Reality Check on TP for Embossement Keys 1

[ModulusCT]

Hi guys. This one's tough. The drawings governing the parts in question are old, and were done by people I no longer work with... To make matters worse, some of the dimensioning methods are 'interesting' to say the least. In any case, a new supplier has come to me asking for a TP tolerance of .018 instead of the .010 that we've apparently required since the beginning of time. No one bothers to figure the stuff out, they just say, "Yeah, 18 mils sounds good" and I cringe. So I asked to take a look at the tolerance analysis (none yet exists). Anyway, I'm having trouble, and if anyone can help, even superficially, I would appreciate it.

So what I've got is a cap (image gallery link below). This cap has a few bosses on it's outer perimeter that act as keys for the mating part -- a receptacle we'll call it (also below) -- So my question is this... What is the allowable TP error for the radii shown on the receptacle (scallops) taking into account the size and location error allowable for the bosses on the cap? Any ideas, suggestions or improvements are appreciated.

http://imgur.com/a/F2zf9

p.s. I'm not sure off hand what the width of these bosses are, but I believe it can be figured out by using the dimensions on the drawing. We would need to find the arc length based off of the 10°15' dimension @ the radius given with added tolerance (so @ r1.4095). Right? This is a fun question... No. No, it's not.

Mod

 

[pylfrm]

ModulusCT,

Don't you need a feature of size to apply a position tolerance? You only have about 60 degrees worth of arc there.


- pylfrm

 

[MikeHalloran]

TP = Toilet Paper?

Perhaps you meant True Position. Say so, please.

Where are Datum Features B and C ?



Mike Halloran
Pembroke Pines, FL, USA

 

[ModulusCT]

True position, yes lol. TP is common shorthand for true position, just as runout is common shorthand for circular runout. But sure, I don't mind spelling it out.

Datum C is the 2.734 diameter seen in the first image. Datum D is the base of the cap if you will, and is perpendicular to the axis of datum feature C. I don't know that runout is needed on this feature in the first place, so I chose to ignore it in this case. It was not an accidental omission.

pylfrm - Not sure I follow you... The radius is being controlled, right? It's position is variable... Maybe I'm not understanding what you're saying.

Anyway, thanks for the help fellahs.

Mod

EDIT: I apologize if the question isn't clear enough. Let me know what other information is needed if you can. The true position of the receptacle scallops is easy, it's the size and location variation of the bosses, defined by degrees that's giving me such trouble I guess.

 

[ModulusCT]

Could anyone just suggest a first step in solving this problem? I know a bit about GD&T, but this particular task is just a little too complex for me to tackle at the moment. Any help would be appreciated.

EDIT: So the arc length of the boss (including the fillets) is .2522" -- and that includes the quarter-degree size variation. If I add the positional tolerance of the boss (±15' so .5°) into the mix, then we have an effective arc length of .2644 (for 10.75°). OK, that's a step in the right direction I think.

 

[MikeHalloran]

Draw each piece to a large scale on tracing paper, and overlay them.
Draw the actual tolerance zones in soft pencil, and adjust things until you are happy.

Sure, you can do it on CAD, but sometimes paper is quicker.

Mike Halloran
Pembroke Pines, FL, USA

 

[ModulusCT]

heh - that's an idea....

Thanks.

 

[pylfrm]

ModulusCT,

First I should note that I am assuming ASME Y14.5-2009 here.

A positional tolerance like you have shown creates a cylindrical zone within which the axis of the unrelated actual mating envelope (UAME) of the feature must lie. You have less than 180 degrees of arc on your radii, so they have no opposed elements. I don't think the UAME is a meaningful concept without opposed elements to expand within or contract around, so this would be an invalid application for a positional tolerance.

I'd suggest you control the geometry with basic dimensions and profile of a surface tolerance(s) instead.

As for how to figure out the proper limits, draw your bosses at all allowable extremes of size and location and then superimpose them. Tolerance the mating part to keep it out of that area.

I'd argue that tolerancing the location of the center of a radius is usually bad practice no matter how you do it, so my profile of a surface suggestion goes for the bosses as well. Also, tolerance analysis becomes much easier if both parts are controlled with profile of a surface.


- pylfrm

 

[ModulusCT]

OK Thanks, that's great advice.

Couple questions though. First, isn't it possible to inspect radii like the one's in the receiving part with gage pins? If you have a .25 gage pin inserted into a .25 radius, wouldn't that locate the center? I guess that gage pin would be good only for checking for size variation?

I work at a GD&T illiterate company. This same company uses GD&T. I know a lot more about Y14.5-2009 and '94 than anyone else here (and I'm learning more all the time), however, it's difficult convincing anyone to do things differently because they're all scared of making a mistake. Many wonder, why change it when the way it was has been working all this time (no mention of how well it's been working). It's very frustrating.

So, I guess my last question is this... The boss is NOT controlled with true position. Only the radii in the receiving part. While it may not be optimal, is there any value to the current setup? Or, rather, how would I work around the current system to ensure 100% interchangeability? I suppose I would need to know the virtual condition of the scallops in the receiving part, the maximum size and position error of the bosses and then just ensure that the scallops are oversized to receive those bosses. I'm seeing the issue I think.

I'll try making a drawing over the weekend.

 

[3DDave]

Unless a feature surrounds a point or a plane, it's not a feature of size and cannot have a position, and therefore cannot be controlled by a position tolerance. You can see this in the limiting case where the radius is infinite and the resulting surface is flat. There is no support to locate a flat surface with a position tolerance. Instead, there is a profile tolerance, which can locate one-sided surfaces.

To see if position is applicable, attach the pin to a fixture and try to move the part. If there is movement the pin does not limit, then it is not a feature that can be controlled by position tolerance, at least not along that direction. To be sure this is the case, use a slightly larger or smaller pin and confirm the motion allowed changes by exactly the amount of change in the pin. Note that the position limit is applied to all the surfaces that are required to limit motion of the pin, not to individual boundary components.

 

[ModulusCT]

That's a great mental exercise to perform 3DDave, to determine whether or not a feature is an actual feature of size. I will write that one down, thanks.