working with a supplier that does not have software such as Smart Profile and PC-DMIS for products that do not require the scrutiny that medical, aviation, and other safety-related products need in order to be efficacious. It gets back to reasonable methods. I plan to take a metrology course soon and I will keep your comments in mind, Belanger. Thanks.

Thanks for reminding me about ASME Y14.8, pmarc. Bob has a copy. I'll go get it.

 

[drawoh]

ptruitt,

I figure that using your complex feature as a datum is perfectly legal, and feasible if you specify that it applies at MMB. This makes your datum_A profile tolerances sloppier. Is that acceptable?

The whole point of datum targets is that everyone who handles your part has the same repeatable datums, even though the part is a fairly inaccurate casting. Datum targets are marginally useful for accurate, rigid parts. They are a good idea for castings.

JHG

 

[ptruitt]

Does not usng MMB make it illegal, drawoh? It does seem that, in most cases, I can, indeed, use it. I'll ponder that while I listen to "Grandma got ran over by a reindeer" for the 1,000th time today.

It is very reasonable to think about this topic in terms of how to inspect something. But also consider the opportunity for some folks to take robust design and process qualification to the point of zero or near-zero inspection... That's kind of funny, now that I think of it. I know of a place (I'm sure you do, too) where they have zero or near-zero inspection down to a science, but forgot the robust design and process qualification parts. Oops.

 

[axym]

ptruitt,

I don't see anything illegal on your drawing, but there are some issues that I would point out.

Datum feature A is a "mathematically defined surface", described in Section 4.13 and Figure 4-28 of Y14.5-2009. So your drawing is legal per the standard, I think, but you're applying one of the brand new tools and you may suffer the pain of being an early adopter. JP is right that it will be very difficult to simulate (I hate that term) datum feature A. Not impossible, but there are some significant practical difficulties. Datum feature A needs to be referenced at some particular "boundary condition". The example in the standard uses the BSC modifier for the mathematically defined datum feature - that is, the datum feature simulator has the basic geometry of the datum feature. Other options are MMB (Maximum Material Boundary and LMB (Least Material Boundary). The actual geometry (or the cloud of points scanned from it) then needs be fitted to the simulator, so that there is contact on the high points. This is not straightforward, even for CMM's with high-density scanning capability. Different softwares will fit the point cloud in different ways, with different datum reference frames as a result. The amount of difference will depend on the form error of datum feature A.

Another issue is simultaneous requirements. There are three profile FCF's that all reference datum feature A only, so all of the features controlled by these FCF's are treated as a composite pattern. It doesn't sound like this was the intent.

There's also another standard that you may want to look into, if you haven't already. It's Y14.8-2009 (Castings, Forgings, and Molded Parts). This standard deals with mold-related issues such as parting lines etc. Here's a link to an article written by Don Day, the chairman of the Y14.8 subcommittee:

http://memagazine.asme.org/Articles/2010/march/Expanding_Standard.cfm

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[Belanger]

Evan -- I also hate the term "simulate." When I wrote my original response I had typed "stimulate" but I'm glad I proofread it! I sure wouldn't want to recommend any of you to stimulate your datums... that sure ain't in the standard

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[drawoh]

ptruitt,

Your feature RFS is legal as per ASME Y14.5, but not totally feasible.

If your datum is a feature of size at MMB, you can make a fixture that simulates that MMB condition. Parts that are not close to MMB can be wiggled around on your fixture, thereby simulating the bonus allowance.

I am comfortable using features of size without MMB, as long as the feature is very much more accurate than the tolerance being tested. Holes for dowel pins are a good example of this. Try sketching out the expanding fixture needed to capture your feature RFS. I think it will be a challenge.

JHG

 

[ptruitt]

Thanks for the comments, everyone. I am having trouble understanding how MMB can be used with non-planar features, though. I looked over the ASME YY14.8-2009 and found much to like. After reading your comments and the Standard, I modified my drawing. I am hoping that it is closer to what many suppliers and inspectors would feel comfortable with. I was tempted to use Datum Q instead of B and Datum R instead of B, but I am thinking that utilizing Q and R would not gain me anything more than what 'implied symmetry' might be doing in another discussion now getting scrutiny. (I do realize that there might be situations where it can be justified, but I can also imagine that it is something I might feel good doing only because I 'like' symmetry, regardless of it's actual benefit.)

Peter Truitt

 

[Belanger]

Well, for one thing, to use datum "Q" would violate paragraph 3.3.2 of the 2009 standard. (Don't you just love rules?)

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[ptruitt]

"Don't ask, don't tell" was repealed on the 22nd, so 'Q's can come out, now.

Peter Truitt

 

[ptruitt]

Ummmm. Maybe I broke another rule in addition to 3.3.2. I sure hope no one takes any offense. My last post certainly wasn't intended to offend anyone at all.

Peter Truitt

 

[CheckerHater]

Thank you Peter,

I liked your drawing, in fact I liked your first drawing more

Unfortunately you did not use Mismatch on your drawing and the limits of using "mismatch" is exactly what I am looking for.

Something like "if it could be described using "MIN" and "MAX" (mostly MAX), then it's mismatch, but if it requires more detailed measurement, use profile"

 

[CheckerHater]

Sorry, pluged it into wrong discussion

 

[SeasonLee]

Peter: Here is my comment on the sketch.

Datum targets B1 and B2 are located on the parting line, this will cause the gage and measurement unrepeatable. Datum targets should be located on features not subject to processing variables, such as parting lines, flash extensions, etc. Datum target B1 and B2 need to be moved or relocated at a suitable place, please ref to 4.3.1 Datum Target Location ASME Y14.8M-1996 for more detailed information

SeasonLee