Cpk of profile tolerance?

[AndrewTT]

I have a cast part that is described with only basic dimensions. There is a profile of a surface (all over) call out on this part.

Some at my work have a desire to require that some of the basic dimensions (size features) used to describe the true profile be critical characteristics. Traditionally, at my work, a critical characteristic requires capability to be shown.

I believe that what these individuals are actually after is some kind of way to track tooling degradation in certain areas of the die/mold and/or give the operator, making the part, a quick caliper check to ensure that the parts are good without having to have a scanner or CMM at their work station.

I have relayed the information that just because the size is within the profile tolerance does not mean that this area of the part is fully complying with the profile tolerance (location & orientation may still be out). They understand this but still want some measurement (less than the entire part) to be made and to have this data used to evaluate the part/process.

1) Has anyone ever required something like this?
2) Does this qualify as “NONMANDATORY (MFG DATA)” (section 1.4 - f)?
3) If not then what would/should this requirement be called? Additional Size Requirement? Supplementary Size Requirement? Tooling check? Etc.
4) How do you show capability of a profile call out? How do you show capability of a small portion of a profile call out?

 

[drawoh]

AndrewTT,

If you are the customer/designer, you care that the part meets your tolerances, in this case, an all-over profile tolerance. Otherwise, you don't care what the fabricator does, short of sacrificing virgins. I see no reason why manufacturing should not take measurements to monitor their process. There may be a dimensional drift that tells manufacturing when to repair or replace tools, even though all the parts meet your specification. Since this is a manufacturing process, I see no reason for it to be on your specification/fabrication drawing.

Would a simple fixture allow them to monitor the process by measuring from datums?

--
JHG

 

[AndrewTT]

The manufacturer can take all the measurements they want/need to monitor their process. As you stated, I don't really care about any of that as long as my parts show up to print. What I am worried about is that we are trying to force a capability requirement onto a basic dimension. This does not make sense to me. I was wondering if anyone had run into a similar situation and how they handled it.

We could have a "simple" fixture to verify the parts but it would only provide attribute data (good part or bad part) so we could not calculate Cpk.

Our part has no datums. The all over surface profile call out is independent ("best fit" approach). When I said "just because the size is within the profile tolerance does not mean that this area of the part is fully complying with the profile tolerance (location & orientation may still be out)." I meant the location and orientation of that region of the part could be out with respect to the rest of the true profile.

 

[powerhound]

1) This sounds like an automotive spec and I can tell you from personal experience that whoever put the <CC> on the dimension is probably not aware that they are also asking for capability. They just waant to ensure the dimension is inspected and must be in tolerance. This only anecdotal information but at my company this exact issue is widespread (drafters not knowing that <CC> also means capability).

2) I don't think so.

3) That's up to your company.

4) The first part of your question: If your process is in control and the distribution is normal then just run it on any Capability calculator. Non-normal distributions may need transforming unless there's historical data to show the distribution is going to normal. The second part of this question: DEtermine which part of the profile you need to analyze then just gather data from there.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[jassco]

Hi, AndrewTT:

A BASIC dimension does not have a capability requirement. Capability only applies to a specification. In your case, it is the overall profile. Well, do what you say and say what you do. If you have an overall profile specification, then you measure it and report it properly. If you only want to measure partial specifications, then you have to establish multiple specifications so that you can measure it accordingly.

Best regards,

Alex

 

[Nescius]

Maybe I'm too philosophical about geometric tolerances, but this issue is related to many others that crop up here on the forum. Attempting to report basic dims on a FAI, for example. Even the perpetual nemesis, "unless otherwise specified", is hiding in the shadows here.

I am a hard-liner. I would say that there is a single requirement for your part. Single. If you want to study the statistical behavior of meeting that requirement, I would argue that you must determine the narrowest best fit all-over profile zone that would contain the surfaces of the part. That may or may not provide some useful data. It would be interesting to see the numbers, but I'd fear they'd be awfully muddy, depending on the complexity of the part. Ball bearing, sure. Complex casting, no.

For a complex part, you could map those numbers onto the part and have a sort of hot/cold color model for variation. I have absolutely no idea how to do this.

My opinion? Leave the geometric tolerance alone to define the part. Don't try to add some weird additional requirement scheme just for the purposes of gathering data. Don't pollute the tolerancing to achieve goals unrelated to part definition. Then, in a separate document, outline what you want to measure and study, the how, the why.

 

[AndrewTT]

Nescius, I think we may have been separated at birth.

I toy-ed with the idea of proposing a "critical region" instead of applying a CC to a basic dimension. This would ask you to focus more measurement effort in this region (and we can argue about how to demonstrate capability later). However, this does not address the other aspect that my co-workers wanted, which was a "quick" way to check parts. But putting this type of info. on a separate document might help this aspect.

 

[powerhound]

Okay, so maybe I missed the mark. I thought you were asking about capability of a profile of a surface callout. If you're asking about showing capability on a basic dimension then it has to tie back to a geometric tolerance like profile of a surface. Did I miss your intent?

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[AndrewTT]

What was desired (by some) was Cpk of a region of the part, even though the entire part was controlled with a single surface profile call out. This desire was initially demonstrated by a CC symbol next to a basic dimension in the region of interest.

I questioned if this was a good/bad idea and how one could (easily) show capability of a binary (good vs. bad) measurement.

 

[dgallup]

1) There is no way to measure the capability of a basic dimension as it has no tolerance.

2) There is no general requirement of capability for a TS 16949 (assuming Automotive) Controlled Characteristic. A specific company may have such a requirement but TS 16949 does not. It's a terrible standard because the definition of Controlled Characteristic is left up to each and every customer. We use 3 Controlled Characteristic symbols, only one requires SPC.

3) Manufacturing is free to put what ever they want on process documentation as long as they include the product definition requirements. QA can put extra measurements on their control plans if they see fit. Maintenance can put out preventative maintenance requirements to keep the process humming. None of these belong on the product drawing.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.