How to define limits on hole position and diameter as worst case orientation of mating part

[TopPocket]

Hi, I've been trying to figure out how to correctly define a hole in relation to a face but all the literature seems to be for fixed gauge pins. This isn't what I want though, I'll try explain...

The hole is perpendicular to the face (which is our datum feature).
The hole is there to guide a rod, so I need it to be: a) perpendicular to the face and b) a good fit on the diameter. The rod can be thought as loose, with the hole being the only thing touching it.

I want the worst case angular offset between the rod axis and the axis perpendicular to the face to be my limiting condition.

Now the tolerances on this is very tight so I would like to do something similar to a MMC between the hole diameter and the perpendicularity between the hole and face.
So a bad perp could be compensated for by a tight diameter and vice versa.
However MMC seems to be based around the RAME, where I would like to base it around the... well I don't know if there's a term for it, the opposite of the RAME?

(by RAME I mean the largest diameter cylinder that is perpendicular to datum A that will fit in the hole - see below)
I want a fixed diameter cylinder that sits in the hole in the position with the worst possible perpendicularity.

I have a way of inspecting this using the CMM and some excel code but I'm struggling to correctly define it on the drawing.
I'm considering using position instead of perp on the hole (see image below). Or maybe runout is what I want... though you can't use MMC with it, tho I can imagine an offset but straight hole failing this, maybe a position to control the axis position, and a runout to that axis (and perp to A)...

Here's a rough drawing with the basic set up (pay no heed to the values)

Maybe it's not possible and I'll have to just stick a note on trying to explain it the best I can with the tols opened enough to allow individual features to pass in all scenarios with the ultimate bottle neck being the formula result. Although I feel this isn't a particularly niche scenario and there should be a way to define this.

If you need any clarity let me know. Thanks

 

[Burunduk]

TopPocket,
The skewed fixed size gage you show on the left side of your first image doesn't limit the worst case angular variation the way you think it does. Imagine the actual hole staying the same size that you show it, but rotating further to the right until the left side of the hole in the shown section touches the gage on the top end and the right side touches at the bottom end. This will allow the rod even more out-of perpendicularity.

What you care about is the maximum looseness between the rod and the hole. Think perpendicularity at LMC.

 

[TopPocket]

Hi Burunduk, thanks for taking your time to help me.

I follow what you say and it does make sense (a very subtle observation).

However that isn't quite what I'm after - an understandable misunderstanding, my dodgy drawings and rambling scrawl no doubt being the cause.

The angle I'm concerned with is the one between the rod axis (let's consider the rod to be a perfect cylinder) and the axis normal to Datum A, not the axis of the hole.

Best to think of it like a tolerance stack up.

The angular offset is a combination of the differnce between the rod and hole diameters plus the perpendicularity of the hole to Datum A.

Note: due to the small angles the rod Ø can just be subtracted from the hole Ø to get the "gap" (ΔØ) without having to calculate the minute differences due to their misalignment.

 

[TopPocket]

Maybe instead of the angle being the limit the difference between the position of the start and exit of the hole can be my limit?

Same thing just different value to try and define... bit weird tho.

 

[Burunduk]

TopPocket,
Perhaps I did misunderstand you.
You were describing some concept you have in mind that has something in common with MMC and the bonus tolerance it provides, but behaves "the opposite of the RAME". The RAME is shown on the right side of your first image, so I thought that the boundary gage related to the concept you were describing was on the left. But apparently I was wrong, and you were only showing how the actual rod would behave in the worst assembly scenario for orientation?

Anyway, you also mentioned that the tolerances and the fit would be tight, so the common way to go about such cases is specifying everything regardless feature size, thus not invoking any fixed size boundaries and not allowing any tolerance exchange between the size and orientation/position. However, the way you described your design intent still makes me think in the direction of perpendicularity at LMC. Did you consider it?

 

[TopPocket]

Burunduk,

Yes, you might be right about LMC being a part of this solution.
I've been thinking about the problem and have narrowed down exactly what it is I'm after:

There is a minimum diameter requirement on the hole.
The hole must reside within a defined boundary: a perfect cylinder that is perpendicular to Datum A. The diameter of this boundary is set by my specs and also gives me the maximum hole diameter.
The position of the boundary cylinder is controlled by the hole itself. (Position relative to Datum B will be a different unrelated spec*)

The left image in following drawing is what defines our boundary. The limits in the right image stem from this.

So maybe something like this:

Looks a bit silly but I think it's what I want.

*the position related to Datum B (2nd to A in the frame) has a much lesser impact due to it only introducing a radial offset. The angular offset is multiplied as the rod is projected so is more critical. To control both together would require knowing the orientation of the angular offset relative to the radial offset and is probably going too deep.

 

[Burunduk]

TopPocket,
What you describe in your latest post is exactly the type of boundary that would be created by a perpendicularity tolerance specified at LMC. To be precise, it is the "surface interpretation" of it according to the ASME Y14.5 standard. This boundary will reside inside the material of a conforming part, and its size will be set by the LMC (maximum) size limit of the hole, plus the perpendicularity tolerance specified in the feature control frame.

The position of the boundary cylinder is controlled by the hole itself

I get what you are saying here, but I would put it a bit differently. The boundary is free to float and is only constrained in orientation with reference to datum A. Its "position" is not "controlled" by the hole. The boundary controls the orientation of the hole, or more precisely, it controls the collective effect of the hole's maximum size and orientation error, without affecting the hole's location. But this is probably just me being pedantic about the choice of words.

Note that everything so far follows the way you described your design intent, and the way you showed it in your posted images. More specifically, you seem to be concerned about the "temporary" orientation error that might be caused by looseness between the hole and the rod's diameters. I say "temporary" because when assembly looseness makes one of the components skewed, it can usually be adjusted manually. For example, the looseness could be used to manipulate the rod position so that it can fit into a hole in another part in a portion that protrudes from your part. You mentioned that "The angular offset is multiplied as the rod is projected" which is correct and many times it's a concern for tight-fitting components such as pins and studs, that are protruding from part #1 into which they assemble with a tight fit, and they must not interrupt with a larger hole in part #2 into which they assemble more loosely. In this case, the increase in looseness with part #1 is not a problem for the ability to assemble, on the contrary. I got the impression that this is not the type of application that you are dealing with... but in case that it is, you need a completely different approach. You may want to look up the "projected tolerance zone" and the "fixed fastener formula" in the ASME Y14.5 standard or in the literature you use.

 

[TopPocket]

I must apologise, I've been getting my standards mixed up. I should be prescribing to BS 8888 (ASME Y14.5 is just so much better explained)

Hopefully it's basically the same: a Perpendicularity of Ø0 with a LMR to datum A. Plus a Position to Datum A and B. And maybe the projected tolerance zone.

You're right, it's a little unusual. The hole is providing guidance for a rod that is to be inserted into a "squooshy material" and I need to make sure the rod doesn't stray too far off trajectory. So I need to design it for the worst case.

I've slapped on the three constraints here, not sure if I need all three or just the 1st and 2nd or 1st and 3rd or just the 3rd...?

 

[Burunduk]

TopPocket,
I now need to add a disclaimer that all my explanations were based on the ASME Y14.5 standard, and the meaning of the callouts might slightly differ per BS 8888, including what we discussed on your other thread on how the datums are derived. If I'm not mistaken, BS 8888 is essentially the set of ISO GPS standards concentrated in one document. Isn't it? I'm less familiar with the GPS standards but, I know they differ from ASME in the way surfaces are treated and simulation of theoretical elements takes place; they deal a lot with "average" features derived from the physical ones and not so much with minimum circumscribed, maximum inscribed, etc. I'm also not sure that per ISO/BS what we call datum feature simulators/ true geometric counterparts need to be at basic orientation (and location) to each other - so all this might affect the relevance of the answers you received in the other thread. There are forum members here who are very knowledgeable in ISO and can shed some light on this.

As for your latest version of the drawing in this thread - you definitely don't need all three requirements. You should specify either just the projected position tolerance or the combination of LMC perpendicularity and unmodified position. 

In your initial post you mentioned that one of the requirements for the application is "a good fit on the diameter", but if the values you show here are representative of what you want to assign - the tolerance on the hole size is pretty generous and this might lead to a loose fit.

Be careful about the tolerance values that you assign. In the case of the LMC perp. + unmodified position combo, you don't want one requirement to make most of the other requirement's  tolerance unavailable, as it is with the up to 0.2 perp. (as a "bonus") tolerance combined with the 0.05 position for the hole axis. The position tolerance also controls the maximum orientation error.  I thought that you cared much more about the type of loose/temporary orientation error that you were describing than for the location of the hole, so assumed the position tolerance would be much more generous and definitely looser than the perp. tolerance.

As I already mentioned, In most cases where some sort of a shaft is rigidly assembled into one part and must stay clear of a hole on the mating part, position or perpendicularity with the projected tol. zone modifier is your friend. It creates a worst case boundary outside of the toleranced part that your rod will not violate, and this boundary size equals the rod's MMC size plus the projected tol. value. Consider whether this may be the more suitable specification for your application.

Have a design review with your colleagues on what approach is more suitable for your application.

 

[drawoh]

TopPocket,

When you specify a positional tolerance, you are controlling the diameter and the angle of a hole or a projecting feature. When you specify a feature of size (FOS) datum, the size tolerance accounts for perpendicularity. Read up on projected tolerances[ ]([Ⓟ]). This might help you wrap your head around all this.

--
JHG