I have a question guys. When the drawing callout is to 3 decimal places example 1.000 +/-.005, for part acceptance should be rounded off to 3 decimal as well or not? if the actual is .9946, can it be reported as .995?
I am looking for something in the standard y14.5 -94/2009 , PWA360 or really any other where it would explain how dimension should be reported and to how many decimals.
I've been reading some older threads, and came across a statement that explains per y14.5-2009 that all linear dim. are absolute regardless of number of decimal used as if they did continued with zeroes. So, if my interpretation of this statement is correct, then for the dim 1.000 +/.005 with a 10:1 resolution rule (of the tolerance), I would have to measure the feature to 3 decimal, and with an actual being .995 not .9946 ?
However if I have a dim 1.000 +/-.002, using the same 10:1 rule, I would then have to measure the feature to .0002 resolution, and with actual being .9946, it would make it OOT correct? and I cannot report it as .995 right?
Taking into account the uncertainty, you have to choose if you're the seller or the buyer..
It's the difference between rejecting a good part and accepting a bad one.
In your case, the main value is OOT, but with the uncertainty, maybe it's not... or maybe it's more OOT.
If the uncertainty is around ±0.002, the real value of the dim is between .9966 and .9926 !!!!!!
If you declare that this part is good with this uncertainty, you will be wrong at 60%!!!!!
The process shouldn't be running that close to the spec limit anyway. Quality is about process control not individual parts. If you have one sample right at the limit, how likely is it that you've found the worst part in the lot?
2.4 INTERPRETATION OF LIMITS
All limits are absolute. Dimensional limits, regardless of the number of decimal places, are used as if they were continued with zeroes.
EXAMPLES
12.2 means 12.20...0
12.0 means 12.00...0
12.01means 12.010...0
Interestingly enough, this question always occur in connection with a minuscle amount OUTOFTOL- it seems no-one *ever* questions the result which is equally minuscle INTOL.
To be 'correct', a band around the toleranced limit, as wide as your uncertainty, should really be colored "UNDECIDED", neither INTOL nor OUTTOL. No easy feat, though, as it is difficult to establish actual uncertainty, and PC-DMIS don't give us color bands on the OUT side...
I see your point. When you referring to uncertainty, do you mean the gage total accuracy ( like they specify on the back of the calipers +\-.002? And MPE for a CMM?).
The way I see it, the only way around this is to have a guard band of the tolerance, in this case a dim +\-.005 would be gaurd banded to +/-.003. Does this logic make sense?
You can change the output....
Put the curser above the dimension
Press enter to start a new line
Start typing "DISPLAYPRECISION" and when it pops up..hit tab to put the code in your edit window. Then put the number you want (how many decimal places it will show/round to)
If something was really 0.9994 and you had a "DISPLAYPRECISON/3" command in there...PCD would report 0.999
That being said.....listen to the warnings above. You don't want to put your company in a bad spot.