yang686526

flatness and parallelism (again)
i've seen a number of posts on the forum about combined flatness and parallelism call outs, but it seemed like the discussion was limited just to whether or not a call-out was allowed. and not so much to the physical result.
lets take two casses,
1) a plate with one side set as datum feature a and a .002 flatness applied to it. then and parallelism tolerance applied to the opposite side, also .002 (referencing datum a).
2) a plate with one side set as datum feature a and a .002 flatness applied to it. in this case the opposite side also has a .002 flatness applied to it. then a parallelism tolerance of .004 is applied between the two (referencing datum a).
(the same basic thickness applies in both cases)
does the practical outcome of these two dimensioning schemes give the same result?
i.e. a plate that varies within .002 on either side, resulting in a potential .004 thickness variation between any two points on opposite sides?
right now i'm thinking that they do, but i've thought about it too much and need a second opinion. any analysis of why or why not, would be helpful.
thanks!
check out our whitepaper library.
i agree with your thinking...
let's see what the pro's have to say.
good to know you got shoes to wear when you find the floor."fff"> - robert hunter

no, they are different.
on 2. while flat .002 to itself, the second face is only parallel to the first within .004. it could be at an angle so the variation in thickness could be .006.
i think, you've gone & got me confused now too.
kenat,
ligo,
a flatness tolerance means that the face is fixtured horizontal, and the face is tested to verify that there is not more than .002" variation in height.
i would have guessed that a parallelism tolerance means that the datum face is fixtured horizontal, and the specified face is tested to verify that there is not more than .004" of variation in height.
i am just now looking at asme y14.5-1994, section 6.6.3.1 parallelism tolerance. the definition is that there are two planes parallel to the datum, within which the centre plane of the feature must lie. the figures show features with flat or cylindrical features located within the two planes. this definition is messy if your parallelism is close to your achievable flatness, because parts of your feature are allowed to be located outside the two planes. the inspection procedure is not obvious to me.
moral of story
do not specify parallelism unless your flatness is much better than your parallelism.
jhg
i'm only learning myself, but i agree with kenat. if you think of the second face being out-of-parrallel as much as possible, the high end will be at the .004 limit. add an additional .002 out-of-flatness and you are closer to .006.
jlang17,
i would have guessed that the total variation would not exceed .004", but based on the definition, the total variation can be .006".
you can define a flatness of .01" and parallelism of .001", and you can draw a feature conforming to this. the inspector definitely will have to kill you.
jhg
thanks everyone,
it was suggested by the instructor of a gd&t class that i took last year, that practically speaking, if one has a plate that is .002 flat on both sides, then it would be impossible to actually manufacture the plate with better than .004 parallelism. which says noting about the technical correctness of a smaller parallelism, and also doesn't help me understand exactly what a .002 flatness on one side combined with a .002 parallelism would actually produce when manufactured.

i dont see it the same way. the flat tolerance is to the surface itself. the parallel does control flat but also orientation to a.
i believe the most angular tolerance is the .004 max to a and within that angle tolerance the surface must be flat within .002.
what would be wrong with the example shown on the attached jpeg? (i just used big tolerances on example so the differences can be visible.
what am i missing here?

designbiz
"quality is in the details"

here's a thought...
on your jpeg the flatness fcf is attached to the parallel fcf, so it's saying "flatness tol zone is within the constraints of parallel tol zone". but i was thinking they were separate, then would the flatness be allowed to break the parallel's constraints?
jlang17,
i interpret them as separate. in figue 6-51 on page 193 of the standard there are 2 fcf's sharing a horizontal fcf line with total runout of .02 on top and under that cylindricity .005. i see the posted jpeg as a similar callout. the parallelism is inspected to a and the flatness is to itself. these are 2 separate inspections right? i dont believe since the fcf's touch that this combines the tolerances. they are separate where the flatness refines the surface beyond its orientation to a.
designbiz
"quality is in the details"

btw people,
i do see that haste makes waste, while trying to put graphics out there to show my point of view. the tolerance zone decimal places dont match the fcf and i forgot the flat 2.0 on datum a. hopefully my point gets across though.
designbiz
"quality is in the details"