ASME Standard for Inspection of an Assembly Drawing

[eng123mech]

Hello,

My company has a case in which part A and part B come together in an assembly drawing. Both parts are on the parts list, identified by find number call out, and are graphically represented on the 2D drawing with enough detail to show orientation of the two components for assembly (fasteners are also called out to put the components together). The issue arises in that the model for part A (which is a commercial item) is not 100% accurate, therefore the visual on the 2D drawing does not match the assembly on the manufacturing floor. More specifically, the edge of part A protrudes further than shown on the drawing - visually obscuring some features on part B (full functionality retained).

Our quality department wants to tag this part as non-conforming to the drawing based on the 2D graphic only. Is there an ASME standard that covers what is inspectable on a 2D drawing? All I've found so far is ASME Y14.24 Section 6.1.3 which gives requirements for an "assembly drawing". From ASME: " [Requirements] d) depiction of items in the assembly relationship using sufficient detail for identification and orientation of the items. Details of a sub-assembly are not normally repeated on the assembly drawing of a higher level". My argument is that based on this excerpt, the only inspectable things should be identification (it's the correct part) and orientation (it's installed in the correct position/location). Is there anything more concrete in ASME or any other standards that can be leveraged here?

 

[3DDave]

That is an unreasonable request from your quality department. Unfortunately I have never found a standard that can be used as a repellent to unreasonable requests.

Drawings are not to be used as shop floor process documents. Creating shop floor process documents is the job of the manufacturing engineering group. They can take photographs of actual hardware for their documents. You need only show the required validation for that drawing level.

 

[jassco]

Hi, eng123mech:

Your part should not be rejected based on the 2D graphic because drawing views are for reference only per ASME Y14.5. Drawing views are not part of specifications.

Best regards,

Alex

 

[eng123mech]

Thanks for the replies! Is there specific verbiage in 14.5 that states drawing views are for reference only?

 

[MintJulep]

Well, part A needs to be inspected against something doesn't it.

Our quality department wants to tag this part as non-conforming to the drawing based on the 2D graphic only.

Part of the quality department's job is to confirm that purchased parts match the documentation. It seems in this case that the part doesn't match its documentation.

You and the quality department need to work together to resolve the discrepancy.

the only inspectable things should be identification (it's the correct part)

In some cases that could be entirely fine. If that's the case for part A, then there should be an "inspection plan" for this part. Something like:

All you need to do is confirm that the part number is correct.

If allowed by your company procedures, that could be as simple as a note on the product definition drawing or document.

Of course, with commercial items, there is always the chance that something will change in a way that's not ok for your application. You have somehow concluded that the differences between the model/drawing and the physical part are not a problem for your application. How? What features and properties are important for your application? Those are the things you want quality to be checking. Those are what belong in the inspection plan.

Imagine that QA does tag part A a non-conforming. The next thing that happens should be the part gets reviewed by the non-conforming material review board. Perhaps you would be a member. You might then say, "It's ok to use this part as-is, because.....". Those becauses are what belong in the inspection plan. Those are the things you want quality to be checking.

What do your company QA/QC processes say about this case? Or does your company just make things up on the fly?

 

[3DDave]

"Part A" literally cannot be non-conforming per the assembly drawing for which it is a component, except if it is not the correct component. It may not be a usable one, but as long as the component number and supplier match, QA/QC must accept it. There are no component qualifications to establish, on the assembly drawing, except the component number. If the component number matches, the component conforms to the assembly drawing requirement for the item.

The part is not purchased to conform to the assembly drawing. They don't send the assembly drawing out for bids to acquire the part based on the appearance of the views of it on the assembly drawing.

It is bad practice not to have a separate acceptance drawing for the component, for the reasons mentioned above, but QA here does not appear to be suggesting that good practice be used, but that a really bad practice be perfected.

The typical approach is to generate a document that includes performance requirements and interface requirements that may include a drawing showing the required dimensional limits allowed for it in the design. This can be as little as an Envelope Drawing or a Vendor Item Control Drawing or a Source Control Drawing or just as part of a specification as most electrical components have. See ASME Y14.24 for information. Some require prior engineering testing to place a vendor part number on the drawing.

It's unclear the path that should happen when QA/QC has been unresponsive to their responsibilities in the design sign-off process and suddenly decide to participate by tagging parts that conform to the drawing requirements without first asking engineering.

No component requirements should appear on an assembly drawing except those that pertain to the assembly effort. If features on one component are used as datum feature references at the assembly level, then sure. If there is match machining on a component that can only be done at the assembly level, also OK. But if the idea is to reject a purchased component for any reason other than it's the wrong component based on the component number/supplier - that should never happen. Any other rejection of a component is only from the component purchase/fabrication documentation.

 

[eng123mech]

MintJulep, I completely understand your perspective here and I understand why quality raised the flag on this one. It is always good practice to raise the question when something doesn't look right. The issue here is that engineering has pushed back on quality and said this is permissible, but quality will not remove the tag unless there is a drawing update to match the visuals to reality. The intent of this post was to find sufficient evidence in ASME that a drawing update (or formal "deviation request") is not required here.

 

[jassco]

Hi, eng123mech:

Plain drawing views are used to visualize dimensions specifications. They are not specifications by themself. I saw this statement either on ASME Y14.24 and Y14.5. I don't remember exactly where this statement is located.

To give you an example, if you draw a cicle on a view, it is NOT a circle until you add a Dia. symbol.

When your QA does inspections, they are supposed to identify all the specifications and create a list of LSL (lower specification limits) and USL (upper specification limits) and balloon them. You can't identify a visual element and report it as LSL or USL.

If I were you, I would ask your inspector which specification he or she uses to reject the part.

Best regards,
Alex

 

[drawoh]

eng123mech,

Is the problem that part[ ]A has a flange longer than that shown on the CAD model? That could be a nasty design problem later on.

What defines your part[ ]A? There ought to be a specifiation document somewhere. You cannot just check if it conforms to the drawing view.

--
JHG

 

[3DDave]

If you give in to this demand you are going to be stopped by QA/QC over and over and over for ever more ridiculous reasons.

I have experienced this as a delaying tactic where some huge screw-up has occurred.

In one example manufacturing complained that part marking was a requirement they could not meet. I offered to load stencils, epoxy ink, thinner, my airbrush and my CO2 tank and regulator and would be in the factory in a few hours drive.

It turned out the factory had not ordered the parts and had none on hand. They were hoping for a week or so of delay as drawing changes were created, checked, approved by engineering, and then sent to the factory for their slow-roll approval, gaining them time to cover their mistake.

 

[MintJulep]

3DDave has posted before about past bad experiences that they have had with QA/QC.

QA/QC and engineering should not be enemies. Work together to solve the actual problem that has led to this situation. (which to me looks like a lack of defined policies and procedures)

If you had a relevant procedure you wouldn't be here asking something in ASME drawing standards to provide ammunition for your side in the fight.

 

[3DDave]

QA/QC in this case is not working with the OP to a solution. They have made a demand that they will not accept components unless the drawing is changed.

I wish them good luck at getting QA/QC to sign off on a procedure they clearly don't want to follow.

There is no need for a procedure to understand that drawings control only what the drawing specifies. I doubt very much that any top assembly drawing for a commercial jet includes the detailed markings on the 10 million rivet heads or the details of windings that might be visible through the frame of an electric motor.

In most assembly drawings many components of subassemblies are excluded and in some cases the majority of some components isn't depicted.

Had this QA/QC been interested they would have made their objection known during the drawing approval process, prior to drawing release, and then again as part of the manufacturing engineering initial process planning, where they would do the work themselves to generate the views they want.

QA/QC should be participating in the initial design concept and the procurement of outside parts. They should review engineering work in process while the manufacturing engineers are creating preliminary plans.

I have no doubt there are companies where that takes place - had one such interaction with a jet engine maker that started interdepartmental meetings before a concept was accepted for development**. Then I ran into a bunch of suppliers where either QA/QC was obstructionist or rubber stampers, in particular in our factory where QA/QC worked for manufacturing engineering, so any production problem became a design problem. Manufacturing and QA/QC would sign off drawings without looking at them because they didn't have to care. If a production stoppage is a design problem then management turns to design and say "Why didn't you design something they can make?" And we would say, "We asked what they could make" and they said "None of your business." Then they would go run off 100 discrepant parts and ask for a drawing change to accept them.***

I even asked for process information to better tune designs so that outside suppliers would be at a disadvantage to keep production in-house. Same thing - "None of your business."

**It was on a CAD blog where someone was asking how to convert a weldment into a casting and this jet engine guy got unhappy that such a thing could happen and related how his company functioned. I related how our company functioned - had a 300 pound part designed as a casting, until the casting vendors either no-bid or had lead times too long. Management said, do a weldment. Was redesigned as a weldment until management found out what x-ray inspection of all the welds on every part was going to cost on a human-life-critical part. Demanded it be a hog-out from billet. Was redesigned until management found out what the materials and machining time would cost. So they demanded it be a casting. The jet engine guy said that he honestly didn't know companies could do that.

*** Not just me. Spirit (Boeing supplier) isn't quite that bold, but how did they manage to be the only one of 3 companies making bulkheads to snowman the rivet holes? Should there be a policy that QA/QC verify the tooling and processes before making hundreds or thousands of irregular holes?

 

[jassco]

Hi, eng123mech:

Your QA/QC is nit-picking. If he or she wants to reject this part A, he or she can try to "reject" it (part A) based on print of part A. Even that, it may not be rejectable depending if there is a specification on the component level.

Best regards,

Alex