How do you GD&T on interface drawing when part datums are no longer available?

[2JL]

Hello everyone

I think the question stated above said it all but a little backstory anyway:

We are about to provide to our customers interface drawing to depict maximum outline, mounting points and other specifications of our finish products. Until now the idea was to use a lot of reference dimensions and very few GD&T and hard dims for mounting points.
But since eventually, we will have to do a first article inspection as we ask to our suppliers to do. We want to make sure what is specified will match as closely as possible the reality of the finish product. And this is raising two issues:

- When we initially design the product, there was not a clear definition of the interface. We roughly imagine what will works for potential customers and from there design all the parts. Detailed drawings for each of these never took the interface as primary focus. Our main concern was to get functional and working end product. Now to defined the customer mountings points, the original datum used at the part level are not longer available at the product level. we have to use surfaces relatively well related for that purpose.
so the question is how do I work out the corresponding tolerances? I think a quick stack-up analysis could lead me to a close approximation but what about the other uncertainties related to deformation during assembly, handling, transportation...? Is there some rule of thumb I can use to bonify the value I will get from the stack up analysis?
Refer to the attached sketch. Hatched surfaces are the functional surfaces available.

- How will the reference dimensions processed? will the inspection results considered always to be in-specs since by definition there is no criteria to decide otherwise? I found these two threads and it seems to lean towards using more hard dims and less reference dims (Link, Link).

Sorry for the long post and thanks for your suggestions.
2JL

 

[KENAT]

So the part is really located by the threaded hole patterns?

How about making them/one of them the pseudo datum(s) and putting reference dims to the outline from that or similar?

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[drawoh]

2JL,

If the datums on your installation drawing are not from fabrication datums, then the numbers you quote are controlled by tolerance stack-ups. Have you accounted for this?

Good practice in GD&T is to use mount features as your datum. This means your mount face and front face are logical datums. I am not so sure about your front outside diameter. Is this accessible, even with those screws sticking out? KENAT's suggestion is probably a better way than what you are doing.

You are calling up your datums incorrectly. Datums primarily are fixturing points, not dimension origins. If datum[ ]A is your primary, always call up A[ ]first, on all views, and on all feature control frames. Datum[ ]A is a face. Datum[ ]B is defined by the two points that extend forward, perpendicular to datum[ ]A. Your Ø60mm[ ]boss is datum[ ]C, and it prevents your assembly from moving from side to side,[ ]only. I would fixture to your datum[ ]C with a vertical slot. Up and down motion is controlled by datum[ ]A. This is all the more reason to use KENAT's scheme.

--
JHG

 

[2JL]

Thanks Kenat!

Yes, when mounted using the side pads, the part ends up located by the threaded holes. But from the front, the central pilot diameter will take precedence. I like your idea but I fear the reference dim will not be well received here since it will lead to an entry in the inspection report and some here will somehow fill the need to assess his conformity.

2JL

 

[2JL]

Thanks drawoh


I'm just doing that at the moment. But what is worrying me is the other considerations I mentioned above (deformation during assembly, handling, transportation...?)


Sorry for the fuzzy sketch, the screws are not sticking out like I depicted, the front side is easily accessible. There is actually two scenarios for mounting: either using the side pads or the front face. I defined (or tried to?) the datums accordingly, and I don't frankly understand why A should be primary globally.

2JL

 

[KENAT]

They don't have to be referenc dims OP, but putting non reference dimensions at an assembly level where there isn't anything you can change to meet those dimensions is problematic.

I also doubt if it's appropriate to inspect to an interface drawing at all - except maybe receiving inspection at the vendor but this is still iffy.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[3DDave]

If the component doesn't conform to the interface drawing, what's the point of an interface drawing? It's what the engineer's based all their tolerance calculations on - it should definitely be used for inspection.

The fight over assembly dimensions is an interesting one. It hinges on the interpretation of the drawing as a manufacturing process guide instead of an inspection document. It also wanders about the idea of who pays if the item doesn't comply - those who fabricated the parts or those who assembled them.

 

[KENAT]

We covered some of this ground previously and I apparently don't spend enough time thinking about it these days to be consistent.

However, I still think this is an area where you need to be careful about effectively 'dual dimensioning' but at different drawing levels.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[drawoh]

KENAT and 3DDave,

This just dawned on me.

Lots of suppliers provide the dimensions of the interfacing structure, complete with positional tolerances for the holes, and surface flatness. This may be a better approach than going full GD&T on the assembly dimension drawing. The interface drawings provide a useful design specification for anyone reviewing the fabrication drawings.

--
JHG

 

[3DDave]

One of the more interesting interface drawings is for the Bern(e) Tunnel; at least that was the name for the one used for military transport portions of contracts. See

https://en.wikipedia.org/wiki/Berne_gauge

It is an ICD for the entire tunnel and is supposed to be more restrictive than any other railroad tunnel that the US DoD would use to move equipment, and is called for in military contracts as the ICD for all tunnels. It includes both local section and section reduction due to track curvature.

In a way it is double dimensioning all the railroad tunnels in Europe. It would be a disaster to have to go with foot-by-foot track&tunnel drawings to get the same information.