FOS (Parallel edges as Feature of Size) ???

[dtmbiz]

Input please....

(See attached drawing)

A drawing uses a position tolerance control for centering a flat by using 2 edges identified as a Feature of Size.

Do straight line edges constitute “opposing elements” as mentioned in the 1994 version’s definition of feature of size?

 

[fsincox]

"What happens as the "V" increases in width along the length of the part? The point-to-point averaged centerplane (derived median) would still be straight, but the V would be useless."

Jim,
I think the "V" would still function fine it would just be a knife edged locator if I have it pictured right in my mind?

"Arguably, either profile FCF could reference the secondary (centerplane) datum as well, though that will inevitably bring up the self-referencing datum, and isn't worth the effort in this case."

Jim,
If I am picturing this one correctly, this is very similar to the parallelism callout we use in my thread on orientation locates?
thread1103-315082
Frank

 

[DeanD3W]

Everyone seems to be speaking of the angled sides of the slot... dtmbiz shows a drawing that applies a width (size) dimension to the face at the bottom of the slot with a position tolerance, so the OP is not about v-blocks or slots. I think it's about whether a 2D feature of size exists.

If one is dealing with a so-called "width", so a slab of material with parallel vertical sides and say, horizontal top and bottom bounding surfaces, I think we will all agree that the two fully opposed vertical faces of the slab with a directly toleranced size dimension between them form a a feature of size... Now, please make the slab a bit shorter... I think we would have no argument regarding the fact that it is still a feature of size... Keep making it shorter until it is 0.05 mm tall. It is still a feature of size, I believe. Now make it so short that its height is not measurable... Convert it to ink lines drawn on a piece of mylar... It is a 2D feature of size.

If you think that a 2D width cannot be a feature of size, then please tell me just how short a feature of size can be before it becomes too short to be considered the same type of feature. Where do you draw the line?

As with all measurement, there will be uncertainty in determining the intersections that form the boundaries of the 2D feature of size. A CMM could be used and a vision system can make this sort of measurement pretty easily.

I would probably recommend profile of a surface for the sides and also bottom face of the v-slot, but to respond to the question in the OP, I think the 2D feature of size and position of its "center line" (a 2D center plane..?) is valid.

Dean

http://www.d3w-engineering.com

 

[fsincox]

Dean,
You are correct we did get off of the OP. I did not want to comment on the FOS part originally because I am not a FOS expert (kind of guy!). I do like your concept of FOS and hope you win the day. My guess would have been NO from the conversations we have been in on positioning gage lines of cones, because it seems if the concept is valid for this part then it should work their, too. It really is the same concept, just wrapped.
Frank

 

[MechNorth]

Dean, the "datum center plane" established by Datum-B would be perfectly perpendicular to datum-A. The "2-point FOS" at the base of the inclined sides is not adequate to establish a centerplane; a center-line, yes, but not a centerplane. You need some vertical aspect to establish a plane. As to how much, that's a good question for debate; the taller the surface, the more accurate the centerplane.

Sorry, Frank; I started to follow the Spinoff2 thread as it unfolded, but didn't get involved due to time. As for a knife-edge locator, you'd need it at two locations to have a controlled centerplane, and again the centerplane would be subject to the actual surfaces ... hence, profile. A line-contact V establishes a single centerline which has planes associated with it; the rotational constraint of the plane thru the centerline is wide open though.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[fsincox]

Jim,
I figured this was only one of the (2) "V" blocks as it's length is not signifigant unless the part is equally as short, then I might agree with your concern.
Frank

 

[DeanD3W]

Jim,
I wouldn't call the feature in question "2-point", but I would call it "2D"... Since it is 2D of course there is no center plane. A position tolerance, with the tolerance zone having 2 parallel planes should be no problem to control its center line. It may be unconventional, but a 2D feature of size should be acknowledged as valid, otherwise someone needs to identify just how short a feature of size can be before it is no longer a feature of size.

We may be left without a clear definition of how to establish the center line of the feature... If it were a slot then an "unconstrained AME" would be fit to the feature to find its center plane. We would be better off with this example if Y14.5 clearly described an "unconstrained AME" for a 2D feature. How to define the mating envelope's orientation is an issue. The same can be said for a very short 3D feature of size.

Regarding a 2D feature of size, in general, how about a circle printed on a planar surface... With a size tolerance applied to, for instance, the circle's outside diameter is there a perfect form boundary at MMC? In order to accommodate the form error of the planar surface the boundary would need to a cylinder rather than a circle..? If instead we were dealing with a cylindrical feature that was 0.05 mm tall, we would probably need to get the mating envelope's orientation from the surface that the very short cylinder projected from, or from a bounding surface on the other end of the short cylindrical feature. I think any difficulties with a 2D FoS also exist for a very short conventional FoS. There is an extension of all of this into Y14.8's material... When a size tolerance is applied to the mold line of a casting, is it essentially applied to a 2D FoS? Is there a perfect form boundary at MMC? If a position tolerance is placed below that size dimension does it apply to the center line between the two mold lines, or to the center point if the mold line is circular? I don't think any standard addresses whether that position tolerance applies to a 3D element such as a center plane or axis, or to a 2D element, such as a center line or center point (Y14.8 is currently discussing this issue, and in my opinion it is the best standard to resolve the issue as it applies to tolerances applied to mold lines).

For the example in the OP, is there a perfect form boundary comprised of two parallel planes separated by .850? I say planes rather than parallel lines to acknowledge the form error of the two intersections that form the "2D FoS". None of this is absolutely and explicitly addressed in Y14.5, but for the example of the OP I'd have a hard time arguing that there were not "opposed elements" present to which a size tolerance can be applied. They can't constrain a mating envelope (other than I 2D mating envelope, possibly)... Anyone have thoughts I this?

Dean

http://www.d3w-engineering.com

 

[MechNorth]

Dean, I still don't buy the 2-D FOS at the points; technically questionably legal (IMO), but functionally not useful in any way. On the other hand, perhaps I have been thinking of this in the "old fashioned" way. Looking at '09, we have the option of using / defining an irregular FOS. Perhaps that is the way to go? Thoughts?


Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[CheckerHater]

I am uncomfortable with the idea for various reasons.
First, I am not sure about corners being perfectly sharp, so dimension is taken from "virtual sharps", that is, distance between 2 imaginary lines. In most GD&T texts "feature" refers to physical object, something you can touch.
Second, I don't see how said 2 imaginary lines establish "envelope" to embrace or to be embraced.
Third, I believe (and you may argue with that) that to be "opposing" elements somehow have to be directional, as in "opposing sides" or "opposing faces". No luck either.
So, I don't see "opposing", I don't see "feature", and sadly I don't see anybody asking "how mating part looks like?" And yet everyone has some idea on how to gage the part in question.
Let say, we CMM sides of the V and bottom flat. Then we can possibly calculate position of "virtual sharp" (I am not an expert on CMM, just assuming). Or we can go old-fashioned way and drop round pin into V, and then treat that pin as genuine FOS (works for decades for threads and gears). Both methods will require dimensioning scheme different from shown in OP.
To summarize, right now drawing looks like a puzzle and pieces are not falling together (yet).

 

[DeanD3W]

CheckerHater,
I think all of your points here are good ones.

On the other hand, all the same problems exist with very short features of size... They're not really physically capable of orienting an axis or center plane in a repeatable manner, but we can still apply a position tolerance to them.

How short can a feature of size be before it cannot be considered a valid feature of size?

If a cylindrical hole is punched in a sheet metal part, it's pretty common to see position applied, which will control the hole's location and (?)orientation(?)... Does perpendicularity get applied often to a hole in a sheet metal part? Through about half of its depth its a cylindrical feature isn't it? Why can't perpendicularity be applied to it? Maybe because the cylinder is too short, so it can't really orient an axis..? Why then can't position be applied to a circle printed in ink on a planar surface?

Dean

http://www.d3w-engineering.com

 

[CheckerHater]

Dean,
I am OK with Position applied to circle.
Will you apply Position to 2 dots printed on surface?

 

[DeanD3W]

CH,
If I applied a size tolerance to the diameter of a dot printed on a surface, I think a position tolerance could be used to control the dot's center point. I acknowledge possible issues with the orientation and form imperfections of the surface that the dot is printed on... Some explanation may need to be added in a flag note.

Since you're asking about two dots, I assume you may be referring to a dimension applied between their outer or inner tangencies (not between their centers though)... Maybe those could be considered opposed elements, but if there was a desire to control the center point between the two dots I would be very careful and at least add a flag not to explain the feature element being controlled and the tolerance zone. If the center point between the two dots was the desired entity to control I think with enough care, position might be the best choice... There are also times that a given case is so specialized that GD&T isn't the best method. I've seen this once on a an ink jet printer assembly and once with an optics application... Can my answer be "it depends"? .

Dean

http://www.d3w-engineering.com

 

[CheckerHater]

No, I mean will you consider 2 dots being 1 "feature of size"?

 

[DeanD3W]

If a slot is to be considered a feature of size, it does need a size tolerance applied between the two opposed planar surfaces... For the two dots, I wouldn't consider them to be a feature of size unless a size tolerance were somehow applied between their outer or inner tangencies... How to you want to apply the size tolerance between the two dots?

 

[CheckerHater]

Virtual sharps as in "objects not present in real part".

 

[fsincox]

Dean,
I found it. OK, should we use position for the sheet metal hole features at all, I know I want to say yes. I do see the problem with measuring perpendicularity of sheet metal holes, like Dean says, how do we make it so the shop doesn't scream? I admit in the past I have skipped using GD&T on sheet metal.
Frank

 

[Belanger]

Perpendicularity is applied to a hole in sheet metal not because we are literally concerned about the tilt of the hole. It is because (usually) that hole serves as a secondary datum feature, and we want to tie it back to the primary datum surface.
So it's really about providing a relationship between the datum features -- especially is a functional gage will be used.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[DeanD3W]

John-Paul,
Now we're back to touching on Frank's "Orientation tolerance controlling location" thread...

A tolerance on a secondary datum feature is not needed just because it's a datum feature... The relationship between the datum feature simulators is present anyway. For the sheet metal part the simulator for the hole as the secondary datum feature is required to be perpendicular to the simulator for datum feature A (for the part described above).

Dean

http://www.d3w-engineering.com

 

[fsincox]

I may owe someone an apology big time! So Dean you are saying using datums without tolerances is OK it is not really cheating at all? I really have resisted the orientation tolerances only if we are going to use a gage, stuff, because in small production run companies we don't use gages for anything.
Frank

 

[axym]

Hi All,

Here are some thoughts.

Many of the problems with applying Position or Perpendicularity tolerances to a very thin feature stem from having to find the feature's unrelated AME. In many (almost all?) applications, we really don't care about what the unrelated AME is doing - we care about what the surface is doing. So we should use the surface interpretation of Perpendicularity or Position and leave the AME out of it.

Let's look at the case of a nominally perpendicular hole in very thin sheet metal. As J-P alluded to, a typical function of this type of hole would be to act as a secondary datum feature with the sheet surface as primary. So the primary datum feature will constrain u and v rotation, and the hole only constrains x and y translation. The hole will be interacting with a pin that is exactly perpendicular to the primary datum, and probably a fixed size (if the secondary datum feature is referenced MMB). To function properly in this capacity, we need to control the hole's related AME so that it doesn't get too small to fit over the pin. In many cases the angular tilt would be huge before the related AME got too small, but we need to have a tolerance. So we're not applying a tolerance just because it'a datum feature.

It's a single hole and all we can control is orientation, so we specify Perpendicularity (not Position, right J-P). Zero at MMC is appropriate, and a functional gage could be used to check it. Again, we're controlling the orientation of the hole because we care about the surface and not the unrelated AME.

So I don't have a problem with specifying Perpendicularity on a hole in very thin sheet metal, provided that it's referenced at MMC. It may be that the process used to make the hole is so capable that the chances of the hole being tilted too much are near zero. I would say that the Perpendicularity tolerance still needs to be specified, and the person doing the inspection planning can choose to not have it inspected.

If the design is such that we actually need to find the unrelated AME of a hole in very thin material, then this is a red flag and the design must be questioned. This would occur if the Perpendicularity tolerance was referenced RFS, or if the hole was referenced as a primary datum feature at RMB.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[MechNorth]

The choice of putting an orientation control on a secondary datum fos on a thin part (metal, plastic, composite, etc.) comes down to the reality of the functionality. If the hole acts only as a locator for whatever is going to poke thru it, then the reality is that the minimal height of the hole's cylinder will not establish an accurate axis anyway ... thus perpendicularity is irrelevant. If the material is deep enough that the projected position of the mating part will be affected by an orienatation error, then you put the orientation control on.

I had a student argue this once in class. He had been taught by his college instructor that it is an absolute that the orientation of the secondary datum feature (a hole) must be controlled wrt the primary (planar surface of sheet) because otherwise the parts would not function properly. We went thru the assembly to see what each of the components did, how they interracted. The 2"dia hole had an aligned hole several inches apart, after fabrication. A rigid tube passed through both holes with a light-press fit. The entire workpiece shifted to accommodate the tube. There was no issue with the alignment that the hole could have contributed to in any way.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com