What is the criteria for inspection of a dimension.

[lutky35]

I am working with an existing design of a machined plastic part. The tolerance block states, .XXX +/-.002. Would all 3 place dimension be considered tight toleranced and show the need for inspection.

 

[HGMorgan]

That Title Block Tolerance would apply to all 3 place dimensions on the drawing that don't have a specific tolerance call-out.
Normally, tolerances that close would need to be inspected.

Harold G. Morgan
CATIA, QA, CNC & CMM Programmer

 

[Cervantis]

I agree with HGMorgan. Anything that is running that close should be looked at. The only things I miss are the ones I dont look at. "Trust but verify" is a rule of thumb I use.


Cervatis

 

[MikeHalloran]

That is an extremely tight tolerance for plastic.
To meet it, you not only need to measure every dimension, but you need to do so in a temperature controlled environment, and stabilize the plastic objects for a few days with controlled humidity, too.
If the part is sourced outside, your metrology department and the supplier's will be constantly battling over every lot of the parts.



Mike Halloran
Pembroke Pines, FL, USA

 

[KENAT]

Wouldn't it be better to compare the required tolerance to the process capability for your particular type & grade of plastic and base your inspection plan on that?

That said, it does sound tight, so will probably be pushing the process capability, so will probably need inspection.

Once you get a real feel for your process capability on that specific part, you may be able to reduce the sampling rate on some dimensions - if your production rate/total volume is high enough to justify it.

Posting guidelines faq731-376

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

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

 

[MikeHalloran]

It would be better to set each tolerance to what's actually required for the part to function properly in the assembly/ product.

Setting tolerances to process capability is my second choice.



Mike Halloran
Pembroke Pines, FL, USA

 

[KENAT]

Mike, aren't we talking about manufacture of an existing part to an existing drawing, not designing a new part/changing drawing?

Sure, for designing a part you should probably tolerance based on function while doing your darndest to keep within the process capabilities (of acceptable cost processes).

However, from an inspection point of view, don't you normally inspect to what the drawing says. My point about inspection based on how close the tolerance is to process capabilities is that just arbitrarily inspecting all 3 place decimal dimensions probably isn't the optimum path.

Posting guidelines faq731-376

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

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

 

[MikeHalloran]

I would interpret an order to evaluate/synthesize the inspection plan for a given part, as permission to examine the assembly stackups and the tolerances of all the parts.

If you can complete something like that before you get caught, it can explain a lot of mysterious "parts don't fit" assembly problems and "assembly doesn't work" performance problems.

If you get caught before completion, well, it was a nice job while it lasted....







Mike Halloran
Pembroke Pines, FL, USA

 

[KENAT]

Maybe it's because I come from a background where changing a drawing usually took 2 acts of God and a year or two. It was a build to print world.

Posting guidelines faq731-376

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

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

 

[MikeHalloran]

I'm more used to being God.



Mike Halloran
Pembroke Pines, FL, USA

 

[KENAT]

Well, me too, my drawings always have the correct tolerance scheme on from the start and I don't appreciate some upstart QA inspector second guessing them.

;-)

Posting guidelines faq731-376

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

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

 

[lutky35]

Thanks to all of your comments. They reassured me that I wasn't concerned for nothing.

A little bit more on the existing part print is it was dimensioned to facilitate how the machinist wanted it dimensioned. So because dimension were coming from every edge made the entire part critical.

Also the part is being machined out of Ultem 1000 so the stability is better than most plastics.

My direction on the print is I want to dimension it for functionality of the part. By using datums and a few GD&T I feel I can loosen up some of the outside geometry while still keeping the relationships on the critical geometry. I brought the number of dimension that were critical ( tighter than +/- .005) down from 36 (where they only inspected 6) to 14. However that was still too many for the QA department. Telling me that Project Manager will write a justification on why some dimension need to be check and some with the same tolerance do not need to be checked.

Not feeling to good about releasing new revised drawing.

Thanks again.

 

[MikeHalloran]

Wait till they discover the concept of 'critical dimensions'.

You'll puke.



Mike Halloran
Pembroke Pines, FL, USA

 

[KENAT]

Critical dimensions, what bunk, my stomach hurts already Mike. Argumeng being that if the tolerancing is done properly, then most if not all dimensions to the stated tolerances are 'critical'.

So lutky, you aren't really just coming up with an inspection plan, you're effectively doing a manufacturability review, including quality plan, with scope for design changes (at least on tols).

Seriously, take into account process capability, not just how tight the tolerance is. +-.002 on a single set up on a CNC mill may be easier to achieve than +-.030 on the outside that's done with a flame cut or something. Assuming they are toleranced correctly, then which would you rather inspect?

Posting guidelines faq731-376

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

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

 

[lutky35]

Kenat,

I understand your last statement.

This gets back to my original question about what is considered to be a tight tolerance and should it be inspected.

With the original print everything was +/-.002 with a few exceptions of +/-.001.

My new print made all .XXX +/- .005 unless specified.

The machinist can still make the outside of the part +/- .002 if he needs to, to establish my critical dimensions. I just want my print to state that they are not critical to function of the part.

 

[KENAT]

What is considered a 'tight tolerance' is driven by the process capability. You don't say what the process capability of the machine(s) you use to make these parts is.

Posting guidelines faq731-376

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

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

 

[dingy2]

"I am working with an existing design of a machined plastic part. The tolerance block states, .XXX +/-.002. Would all 3 place dimension be considered tight toleranced and show the need for inspection."

ALL dimensions need to be inspected once whether the tolerances are tight or not.

On an ongoing basis, find out how the part fits and functions. Those dimensions affecting the fit and function require ongoing inspection on a ongoing regular basis.

I have found in the past, there may be a tight tolerance on a dimension that has no bearing on it fit or function while a less rigid dimension may have a bearing.

Dave D.

http://www.qmsi.ca

 

[KENAT]

So dingy, you don't take into account how easy it is for a tolerance to be met by the relevant process?

Posting guidelines faq731-376

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

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

 

[dingy2]

Kenat:

If there is a dimension on a drawing, it must be checked even if it could be easily met by the relevant process.

If the sample was submitted to the Customer, it would come with a marked up drawing (numbers on all dimensions) and a full measurement report reflecting all the dimensions. That is the easy part.

The difficult part is the ongoing control. I would love to say that if GD&T was applied to only features that affect the function and mating relationship, it would be pretty easy but, as you know, some Designers and most trainers in that subject want it applied to most dimensions whether the dimension has impacted the part function or not. What should one do now?

For ongoing control, I would suggest that the Quality Engineer (or whoever does this function) should find out how the part functions and mates to the assembly. Base ongoing control to characteristics or features that could affect it function and mating relationship.

On ongoing control, if the process easily meets the specification of a feature, then lower the frequency of the inspection. On other features where the process is having difficulty meeting the spec. increase the frequency. In all cases, gauges that are easily read and understood should be on the shop floor for Operator use.

Hope this helps.


Dave D.

http://www.qmsi.ca

 

[KENAT]

Dingy, I was assuming (perhaps incorrectly) that this question pertained to ongoing inspection not first article.

I agree on first article I'd generally expect all dimensions to be checked.

I'm always hesitant to buy too heavily into the 'check based on function' line of thought. Maybe it's the kind of parts I've worked on but typically most of the dimensions if exceeding the stated tolerance (assuming this was assigned properly - which sadly is a big assumption) could cause problems. In this case most of the dimension would affect function, so require high sampling inspection based on my understanding of what you put.

It gets a bit to close to the idea of 'critical dimensions' for my liking.

Posting guidelines faq731-376

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

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