Failure to measure conformity of product

[Thor1]

We are an aluminum mill that manufactures coil and plate for the automotive and aerospace industries. I recently issued an internal audit nonconformance to our Plate final inspection department against Clause 8.1 for failure to demonstrate conformity of the product. Currently the tolerances for width and length on the lot tickets have specifications that are four decimal places. They are using tape measures that are only graduated to 1/16". The fractions, of course, can be broken down to four decimal places but they would need to measure to 1/32" to be able to mesure the spec., the way I see it. This is currently causing a big argument on what needs to happen. They are saying the graduations are too small for their guys to read on a 1/32" tape measure. Has anyone faced a similar situation on the measuring problem and if so, how did you solve it?

[Jeff Frost]

Actually you would still not be conforming to AS9100 measurement requirements because you would need to maintain 4 to 1 ratio related to tool accuracy and minimum measurable tolerance.

Yes you can make measurements to 1/32 Inch (0.03125) with the magnification but the bigger question would be are they supplying finished product to the customer?

[Al Rosen]

Originally Posted by Thor1

We are an aluminum mill that manufactures coil and plate for the automotive and aerospace industries. I recently issued an internal audit nonconformance to our Plate final inspection department against Clause 8.1 for failure to demonstrate conformity of the product. Currently the tolerances for width and length on the lot tickets have specifications that are four decimal places. They are using tape measures that are only graduated to 1/16". The fractions, of course, can be broken down to four decimal places but they would need to measure to 1/32" to be able to mesure the spec., the way I see it. This is currently causing a big argument on what needs to happen. They are saying the graduations are too small for their guys to read on a 1/32" tape measure. Has anyone faced a similar situation on the measuring problem and if so, how did you solve it?

Why is it necessary to measure to 4 places? Although a fraction can be represented by a 4 place decimal, a measurement using a tape can't be resolved to 4 decimal places. Change the requirement to fractions or use a gage more suitable to the requirement.

[Wes Bucey]

Originally Posted by Al Rosen

Why is it necessary to measure to 4 places? Although a fraction can be represented by a 4 place decimal, a measurement using a tape can't be resolved to 4 decimal places. Change the requirement to fractions or use a gage more suitable to the requirement.

Unless your mill is doing something I have NEVER experienced a mill doing in my entire career, I have no idea why a MILL is measuring to closer than one decimal of an inch on the length and width of plate, coil or sheet. The thickness, however, is an entirely different matter. As a practical matter, strap marks from bundling, chain and cable marks from handling, etc. ALL produce defects larger than five ten thousandths of an inch on edges of such stock.

Possible explanation and hint for "root cause":
We sometimes ran across the phenomenon of "overtight tolerances" on requests for quotation in our machining environment. It turns out the cause was simply the fact that engineers and draftsmen routinely created part and product drawings in CAD which defaulted at four decimals and never bothered to convert back to three decimal tolerances they really required. I suspect you have a similar situation and should investigate that first before creating a new inspection method and buying new instrumentation.

This all falls under CONTRACT REVIEW and tolerances on orders should be reviewed by Quality and Inspection people to assure they have capability, capacity, and equipment to perform the inspection at the tolerances requested.

[Kevin H]

Overall, I've got to agree with Al & Wes, as a metallurgist with 30 years experience, much of it in steel mills, I'm not aware of any AMS or ASTM speecification/standard that requires width & length to be measured to 4 decimal places, especially with as rolled material. If material is slit, then width may become critical, and measurement with calipers or micrometer may be needed, depending on the customer specification accepted - again, a contract review issue. If provided as cut to length individual sheets rather than as a coil, length becomes more critical. However, most of the specifications for width and length I've seen over the years are +/- fractions of an inch that can legitimately be measured with a precision tape measure. As Wes notes, thickness is another issue, and for sheet typically a digital electronic micrometer would be recommended.

I'd check tolerances accepted from the customer, and how they're transferred to you. If they are coming to from the customer as 4 decimal places and your company is accepting them, then you'll need to be able to measure to them.

[Thor1]

We don't have a problem measuring coil. Plate is where we have a problem. This is thick plate, too. From what I can gather, the customer orders the particular length, width and thickness. They don't issue tolerances ,they just say to use the Aluminum Association tolerances. The AA is the one that gives us a tolerance of 5/16" for example. We are debating if we can change the 0.3125 tolerance to 0.313. If so, we would be OK. We would exceed the requirement without limiting ourselves too much. This is an old mill. The customers have never had an issue with this. I just stirred the pot when we decided to go for AS9100 in the Plate track. We currently are QS9000 plantwide and going for TS16949. Our registrar seems to concentrate mainly on the automotive processes and just skims the Aerospace process. Therefore, many things went left untouched.

[Jim Howe]

This steel thickness thing struck a nerve. Recently i was tasked with finding why our rib spacings were varying. One thing I had to check was the thickness of our bearing plates that are welded to either side of the ribs. These bearing plates were burned from 5/8" steel plate. My audit of 9 pcs, revealed that the thickness ranged from 0.652 to 0.629. Conclusion was that the thickness of the bearing plate could throw the rib spacing off by as much as 1/16". Our tolerance on rib spacing is +/-1/8". NO PROBLEM!!!
Just thought I would throw in some real life dimensions for a large consumer of steel plate. We typically use thickness from 1/16" up to 6.00".

[Wes Bucey]

Originally Posted by Thor1

We don't have a problem measuring coil. Plate is where we have a problem. This is thick plate, too. From what I can gather, the customer orders the particular length, width and thickness. They don't issue tolerances ,they just say to use the Aluminum Association tolerances. The AA is the one that gives us a tolerance of 5/16" for example. We are debating if we can change the 0.3125 tolerance to 0.313. If so, we would be OK. We would exceed the requirement without limiting ourselves too much. This is an old mill. The customers have never had an issue with this. I just stirred the pot when we decided to go for AS9100 in the Plate track. We currently are QS9000 plantwide and going for TS16949. Our registrar seems to concentrate mainly on the automotive processes and just skims the Aerospace process. Therefore, many things went left untouched.

Just as a mere aside:
converting a fraction to its corresponding 2 to 5 place decimal does NOT mean that the tolerance has been converted to a decimal as well.

If someone gives ME a tolerance of +/- 5/16 inch, I use an instrument (tape measure?) which has a discrimination of 1/16 of an inch and I look to see whether the target dimension is within 5 of those marks on either side of the specified dimension. If the reading is a "hair" under 5 marks past the specified dimension, it passes; if over by a "hair" - it fails.