【转帖】sig figs and tolerance

yang686526

sig figs and tolerance
settle an arguement...
there is a dimension of 6.7mm +/- 0.25mm
ignoring the sig figs in the dim the limits should be 6.95 and 6.45 however since the dim only has 2 sig figs the limits are 6.5 and 6.7.  agreed?
now what if you have multiple tolerance?  do you round the tolerance (e.g. the tolerance become +/- 0.3mm) and apply it to all the dims or do you add them all up and then apply the rules of rounding.  (e.g. 2.0 + 2.0 +2.0 = 2.25 + 2.25 +2.25 = 6.75 = 6.8?)
thanks,
sean
eng-tips forums is member supported.
the only sound solution is to match up sig figs on tolerance and dimension.  the safe assumption is to add a zero to the 6.7 dimension.
"ignoring the sig figs in the dim the limits should be 6.95 and 6.45 however since the dim only has 2 sig figs the limits are 6.5 and 6.7.  agreed?"
no.  based on the stated assumed limits, you would be rejecting good parts. if ±.2mm was the design intent, the drawing should state such.
i agree with thetick.
when the people fear their government, there is tyranny; when the government fears the people, there is liberty.fff"> - thomas jefferson

seandotson,
   significant figure are irrelevant here.  asme y14.5m-1994 requires that you delete trailing zeros from your dimensions.  in any case, the 6.7 is a requirement, not a measurement.  significant figures apply to measurements.
                       jhg
what drawing standards are you working to?  asme, iso or some other.  not sure it will make a difference, i'd interpret it as 6.70000000000000000000000000000000000000000000000... +/- 0.25000000000000000000000000000000000000000000000...
can't recal the exact reference but probably somewhere like asme y14.5m section 1.6.1
kenat, probably the least qualified checker you'll ever meet...
thanks for the chapter and verse, drawohfff">.
i just found out the ref to back up my post, asme y14.5m-1994:
quote:
2.4 interpretation of limits.
all limits are absolute.  dimenisonal limits, regardless of the number of decimal places, are used as if they were continued with zeros
however, off the top of my head i don't know what iso may say, but i'd expect similar.
kenat, probably the least qualified checker you'll ever meet...
yup.  max .25 means .25000000000000000>infinity.  if a measurement is .250000000000000000001, the part is out of spec.  if someone feels .250000000000000000001 is acceptable, then the spec should be changed to allow it.
matt lorono
cad engineer/ecn analyst
silicon valley, ca
thanks for the info guys.   
the answer is correct for both metric & inch, however number of decimal places is a different thing. in metric you truncate trailing zeros from both the nominal and tolerance value, which means that you may end up with a different number of decimal places in each.  for inches, you must have the same number of decimal places in both the nominal and tolerance value.  that wasn't the original question, i know, but it seemed to be the source of sean's confusion.
jim sykes, p.eng, gdtp-s
profile services
cad-documentation-gd&t-product development
jim,
you are hijacking this thread.   anyway, i do think your point was covered above in part.  again, there's no law about this and as long as rules are established in advance, a company can choose to use .xxx .xx style dim'ing for metric.
matt lorono
cad engineer/ecn analyst
silicon valley, ca