How to do true position with a negative projection

[geesamand]

I have a gearbox housing half that is bathtub shaped. In the floor of the housing half are some bearing bores and on the split (top) face are dowel pin holes where the housing meets the opposite half.

We set up a datum structure with A as the housing split face, B and C as the dowel pins. So the bearing bores are toleranced with true position according to datum A|B|C. The problem is that the bearing bores are in the bottom of the housing and according to GD&T rules, true position treats them as cylinders and projects the location onto the A datum plane. The projection error is far too great in practice to use this definition, so we instruct CMM inspection to evaluate the true position of these features as circles projected normal to A. But the drawing has no special callouts to advise this is what we're doing!

I don't want to inspect projected locations - for the positional tolerances I want to measure a cylinder's worth of each bearing bore, calculate a center point, ignore the feature's cylinder axis vector, and use (There is a separate perpendicularity control for each bore.)

I looked into projected tolerance true position callouts but since the definition appears to only support a projected location that is above the part surface, so this option will move the projection plane further from the bearing bores, not closer. (I would need a "negative" projection to counter the projection error).

Am I missing something obvious?

What kind of callout would work to control feature location without projecting along the feature axis?

Suggestions appreciated.

David

 

[Belanger]

It's tricky to visualize this one, but realize that you can have other datums mentioned on a print besides A, B, C. Maybe you want to simply add another datum "D" to the recessed surface, and mention that in the position feature control frame with the projection. Yes, the standard shows the projection always going up, but simply move the primary datum to where you want the projection to begin, and then don't forget to somehow indicate the distance of projection from there (always away from the feature being positioned, not into it).

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

http://www.gdtseminars.com

 

[geesamand]

Can I declare such a datum without the geometry being there? Such as a datum plane D that is defined to be parallel to the split face A and offset by 6 inches?

 

[Belanger]

Well, anything can be done if you add a note. And maybe in this case a verbal note under the feature control frame for position might be best.
It is possible to have a datum plane exist off of the physical part (offset by 6 inches) but I just don't know if everyone would realize that that's where the projection begins.
So maybe combine these two ideas -- show datum D with the regular triangle symbol offset from the surface and then use a note to say that "projection begins at datum D."
I'm going out on a limb with this stuff, so maybe others can jump in with suggestions?

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

http://www.gdtseminars.com

 

[pmarc]

I am just wondering. If you are saying that there is a separate perpendicularity control for each bearing bore (I assume with relation to A), and A is trully functional datum feature, shouldn't specified perpendicularity tolerance value be equal to a maximum value satisfying your projection requirement?

 

[Belanger]

Righto, pmarc. Why didn't I think of that -- just add a perpendicularity callout. You have to do a little trig to figure out the value, but I think that would avoid the awkwardness of the projection being asked for.

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

http://www.gdtseminars.com

 

[geesamand]

A fair idea. I ran the numbers and given this 2.00 dia bearing bore is 0.67 deep and 6.00 from the split face, in order to keep the projection element of the true position under .001" (that's 50% of the .002 TP tolerance), I'd need to impose a perpendicularity of .0001" over that 0.67" bore depth, and that's not manufacturable.

One option using a note is a datum plane floating in space that is defined as parallel to the split face and offset to a location at the middle of each bearing bore. Then use that as my primary datum. But that could be confusing since I'll one of these for each bearing bore.

The other note-based option is to put a note on the true positions stating that the bearing bores must be measured over the entire 2.00 dia x 0.67 deep cylinder but evaluated as a circle. This might be the best option since CMM can do this very easily and many of my vendors don't CMM these bores as more than 4-pt circles in the first place.

 

[powerhound]

geesamand,

Keep in mind that the tolerance zone only exists over the length of the bore being toleranced, not all the way back to datum A. I hope I'm understanding your post correctly; otherwise, what I just posted may not make any sense.

John Acosta, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[geesamand]

John,

Hm. I reviewed my GD&T documentation and I'm not quite sure whether the standard formally states that a cylindrical feature true position is projected onto a distant datum plane or simply evaluated as you describe.

What I do know is that CMM softare (Zeiss Calypso for example) will always use the feature axis for projection if the feature is measured as a cylinder. you have to take that cylinder-worth of data and reprocess it into a circle feature if you want Calypso to ignore that axis and project normal to the primary datum plane. I argued with them about it for a while but that's not going to change the software.

If most other CMM softwares behave like this then the literal meaning of the standard is a moot point to me. I'm just trying to convey the requirements as clearly as possible.

 

[pmarc]

geesamand,
Out of curiosity, if you had a chance to apply projected position or perpendicularity tolerance 6" high in normal way (that is, in positive direction "towards" datum plane), what would its value be?

 

[powerhound]

geesamand,

If you are working to the ASME Y 14.5 standard then the applicable paragraph is in section 1.4,(m) in the 1994 standard, in the 2009. It starts with "Unless otherwise specified" so if the tolerance is specified as projected all the way to datum plane A then disregard my post.

John Acosta, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[geesamand]

powerhound: That's very useful to know, because "unless otherwise specified" readily lends itself to mean that there is no symbol to do it but a simple note would take care of the matter.

pmarc: Speaking approximately, this 3" dia bearing bore would have a TP of 0.002" relative to the next bearing bore since the center distance must be within .002". I would put a perpendicularity of .001" on the cylindrical part of the bore and a perp on the adjacent shoulder of about .001" give or take depending on bearing type. These housings are machined using boring bars on a milling machine, so a perp of 0.0001" is not be achievable.

David

 

[powerhound]

What it says is that if you have a 1/2" deep bore that is 6" above a datum plane, and you control the perpendicularity of the bore axis WRT that datum plane, your tolerance zone is a cylinder that is 1/2" long, oriented perfectly perpendicular to the datum plane, and exists inside the bore. This is the default condition. Any projection of the zone all the way down to the datum plane would require a note to say so.

John Acosta, GDTP S-0731
Engineering Technician
Inventor 2013
Mastercam X6
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[geesamand]

Oh, well then in that case I need to get the current copy and my whole line of questioning is moot. I can then take that page and pound Zeiss tech support to answer how to execute true position to that standard.

I understood Zeiss' point that you can measure a shallowish cylinder either as a circle (no normal axis) or a true cylinder (normal axis). Zeiss takes a true cylinder and does various things that include projection and converting the best-fit cylinder to a maximum inscribed (even if it's an RFS feature).

David

 

[pmarc]

pmarc: Speaking approximately, this 3" dia bearing bore would have a TP of 0.002" relative to the next bearing bore since the center distance must be within .002". I would put a perpendicularity of .001" on the cylindrical part of the bore and a perp on the adjacent shoulder of about .001" give or take depending on bearing type. These housings are machined using boring bars on a milling machine, so a perp of 0.0001" is not be achievable.

From your description I am not sure which tolerance, position or perpendicularity, projected tolerance zone would be applied to, but this does not change a fact, that to me it seems that projected tolerance zone is not the way to satisfy your requirement. If the concern is to assure proper orientation of bores axes with relation to split face, you have to focus on the axes inside the bores and not on what is happening with them outside. After all this is how the part will function, if I understand it right - the bearing bores and the bearings inside must be as perpendicular to A as possible. If you are saying that the requirement is to control perpendicularity within dia. .0001" and this value is not manufacturable, better start looking for a supplier that is able to manufacture it or loosen the requirement.

 

[geesamand]

I wish to control the part as you describe - using the projection only along the length of the bore.

The fact that I have misunderstood the standard until now is probably misleading everyone who's been helping. Thanks for your patience.

I have found our copy of Y14.5M (it's an old 1982 version) but looking at a couple of diagrams makes it clear that 1) the projection axis is always normal to the primary datum, not the feature axis, unless 2) you're specifying a projected true position.

Additionally, on figure 153 of this edition it looks like I can specify a surface for evaluating the true position that is offset from the primary datum by declaring that offset plane (say, SURF D) and putting a callout [(TP)|(tol)|A|B|C] with "AT SURF D" below it.

I think I'll have to have another round at Calypso and see if I can determine whether it's all being done exactly to the standard.

David