+/-.0025mm tolerance on an OD - What should I use to measure it?

[Kevo25]

Hey Everyone,
I am looking for some suggestions on how to measure an od. The dimension is 4.395-4.400. I was thinking a 4 decimal place mic and a spc processor. We also have a sheffeild gauge which reads to millionth of an inch, but I dont find that gauge repeatable or user friendly. I know which ever method I used I will have to preform a gauge R&R, but I would like some other input.

Thanks,
Kevin

 

[Jim Wynne]

Hey Everyone,
I am looking for some suggestions on how to measure an od. The dimension is 4.395-4.400. I was thinking a 4 decimal place mic and a spc processor. We also have a sheffeild gauge which reads to millionth of an inch, but I dont find that gauge repeatable or user friendly. I know which ever method I used I will have to preform a gauge R&R, but I would like some other input.

Thanks,
Kevin

You'll need some kind of laser or "super" micrometer, and you may the cost prohibitive. Is this part from an external customer? If so, you'll need to find out how they're measuring it and match the method as closely as possible. Once you get out to the fifth decimal place (in inches) things can get very cloudy. In my own experience there are two types of machining houses: those that thrive on "finesse" jobs and those that probably won't be successful with them, but do "run of the mill" (pun intended) jobs well. The former category is experienced with, and equipped for, these kinds of jobs. As I said, I think your best hope is to find out how the customer measures the part, and get in sync with them.

 

[bobdoering]

You might want to consider an air gage. What type of process are you using to generate this OD?

 

[Denis]

In a previous role we measured the width of slit foil.

The foil thickness was 100µ , and the critical variable was the width of the foil.

Typical width dimensions would be 30.00mm +/- 0.001mm.

We contacted Zwick instruments and used a laser micrometer. To remove the influence of operator error, we also made a jig to rest the sample in whilst it was being measured.

I'd suggest that you could measure this aspect with a similar set up. But it won't be cheap to implement. We had been through the usual route of all sorts of verniers etc, but found that the MSA would fail due to operator errors and varying techniques.

 

[bobdoering]

You mentioned an SPC processor. If you are precision machining the OD (grinding, CNC, etc.), I would say getting data acquisition is OK, but do not waste your money on automated SPC - most of the systems are incorrect for precision machining (especially X-bar-R, the worst) See Statistical process control for precision machining

 

[Kevo25]

Hey Everyone, Thanks for your input. Im sorry I havent responded, My internet activity has been very limited the past few days. I found out after some digging around, cause we all know that quality is last to find out the details, We are currently subbing the machine out to a shop that grinds the od. Apparently they are using a micro cator? I suppose its micrometer type device with a very fine resolution. I am going to try and visit this shop and see exactly how they are doing it. I used to work in automotive and we has some tolerances similar to this one and we used a cmm. looking back, I doubt that was the correct method as Cmm's are generally not accurate enough for these tolerances. Its a challenge for sure to measure this part in a company who does not want to pay much for gauging. Same old song and dance tho I suppose..lol
Cheers,
Kevin

 

[bobdoering]

You may want to share the above link information with your subcontractor - especially if you require SPC from them. It will help you both out!

 

[Kevo25]

Thanks Bob! Thats some really great information.

 

[George Holz]

This requires some kind of laser micrometer. No conventional micrometer and certainly no CMM can measure to the point of passing Gage R&R. It's likely that these tolerances are an "engineer's dream". Please let us know if these tolerances are really being held with any proven capability.

 

[bobdoering]

Yes, there may be devices out there with resolution to read .5 micron, but to get 10:1, you are talking 500 nanometer resolution. Now, consider lobing, surface finish and temperature, ponder what variation one find in the 500 nanometer range? Does sound like a dream....