Reporting decimal values 4 places on a 3 place tolerance dimension 1

[notats]

Is it common practice to report/record 4 decimal places on a 3 decimal place dimensional drawing callout?
Example: .075 +/- .001 , would you record to the 4th decimal place ( .0753 ) to show all deviations?

 

[ctopher]

Anything, regardless of decimal point, in the range of .074-.076 is acceptable. Depending on what measuring tool is used, some read up to 4-5 decimals.
I would record whatever the tool indicates.

Chris, CSWA
SolidWorks '15
SolidWorks Legion

 

[RoarkS]

If your equipment is calibrated to do so and does it automatically why not. My CMM reports in 4 places on everything... it's good for .0001"
If using hand tools... I would look at ya funny for using micrometers out to 4 places on a Ø.063±.005 dim.

But if it's Ø.063±.005 and you're close to the top end and don't want to scrap the parts... you pull out the micrometers and you measure .0679"... beautiful. Just make sure your stuff is calibrated/certified to read what you are reporting.

Remember Ø.0680000000000000001 is an out of spec part for Ø.063±.005 per ansi y14.5 which shoots ya over to ASTM SI10.

BTW... I would love it if someone proved me wrong on that one because I hate saying that I measured something with a .0001" micrometer that says .0680" is a bad part just because you can't prove it's not over.

 

[3DDave]

If the makers of the part are playing that close to the border, get new ones. In fact, parts that are within reach of the tolerance limits based on the ability of the measuring device should probably be rejected as well. Whatever the error or least reading, or both, should be subtracted from the tolerance limits to avoid accepting out of tolerance parts.

The day that happens is the day someone wakes up to all the 90 degree angles between surfaces of parts that they haven't been adding into the reports either. A rectangular block has 12 of those to measure if it isn't controlled with FCFs.

 

[dgallup]

For your measurement system to be sufficiently accurate, it needs to have at least ten divisions on either side of the nominal. So for the example given by the OP (0.075+/-0.001) the measurement system MUST read to at least 4 decimal places. There is nothing wrong with having 5 or 6 decimal places but that kind of accuracy does not come cheap.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[JNieman]

In-house and unless otherwise specified, we report as many decimal places as the tool is calibrated to measure to. If it's a single place dimension but we use a vernier micrometer, we'll take it out to .0001 increments. We at least want to see one place further than the dimension itself.

We do have some customers that demand all quality reports show 'actual values' to the same precision as the dimension. So a 3-place dimension gets reported as a 3-place actual condition. It's up to us to know that .7605 is not acceptable to report as .760 if the tolerance is .740-.760 - I think that's a risk they take, in my opinion, but it's not my place to say. I am betting they have other suppliers less-rigid as we are, who may just round it and call it good.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5