True Position - Drawing Issue?

[[Br...]]

Please sign in to view this quote.

That's probably true. If a lawsuit was filed, the company who made the drawing is most likely going to lose, but by then, you've probably lost a costumer unfortunately. It's best to try and communicate with them to find the intent and go from there. The purpose of the standard is to be clear an objective, but if that hasn't been conveyed properly, then you probably need to ask questions to avoid any future issues.

[[An...]]

Please sign in to view this quote.

This is a very bold statement,can you prove it please?

Designer's decision #1:
Not Poka Yoke
Rectangularity and parallelism of the four relevant planes within small tolerances.
Effect: high production costs.

Designer's decision #2:
Poka Yoke
Rectangularity and parallelism of the four relevant planes within large tolerances.
Effect: low production costs.

It's up to the designer to decide,if the costs are low or high.
And the poor CALYPSO people only do what the designer wants. (n'est ce pas?)

[[Ja...]]

Please sign in to view this quote.

Derek, I agree with you. In my case, I strongly believe that I can program all of the configurations. The problem that would arise is some of the reported dimensions will be bad in some cases. The inspectors simply run the part and that's that. Selective reporting can quickly become a very slippery slope (and I am not implying that this is what you are doing, I think our situations are slightly different). One of our engineers suggested that, in house, we hold each side to the same position (allow C/L deviation to affect position). This will be a tighter hold, so we can say with certainty that if we see a good part, that part is good. Unfortunately, the higher-ups are seeing this as more bad parts down the line, and will not allow it.

I am at a loss, I will not be confident in the program.

Please sign in to view this quote.

Brett, this is my point. Our lack of communication may very well lead to this one day. I am trying to make it a quality issue as Derek suggests, but I am not the one to reach out to the customer for clarification. I would lose my job with how many toes I would have to step on.

I think we're all agreeing that this is a design flaw that may lead to miscommunication. The standard only suggests that datum features be identified on a symmetrical part. I guess without the rigor that I am seeking, I am not the one to make the change. This may always plague me.

Thanks everyone.

[[Br...]]

Please sign in to view this quote.

Its not bold. The drawing has a spec, the spec was met. Its really simple. I'm not going to keep repeating myself.

[[An...]]

Please try to understand what POKA YOKE means!

[[De...]]

In our case we aren't selective at all, if any orientation has something wrong then its considered a bad part.

That was what I was attempting to explain. When we began making the parts we made sure our process was able to make the parts so they came out good regardless what orientation they were checked in. We still check all the orientations in one program and we report the worst case.

If we find a part that has 11 orientations that come out good and one bad we consider it a failing part.

It comes down to whether the customer will work with you or not.

If they will, then great - they can update the blueprint. Unfortunately in many cases the engineer for the customer is either unwilling or not allowed to make changes to existing prints and that leaves us dealing with the problem of how to interpret the ambiguous call out.

We have made the determination that in order for the part to be good ALL possible orientations of the part must be good, like someone already said - someone needs to assemble this part and it is unlikely they will do a full inspection prior to doing that.

[[Ja...]]

Derek,

I interpreted what you wrote incorrectly. I would like to have the ability to do what you do, as I feel as though it is the safest way.

[[An...]]

Please sign in to view this quote.

Definitely, yes!

[[Me...]]

But the part hasn't changed. It's the same part. The drawing shouldn't allow for 11 configurations when a part has not physically changed. If it passes 1 it's good. You can't change the datums for each check.

In the end it's all about the customer. I would make my case, bring up cost and time, and then it is what it is.

[[Me...]]

Please sign in to view this quote.

And if the Datum feature isn't clearly marked on a symmetrical part then it either needs to be identified so that the correct datum is used each time or you're allowed to use whichever you choose. It's the same for any symmetrical part. If you have a square block with a symmetrical hole pattern, and the datums are the top and 2 sides, you don't have to check the part 4 times, rotated 90 degrees.

If you're inspecting your part in 4 different configurations, 3 of them are wrong, because you're changing the datum features each time. There can't be multiple definitions of a datum feature on the same part. So which configuration is right?

Food for thought.

[[An...]]

Please sign in to view this quote.

..... may be necessary?

or

..... is mandatory?

[[Ja...]]

Please sign in to view this quote.

Does it always make sense to take the configuration that measures best as the correct configuration? I think, as Andreas is saying and which ASME does not specify, that the datum features must be physically identified to ensure that everybody measures the same thing. If you measure it OK and ship it to me and I measure it NG, then why am I wrong and you are right? I would think that you're wrong and I am right. Who is the ultimate judge? Are we both right? Both wrong? This situation needs to be defined.

There is an inherent possibility of large variance between measured results. This should be stomped out, and it's as easy as the ASME community getting together and changing the word "may" to the word "must". We would have consistency between results and we would be forced to machine better parts/not be sneaky by measuring until it's good.

[[Br...]]

Please sign in to view this quote.

ASME Y14.5-2018 Para.1.4.1
(a) The word “shall” establishes a requirement.

Para.7.8
"...Datum features shall be readily discernible on the part."

However you make them readily discernible, is up to the designer. One possible way on a symmetrical part is a physical identification. So if you have a better way to make the datum features readily discernible, that would be fine. It's an example of how to make a datum feature discernible on a symmetrical part.

[[Ja...]]

Brett wins.

ASME requires that datum features be discernible on the part. I didn't see 1.4.1.

Now this situation is resolved. In my case, the datum feature is not readily discernible. None of them are on our symmetrical parts. I need to talk to our/their designer.

Thank you everybody for pitching in, and thanks Brett for digging up 1.4.1.

[[Br...]]

Please sign in to view this quote.

I guess that's my fault for not clarifying that earlier. That's been the argument the entire time. I think we got a little side tracked with physical identification part, instead of asking why that is necessary. Aaron earlier quoted this from the standard as well. So I guess I assumed that was understood.

Really though, the problem for you I still think might be solved by doing what Derek suggested. If its not too much to ask of your machine shop, it would probably avoid a lot of headache with having this exact same debate with your costumer.

[[Ja...]]

Please sign in to view this quote.

Well I saw his quote for 7.8 but I didn't see their definition of their usage of "shall". I thought ASME implied that it was optional.

But yes, I agree, the way that Derek suggested to measure would avoid much of a headache. I will shoot for this, and I will see where I get.

[[Br...]]

Please sign in to view this quote.

His quote was from the 2009 standard which actually uses the word "Must". We can infer that "must" is a requirement, but they further clarified it in the 2018 standard by changing it to shall. All the Y14 standards use these naming conventions now.

[[Ja...]]

Please sign in to view this quote.

2009 goes on to say

Please sign in to view this quote.

Which implies that it is optional. Unless they are allowing for another form of identification, of which I cannot think of.

[[Br...]]

Please sign in to view this quote.

The practice of physical identification is optional, but it's a requirement that the datum features be readily discernible. You cannot discern which features are the datum features by the drawing you have. When you hold the part in your hand, you cannot point out which side is datum feature B. There's not enough information on the drawing for you to make that distinction. So they go on to explain that with a symmetrical part, you may need to put a physical identification to make them readily discernible.

[[Ja...]]

I assume with their wording (may) that it is optional which datum is called which. But as Derek has suggested, to avoid any potential conflict with the customer, it may make sense to run it each way and only accept if all pass. We really only need to run once, create a second alignment, and some extra results. I just need approval, which they do not want to give. I was hoping that there was some ASME/GD&T rule that I could bring up to ensure approval. So I may just be keeping things the way that they are.

ASME should define what they mean by "may be necessary". The word may implies a conditional, which requires further definition.

[[Br...]]

Please sign in to view this quote.

It's a requirement to make the datum features discernible. Its clear as day in the standard. Forget about the physical identification part. This has not been done on your drawing. It doesn't follow the standard. They do define what "May be necessary" means. Its an allowable practice. But what violates the standard on your drawing the sentence before that.

[[An...]]

Please sign in to view this quote.

Thumbs up! It definitely makes sense to run it each way
It just takes some time and a lot of paper sheets,so what?

Advice to the designer in order to avoid long discussions in the future:
Bring up a text line into the drawing with the following contents:

***** THIS SYMMETRICAL PART HAS TO BE CHECKED IN 4 ORIENTATIONS *****

[[Me...]]

Please sign in to view this quote.

I totally disagree. You're putting a requirement on the inspection process that isn't defined by any standard, nor by the drawing itself.

[[Ja...]]

Please sign in to view this quote.

I disagree as well, but only because in most of my cases, you measure all of the features in one shot. All you need to do is offset the alignment to match each possible DRF and gather results this way instead of manually rerunning the program.

Robert, ASME Y14.5 says that there should be no ambiguity between datum features. In my case, there is. This is a design issue that should be brought back to the customer. Nobody wants to do that, and they will continue to run the part in whatever way makes it good (some people here agree with that, I assume fix the design so that it won't be possible) and the way that I will accommodate them is to report the best result, thus saving inspection time that we are currently using up. Incorrect? In my opinion, yes. But what is correct when the part is not defined to the requirements of the standard?

[[Me...]]

I totally get why you're checking it the way you are and the customer side of this. I don't disagree with any of that from a vendor standpoint. This is more academic to me and about what a drawing does and doesn't say.

Take datum's and GD&T out of the equation. You have a simple square part with a symmetrical 4 hole pattern to all 4 sides. Everybody has seen this part. The drawing has linear dimensions from 2 sides locating the holes. The hole tolerances are +/- .005. The length and width of the part is +/- .030. Which side do you measure the holes from and why? It's the same issue and the logic that drives my response on the previous part.

As far as the original part, it probably only functions based on a hole to hole distance and the orientation of the bores to each other. So the GD&T probably doesn't accurately express the functionality of the part anyway. But who knows.