【转帖】length and width datums for drafted pocke

yang686526

length and width datums for drafted pocket
folks-
imaging a rectangular pocket diecast into a surface.  the pull direction is normal to the surface.  each side of the pocket has, say, a 3deg draft so that the pocket has its greatest length and width at the surface (and has its least length and width at the bottom).
while i understand that datum targets  may be used to establish equalizing datums (per asme y14.8m-1996 castings and forgings) through the length and width centerplanes, i'm wondering if one could simply assign a datum to the pocket's length and width dimensions at the bottom of the pocket to achieve the same result.  one would probe along the corner lines (two parallel lines for each dimension) to establish the centerplanes.  is that ok?
now what if we added a fillet radius between the pocket surfaces.  can one extrapolate the fillet tangent lines to come up with the lines used to establish the length and width datums?
tunalover
check out our whitepaper library.
a probe with a ball of finite radius at its tip cannot actually touch the bottom edges of the pocket.  doesn't sound like the best datum choice to me.
mike halloran
pembroke pines, fl, usa
for draftet pockets, i prefer to dimension at a section that is a defined distance from the top surface.  this avoids measuring distortions near edges and fillets, and allows for cmm to touch actual measured surfaces, not calculating theoretical points on potentially distorded urved features.
first, you can't datum a dimension.  you can only datum a feature of size (fos) or a planar surface.  for a fos, you have to have directly (i.e. not offset) opposing points.  picture a slide vernier; if your jaws can close on opposed points, it's a fos.  if the jaws can open onto opposed points, it's a fos.  if the tail can measure the depth of a feature, it's just a depth measurement...there are no opposing points, hence no fos, and therefore can't be a datum feature.  sorry, but target datums are your best answer unless you can live with cast planar surfaces as the datum features.  if so, there you go, otherwise it's back to the datum targets.  sorry.
mechnorth-
maybe i chose my words poorly.  when the datum symbol is aligned with the dimension in the drawing and if the dimension defines two parallel planes of a feature of size(e.g. a width) then a datum is created at the centerplane of that feature.  my question has to do with the fact that a drafted feature does not have a feature of size defined by two parallel planes;  rather i'm asking if the datum can apply to the width at the bottom of the opening.
  
tunalover
regrettably, no tunalover, at least within the legitimate application of y14.5m-1994.  
if i'm understanding this, you want the centerplane of the width at the bottom of the pocket to be the datum...correct? unless the sides are parallel, no can do.  you must (that's an absolute must) have opposed points to generate a midplane datum, and guesstimating where the fillet radius is tangent with the sidewall doesn't count.  
so, you have two options that i can see.
1) use datum targets at predefined heigts (line contact would be best) as suggested by thetick, which is likely your most repeatable
2) add a small no-draft section at the bottom of the walls or as a secondary slot on the bottom surface, say 2.5mm deep.  make it shallow enough to not interfere with the intended functionality, but deep enough for your probe to contact and to be protected from damage during use.  this way you can pick it up again later if needed for refurbishment or other reasons.  i've done this a few times on machined surfaces and it worked out great at refurb time.
jim sykes, p.eng, gdtp-s
profile services
cad-documentation-gd&t-product development
mechnorth-
i was using this forum to gage public opinion on the issue.  based on phone calls i had with them a few weeks back, greg hetland and lowell foster seemed to think that the midplane can be established by establishing the side planes and bottom plane with a cmm and then using them to derive the imaginary lines at the intersections of the fillet tangent lines.  using those two imaginary lines the center plane can then be established.  i will present the problem in a cleaner way and send it to them for a second look.  i will advise the outcome.
tunalover