几何尺寸与公差论坛

 找回密码
 注册
查看: 1033|回复: 0

【转帖】round off rules

[复制链接]
发表于 2009-4-29 21:46:49 | 显示全部楼层 |阅读模式
round off rules
what round off rules do you all prefer for decimals? if the decimal ends in 5 we have always rounded up if it comes after an odd number and down if it is even.
for example, a length is 1.125 inches. if from the start i intend it to be a 2 place decimal to use default tolerances, i draw it as 1.12. however, if later on i change my mind i can either go back and redraw it at 1.12 or accept the cad default of 1.13. unfortunately, that bugs the hell out of the checkers because it breaks the rules.
with autocad or solidworks i don't think you have the option of changing the rule. i would just as soon take the default as long as it does not affect the function of the part.
(related to this, you would not use .70 for a dimension when you could use .69, because it is equivalent to 11/16. i don't see a lot of reason for perpetuating this.)
there is a standard covering this.  it used to be an ansi standard, but now may be a asme standard.  sorry, i don't know the standard number.
round off rules are covered by an asme standard when it comes to engineering drawings. the dimensions is always rounded to the even number.
if your cad system doesn't conform to asme drafting standards, get a better system.
"wildfires are dangerous, hard to control, and economically catastrophic."
"fixed in the next release" should replace "product first" as the ptc slogan.
ben loosli
cad/cam system analyst
ingersoll-rand
what is a better system than solidworks?
there might be a way to change it but i haven't found it yet. still learning.
i was always taught to round down for 1,2,3 and 4 and round up for 5,6,7,8 and 9, as far as i am aware all cad systems do the same.
a quote from ug help documentation (which follows the asme standard):
if the numbers after the required precision (number of decimal places) is less than 5, then there is no change in preceding digits (round down). for example, 2.4634 rounded to three decimal places would be 2.463.
if the numbers after the required precision is greater than 5, the preceding digit is increased by 1 (round up). for example, 4.37652 rounded to three decimal places would be 4.377.
if the number after the required precision is exactly 5, round off to the nearest even number. for example, 8.36500 becomes 8.36 when rounded to two decimal places. 8.35500 also becomes 8.36 when rounded to two decimal places.
according to ansi y14.5m - 1982 section 1.6.4 conversion and rounding of linear units....see ansi z210.1
i'm not sure if it's covered in asme y14.5m....my copy is at home since my company hasn't updated.
best regards,
heckler
sr. mechanical engineer
sw2005 sp 4.0 & pro/e 2001
dell precision 370
p4 3.6 ghz, 1gb ram
xp pro sp2.0
nivida quadro fx 1400
      o
  _`\(,_
(_)/ (_)
"there is no trouble so great or grave that cannot be much diminished by a nice cup of tea"  bernard-paul heroux

asme y14.5m-1994 section 1.6.4 conversion and rounding of linear units.
for information on conversion and rounding of u.s. customary linear units, see ansi/ieee 268.
"wildfires are dangerous, hard to control, and economically catastrophic."
"fixed in the next release" should replace "product first" as the ptc slogan.
ben loosli
cad/cam system analyst
ingersoll-rand
engjw,
im not sure that i correctly read your statement about not using .70 when .69 equates to 11/16. but,if i did,i would disagree with that statement.  also with regards to the length of 1.125.  how did we arrive at that when 1.20 would be easier to deal with?
i have been retired for a few years now, but i seem to recall a standard that we should design to decimal parts of an inch rather that fractional equivalents.  has that been superceded or just merely forgotten and discarded.  it certainly seemed to make sense at the time?
it would tend to eliminate some of the confusion with the rounding up or down also.
mr. ringman,
most of our old drawings were based on fractions. if i design a part based on the old one and see an 11/16 dimension, i usually round it off to two places and make the new part .69. our default tolerance is plus/minus .010 for two place, so i make everything two place unless more precision is needed. now, for a totally new part there would be no reason to use .69 when a nice round number of .70 would work. however, some checkers who think in terms of fractions seem to have fits about this.
one guy will take 11/16 and draw the part as .6875 and let autocad round it off to .69. i can't see designing a part out to 4 places if that kind of accuracy is not needed, but that's just my preference.
actually i like designing in metric better. i do everything in increments of .25mm which is .010 inches. it still gets the default tolerance, and no more hangup about fractions.
if your part is meant to be 11/16, then designing it at .7 is out of tolerance, if your tolerance is +/- .010.
always design the part in a cad system to the actual nominal dimension of the part. tolerenaces are for manufacturing, not for design.
"wildfires are dangerous, hard to control, and economically catastrophic."
"fixed in the next release" should replace "product first" as the ptc slogan.
ben loosli
cad/cam system analyst
ingersoll-rand
您需要登录后才可以回帖 登录 | 注册

本版积分规则

QQ|Archiver|小黑屋|几何尺寸与公差论坛

GMT+8, 2024-12-23 02:37 , Processed in 0.036838 second(s), 19 queries .

Powered by Discuz! X3.4 Licensed

© 2001-2023 Discuz! Team.

快速回复 返回顶部 返回列表