Datum Targets

[phil george]

I have a part with 4 target points called out for datum A, two for B, and one for C. Using an iterative, the 4 A points solved within .2mm (best it could do). Now, after completing the job, the customer (an engineer) says I should have used only 3 of the targets for A and that I should have known that. I'm a bit stupified at this claim since the print specified four targets for A but, as always, I could be wrong. Anyone heard of this?

[Matthew D. Hoedeman]

NOPE, just more proof that engineers are idiots, for the most part. If the print says there are 4 a-datums, then you use 4 a-datums. No if's, and's or but's about it. Period. If you don't use all the datums specified, then it is not aligned correctly. If, in your case, you used only 3 of the 4, then the 4th one, the one unsed, would check 0.4mm off (at least). Nope, use them all, just like it is called out on the print. He hasn't got a freaking clue.

[bw_bob]

In my opinion the print calls out 4 target points then he gets 4 target points. If he only wanted 3 the print should be changed to reflect this. All you have is the print to go by for setting up the part. My .02

But everyone one here knows that engineering is never wrong......

[John Kingston]

WMS and WBS

[Peter Warcholyk]

Phil,

Perhaps the "engineer" should go back to school.

Unless otherwise specified, you are correct in using the 4 "A" points.

I have worked in plastics for quite some time, especially dashboard for vehicles. There may be as many as 15 to 20 net pads all relating to my "A" datum.
As for the person telling you that "you should have known" which of the 3 targets to use, this is hogwash.
If they wanted you to use 3 target points, then the print should only have 3 datum targets on it. Case closed.

Regards,
ZydecoPete

[phil george]

I thought as much. Just wanted to get the opinion of the real experts. Thanks.

[MrComment]

That engineer must have worked in production or management before design. He's probably from the school of thought that you just use the 3 that make the part look good. Isnt it obvious??

[Hilton Roberts]

What Matt said.......

Got to REALLY look at the print.

Interesting thing happened here. We had two tools subbed out to a vendor. Right and left hand tool.

Got brought in here and by passed inspection. One tool produced big parts and one tool produced small parts yet the tools are mirror images.....HOWZATPOSSIBLE????

We take the big tool and send it to our "Black Hole" and blue the tool up and run a nominal cutter path around it. If the tool is to design ( what the subber told us ) then we should not take off any material......right...???

Well, we take off material and check the tool on our cmm and now it is good.......hooray!!!!

We decide to remake the small tool and send it to another vendor who subs it out to a job shop.

In our world, most blueprints for symetrical parts state on the face of the drawing, "LEFT HAND SHOWN, RIGHT HAND OPPOSITE"

We had a non-standard drawing that clearly states on the face of the drawing "RIGHT HAND SHOWN, LEFT HAND OPPOSITE"

The tool we want built is a left hand. After two months, we get the tool in house, blessed by two vendors and because of the problems in the past, the powers that be send the tool to me to check out.

I look at the tool, I look at the drawing and the tool looks like the tri-view on the drawing.

I check it anyway and it is out of spec so I squawk the tool and get it out of here.

The vendor wants my inspection data ( which I supplied ) that shows the tool is out of tolerance and they still have not figured out why they gave us a right hand tool and we asked for and contracted for a left hand tool?


I mean who really cares now if the tool is out of spec. IT IS THE WRONG HAND TOOL............

I think about three days of steady drinking will put a positive spin on this fiasco......

[Matthew D. Hoedeman]

Thanks, bro. Have not heard from you in a while. Nice to know you are still out there in Pcdmis land!

[Jeffrey Lovely]

sounds like he did not like the data that you provided and done correctly
You may want to let him know the part will check different in a fixture
eliminating the .2 error
and that the note should be applied to the print
"Part must be measured while restrained in a fixture" - This is common

[T Miller]

Got the same thing 4 points are called out for datum C ,2 on each side of the part. Datum C is called out to be offset in X by x amount. How can that be I ask - oh we only want you to use target c1 c2-to c4 are just for info to see how wide the part is... So you are not alone

[JoeRZ]

Dude - Is reading mind part of your job requirement?

I get same thing all the time...He ll, if I could read minds I'll be in first flight to Vegas!

The design intent specifically points out what is required and shown on the print. The print is your roadmap.

APQP before build - clarifies what's needed to get desired outcome.

[cmmguy]

I have a part with 4 target points called out for datum A, two for B, and one for C. Using an iterative, the 4 A points solved within .2mm (best it could do). Now, after completing the job, the customer says I should have used only 3 of the targets for A and that I should have known that (even though he doesn't say how I should have known which one to leave out). I'm a bit stupified at this claim since the print specified FOUR targets for A but, as always, I could be wrong. Anyone heard of this alledged common knowledge? And yes, he's an engineer.

as Hilton said, you have to look closely at the drawings.

Sometimes when 4 targets are identified, two of them are equalized to each other to form single point. Datum targets are usually intended to be used for the workholding (material location purposes), it is unlikely that 4 targets would be rigid, therefore bestfitting between them seems unlikely. Unless the part is intended to be "restrained" against the 4 targets.

[phil george]

Yeah, these are not the workholding type of targets. They're on an assembled part that's part of a larger assembly. All 4 of the target areas have a mating surface with another part. But in any case, the print just has the 4 datum targets and the engineer in question told me I should have used 3 of them and reported out the deviation of the 4th.

I don't think I've ever seen a case where 2 of them are supposed to be equalized into 1. How is that called out on the print? Isn't the iterative basically equalizing all 4 of them as best they can be?

[phil george]

Since we're on the subject (sort of) of engineers, I was just reminded of an incident about a month or so ago where a whole group (4 or 5) called us to ask how we were calculating true position on our reports (the standard Pc-Dmis report). When we got to the part about multiplying by 2 they collectively lost it and demaded to know why we do that (presumably because not doing so would have made the parts good), and when we explained it to them they wanted some kind of proof, so I copied page 320 out of Geometrics 3 and faxed it to them and never heard from them about it again.

How is this possible?