OK. Lets say a “Widget” is 2.000 long +.06/-0
Now, you wouldn’ make it exactly 2.000 long because it would be too close to going undersize right? So, why do design guys do it? What is the point of the 2.000 dimension?
Why not just make it 2.030 and split the tolerance? I realize that you need a number to start with, but should you always try to hit the nominal dimension and not worry about the tolerance? Is it an after-thought by the design guy? Is the thought process something like… “ok, my part wants to be 2.000 long…great!..ohhhhh, but it would be bad if it was undersized…so I’ll put a +.06/-0 tolerance on the thing…” I see it all the time… a hole size of .125 +.003/-0. Why even put .125 there! It’s obvious that the designer wants a .1265 hole, why don’t they just put that on the print in the first place as the nominal? Is it just poor planning?
And btw, what is the definition of a nominal dimension anyway?
I’m not sure about widgets, but in lumber when you go shopping for a 2x4 you actually get somewhat less in each dimension. The board started out as a 2x4 but was sent throug a planer and dried before marketed as a nominal 2x4.
I’m interested to see a technical answer to this, but I think the tolerance is included for a couple reasons.
(1) If you go outside this tolerance, the widget won’t fit/work properly. The designing engineer has allowed for some wiggle room and/or a safety factor.
(2) Setting the tolerance in the design indicates what kind of manufacturing process and what kind of skilled labor is needed. Obviously, smaller tolerances are more difficult and more costly to achieve.
Ideally, the designing engineer would like the 2.00000 dimension to be met exactly, but eventually the precision becomes too difficult/impossible to achieve with manufacturing methods.
In this case, IF the widget came out to be 2.000 it would be acceptable after inspection. That means, you can use the part. Tolerances are meant to describe what is acceptable after inspection. Not how the part should be made, or what size it should be made.
**
Not that simple my friend. It all depends on what the part is used for, what it mates with, and what material you are using. Granted, +/- tolerancing is used incorrectly by many designers. That is why geometric tolerancing is much more useful. But the main issue is that tolerancing is basically a statistical requirment. If I want a shaft or hole which is .281 inches, I know that some of the time I will get a .278 feature and sometimes I will get a .282 feature. What can I accept? I put that into the tolerance.
** The designer may intend/want for the machinist to use a .125 drill bit, but if the hole comes out smaller than .125, then the machinist better bore it out…In other words, the designer is saying “make it .125, but just don’t make it any smaller than .125.”
Nominal means ideal. It probably can’t be achieved though.
The latest thing is geometric tolerancing. Where position of features relative to datums is specified, instead of straight up +/- tolerancing. Then the size of each feature is dependent its position relative to the datums.
Nominal dimensions are “named” dimensions, rather than actual dimensions.
A 12" pipe has an OD of 12-3/4" OD. In old days, the ID of 12" pipe with “standard” wall thickness was 12" but that is no longer the case.
Tolerances with a +0 or -0, are typically used for matching/mating parts. You use the same number for both, but different tolerances, to show they go together.
You also use those tolerances to avoid having tolerances adding up. The part can be machined to greater accuracy than the tolerance, so the machinist does not have to “aim” for the middle of the tolerance range.
nope…you missed the point!
lets say I had a hole on the widget 1/8 inches +0 / -1/32
would you use a 1/8 drill? no, or at least I wouldn’t
but if the hole was called out 1/8 +/- .005 an 1/8 drill would most likely work fine.
my point is this. I think a part made to the NOMINAL dimensions called out on the print should be in the middle of the tolerance already. On a 1/8 hole dimensioned with a +0/-1/32 tolerance, if it was made to exactly 1/8, it would be too close to going over the tolerance and thus being a bad part…so…why is it done?
excerpt from http://www.m-w.com
Main Entry: 1nom·i·nal
3 b : of, being, or relating to a designated or theoretical size that may vary from the actual :
Basically, the nominal dimension is the specified size. The practical reality of it may vary…whether by tolerance limitations or even by what something is called…for example, you may need a water pipe that is 2" wide to get the flow you need, but when you order a 2" pipe from a catalog, it may mean the outside diameter of the pipe instead of your intended inside diameter of the pipe. Just a possible example, as I understand it.
Maybe I didn’t. I agree with you that using a +/- tolerance with a zero should be avoided whenever possible…I was trying to offer an example of why somebody would use a 0.0 in a +/- tolerance. There are rare situations where you might need this. Like I said, many designers use +/- incorrectly.
nope…you missed the point!
lets say I had a hole on the widget 1/8 inches +0 / -1/32
would you use a 1/8 drill? no, or at least I wouldn’t
but if the hole was called out 1/8 +/- .005 an 1/8 drill would most likely work fine.
my point is this. I think a part made to the NOMINAL dimensions called out on the print should be in the middle of the tolerance already. On a 1/8 hole dimensioned with a +0/-1/32 tolerance, if it was made to exactly 1/8, it would be too close to going over the tolerance and thus being a bad part…so…why is it done?
Those tolerances aren’t there for the machinists (in)convenience. They’re there for a reason.
Say you had precision ground shafts for your widget hole and you knew that every one was 1/8" dead nuts on. the hole you spec for them to pass through could be 1/8" or slightly larger, but not smaller. The shaft simply wouldn’t fit. So you as a engineer spec -0, +.005 or something like that to let the machinists know that 1/8" or larger is acceptable, but anything smaller than 1/8" simply won’t work.
In construction, a nominal dimension is an even dimension that makes drawing things easier. For instance, concrete masonry units are 3-5/8", 5-5/8", 7-5/8", etc. wide. But they are called 4", 6" or 8"block, and drawn that way. The missing 3/8" are essentially ignored, and things work fine as long as you dimension to only one face of the wall.
Similary, partition walls are rounded up to the nearest 1/2" to make the drawing process easier. A 2x4 stud wall with two layers of 5/8" wallboard is 4-3/4", and shown as a 5" wall.
Other things, such as 2x4s, and 2" pipe, are holdovers from when they were really that size. 2x4’s aren’t cut to 2" by 4" and planed down, they are cut to be exactly 1-1/2" by 3-1/2" because that is now the standard.
Same for pipe. The walls are thinner now, so that the same interior diameter pipe is smaller than before. You usually aren’t calculating the exact diameter required, you use a table to determine pipe sizes. Those tables are calculated using the standard interior diameters. It’s much easier to say - “Gimmee 8’ of 2” pipe" then to say “I require 8’ of pipe with an interior diameter of 1-3/4” "
The OP seems to be asking why a drawing dimension would be shown as 2.00 +.06/-.00 rather than 2.03 +/- .03, when both mean that he part dimension would have to fall in the range of 2.00 to 2.06 no matter what. The writer seems to understand tolerancing in general.
Showing a non-symmetrical tolerance means that if the manufacturing operation can consistently be held to a tighter variation than the drawing tolerance, then it’s better for the product for some reason to target the operation towards one side of the band than the center. Perhaps it’s a wall thickness that must meet some minimum for strength, but the excess weight of anything above that is undesirable but acceptable.
Also, a dimension can start as symmetrical, but be changed to non-symmetrical simply for ease of drafting without changing the nominal figures (if it’s a complex shape, for instance), if the manufacturing side can convince the engineering side that the original tolerance cannot be consistently met.
Well, yes and no. Your original question was along the lines of “why are parts sometimes dimentioned, say, 2.000 +.060/-.000 instead of 2.030 +/- .030”? Now you seem to be asking, “how close to the end of a tolerence band is too close”? Different questions.
Old machinists trick: Gullible young engineer walks into the shop, all excited. “This widget is half-a-thou undersized!” Machinist: “Really? let me see?” (takes part) “Oh about this other thing…” (stalls for time while heating up part in hand) “Anyway, let’s measure this part… looks good to me.” Engineer: “I swear it was undersized! I guess I read the mic wrong…”
Anyway, to answer your original question, you’ll often see something like 2.000 +.060/-.000 because 2.000 is the ideal dimension, the one that you’d want if you didn’t have to worry about machining tolerences. One good reason to do this is so that, at a later date, the tolerences can be easily adjusted. For example, if your widget wasn’t working quite up to snuff, and you decided to tighten tolerences to (hopefully) increase performance, it’d be a shame to accidentally tighten them to 2.030 +/-.010 when you really wanted 2.000 +.020/-.000.
However, in general, I agree that tolerace splitting (i.e., using a +/-) is preferable, if for no other reason than it makes the drawing easier to read, which cuts down on machining mistakes.
