ImnotfromMars,
Profile of a Surface would be a better choice than Position.
Position applied to a cone is not supported by ASME Y14.5. That said, it is common to apply a size tolerance that applies to one end or the other of a drafted cylindrical feature, so a cone, on a cast, forged, or molded part, using a method from ASME Y14.8, with a Position tolerance feature control frame placed below that "2D" size tolerance. No ASME standard currently clarifies the question of whether the position tolerance applies to only the center point of the 2D size tolerance, or to the entire axis within the conical feature... The cone's axis could be derived using the cone's unrelated actual mating envelope, but Y14.5 provides no explicit support of this approach. A note could be added to clarify what the Position tolerance applies to, but why not just use Profile of a Surface instead?
Is there a particular reason you want to apply Position in this case?
According to ASME Y14.5-2009, position tolerance may be applied to "one or more features of size relative to one another or to one or more datums" (para. 7.2).
So the obvious question is whether a cone qualifies as a "feature of size." That goes back to paragraph 1.3.32, where the new standard talks about "regular" and "irregular" features of size. A cone is definitely not a regular FOS, but you can make a case that it is an irregular FOS according to paragraph 1.3.32.2(b).
So while I've never seen position used on a cone in a real-world drawing, I guess it's possible. The closest example I can find in the standard is Fig. 8-24 of the 2009 standard (or Fig. 6-19 in the 1994 edition), which isn't a cone, but at least it's an irregular shape. Also notice that profile is shown there too. This would probably be the best way to handle the cone (position for location of the boundary, profile for shape/size). Of course, you can use profile as the only geometric tolerance on the cone and have it control location also.
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
Are they coaxial? Why not just use circular runout? As mentioned above, profile also works, however runout would give you the freedom to easily separate the size requirement from the position requirement.
I'd argue you can, however a lot of folks argue you can't so if there is another robust way of dimensioning it for your application, you may be better choosing that technique.
I believe there were also a few others, but at the moment I can't recall exactly.
I also feel that there is a huge gap in Y14.5 about dealing with location of conical features. First of all no clear, unambiguous indication how to treat cones - as features of size or not. Secondly not a single example (correct me if I am wrong) showing a method of locating a tapered pin from other datums (for instance from three mutually perpendicular datum planes).
Y14.8 gives some methods of tolerancing drafted features themself but also does not give many hints on correct method of locating them relative to other datums. Profile of surface is generally recommended method, especially for features vital for part's function.
I tend to accept profile of a surface as a solution, however there is one, IMO, major disadvantage of profile tolerance comparing to positional tolerance and it is a lack of possibility of applying (M) or (L) modifiers to the geometrical tolerance value. Therefore a combination of profile and position of feature's boundary, as J-P suggested, sounds really interesting.
On the other hand at my daily practice (plastic molded parts for automotive industry) I often deal with a method described by Dean - position tolerance to "2D" size tolerance. This has some advantages, however as long as no global standard clarifies "the question of whether the position tolerance applies to only the center point of the 2D size tolerance, or to the entire axis within the conical feature", I see a need for a really smart internal dimensioning policy that would explain the problem.
Dean,
The reason is MMC, the ability to have a greater location tolerance as the size increases. This is as old as the reason for GD&T. Tapered features are used in other applications than just to compensate for tolerance adjustment as in the GD&T books.
Frank
Frank,
While it is common to treat Position at MMC or LMC as an axis/center plane/center point control with a potential bonus tolerance, it might be equally common to treat such a tolerance spec as a surface control with a virtual condition boundary (as I think you are well aware of, based upon some of your other posts). If Position at (M) or (L) really becomes a surface control, then I think it would be better to apply Profile. This becomes debatable, I readily admit, given the current methods of applying profile...
I think one of the things that is most in need of a change in the language is this dual interpretation for Position... I sometimes say there is the common interpretation (axis), then the one that overrules it (surface). I think the situation would be much more straightforward if Position at RFS were always an axis/center plane/center point control, and if there was a desire to impose a tolerance that would ultimately be a surface control {Position at (M) or (L)}, then Profile of Surface, with some enhanced methods to allow exactly what is desired for the +material and the -material boundaries, would instead be applied. Such a change will require a pretty significant Y14.5 revision, but I think we should have only one interpretation for any tolerance spec.
Since Position at (M) or (L) should always be specified with a zero tolerance value (with all the functionally acceptable tolerance moved to the size tolerance), the requirement boils down to is the virtual condition boundary and the local size requirement portion of the size tolerance. With the approach that I would like to propose each of these surface controls would be provided by a slightly revised approach with Profile.
Another thing I'd like to mention is that a cone with a very small included angle may function as a feature of size, since it is very nearly a cylinder, but a cone with a very large included angle will not. With a very large included angle, 178 degrees for instance, the feature will function more like a planar surface than as a feature of size (that behavior being the ability to constrain a mating envelope, which a nearly planar surface really can't do, in practical sense). So, just as with partially opposed planar surfaces, where the line is drawn between these features being "of size", or not has be be drawn by the designer. For a conical feature, or wedge-shaped (wedgical? :^)) feature, this depends upon the included angle, and with partially overlapping parallel planar surfaces this depends upon the amount of overlap. I don't see these two cases of "designer decides how the feature functions, so therefore how the feature should be toleranced" as likely to be discretely defined by any standard.
While I agree with your ideas about axis interpretation vs. boundary interpretation for position, there is still a significant difference when you jump over to profile. I guess this is why you say that trying to equate the two would involve a big change in the standard, because profile inherently covers form.
Profile also requires a basic dimension, which is why I have such a beef with Figs. 8-17 and 18 in the 2009 standard!
I'd also add that position with MMC/LMC can usually be specified with a zero tolerance, but I wouldn't say "always." There are times when that might not work due to weight/mass limits. That's probably a can of worms for another thread, but I couldn't help mentioning it.
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
Hi Jean-Paul,
Yes, when I typed in the word "always" in the post above I kind of thought that might generate some controversy :^). I agree that there can be other considerations that would make zero at MMC not the best Position at MMC choice. This won't change my position regarding the surface interpretation of Position... I will still propose that if a surface is to be controlled then it would be done with profile.
I also agree that figures 8-17 & 8-18 in Y14.5-2009 are less than wonderful.
Dean,
I am wondering if your concerns may shared by others on the committee and one of the reasons they have now introduced the concept of material modifiers for surface datum references.
I know this, if I have a mixed flow impeller (conical shaped fan) on bearings inside a shroud (containing hub) and either part deviate from their MMC there is more clearance in the assembly. The fact we have made it is hard to describe clearly using our standard is NOT the fault of the assembly itself. The ISO has tools to do it and has had for a long very long time, it may not be perfest but few things are really inspected "perfectly" with gages.
While we wait to find the perfect solution the manufacturing base of the country will be gone, I will be retired, and the world will be using the ISO standard derived from the concepts we ourselves put inplace but were afraid to use.
Frank
Frank,
"Material boundary modifiers" (material condition modifiers applied to datum features) were implemented in Y14.5-2009 to provide a means of specifying datum feature simulator behavior/characteristics especially for the case of planar secondary or tertiary datum features that overlap (can be located relative to) the DRF established using higher precedence datum features... This was termed the "Tertiary datum problem", but the issue of a secondary that can be located relative to the primary datum feature is the same situation.
My concern is the confusion and expense that can be caused when a tolerance has a dual interpretation. While I think ASME Y14 standards are quite a bit better than the similar ISO standards, I do think Y14.5 should be changed such that Position tolerances have only one interpretation, so no longer both an "axis" interpretation and a "surface" interpretation. This is especially a concern when MMC or LMC is applied after a Position tolerance value.
I'm interested in the application you describe... I can't quite picture exactly what the issue is. Is impeller-to-shroud clearance the main consideration? How does an ISO standard address this in a better way than the methods in Y14.5 provide? If you can provide more information, even if you think it should be in a new thread, that would be great.
Dean,
The impeller itself looks very similar in geometry to one posted by another in another thread:
thread1103-247655
His was used for mixing, more of a driving screw I imagine, my applications are fans (air flow). These are spinning at a high rate of speed, so ideally I would like it to be concenteric (yes, the dirty word), but since it is balanced as a separate operation I feel it can be allowed MMC. The shsahft is supported on ball bearings (datum A-B).
Frank
Frank,
For the impeller, I assume A and B would be simulated RFS (RMB per 2009 terminology), since they're in bearings. You then need a boundary around the perimeter of the impeller, which I wish could be called an MMB, but to use proper terminology, I have to say virtual condition... Since the outer rotational clearance boundary appears to be affected by both the blade faces and also the "edge faces" it seems that Profile of a Surface could be applied to all surfaces. I can see that that's likely more inspection and less direct control of clearance than you would prefer. If the revenue for the product can cover the cost, maybe a functional gage could be constructed and spherical tooling balls used to check clearance with the functional gage. This would take a note to explain, of course.
Position is the tolerance normally termed to be associated with a virtual condition boundary, but profile's "MMB" (if I can be allowed to mis-use that term here) is really the same as a surface interpretation of Position. The difference between Size + Position and Profile of a Surface comes for the "LMB" (Resultant condition), which is something I hope can be specified as Profile with more "LMB" flexibility in a future revision of the standard.
I don't know for sure that I am picturing the functional need well enough, and I still wonder about the more effective method an ISO standard provides you referred to.
Dean,
I am not sure I said more effective, it is just that it is an accepted practice, not scorned as it is here. They can use position or coaxiality (Concentricity Symbol) with an MMC modifier.
My position has always been not that it is better or more thought out but it was availible as a tool for those that need it, the position taken here is more often if it is not all perfectly worked out it is not ready yet, so I do not have the tool I need to express my real world day to day actual design condition.
I was taught GD&T by a gentleman, now on the committee, who always said: "if it is not prohibited in the standard it is OK, you may need to add a note of explaination, but it is not illegal". I believe this and it seems more like what the ISO is doing.
Here, No offence intended to anyone in particular, the attitude seems more, if it is not already shown in "the book" so that it is "clear" then it is not allowed. All the while they still debate the illustrations that have been in "the book" for 30 years.
We need the tools today, not 20 years from now (10 for the next standard, 10 more till people really use it, my company is still at the 1982 version).
Frank
Frank, I agree with your philosophy: just because it isn't in the standard doesn't make it wrong. I am constantly telling folks in my training that GD&T is a language; like any language you may sometimes have to twist things here and there to accommodate a special need. That's OK as long as it isn't prohibited. And if it helps, textual notes are certainly allowed to help nudge the reader.
John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems
