GD&T for bolted flange of two rotating parts

[New to this]

Hi all,

I'm designing a shaft that has two halves that connect via bolted flange. I want to ensure that they're concentric, so I added a centering lip on one half to locate via LT1 fit in the other half. I'm concerned about the alignment of these two parts, so I started to look to GD&T to achieve some of my objectives. For example, I don't want the centering lip to cause the bolt patterns to be misaligned. However, I've never used GD&T, only basic dimensioning, or even taken a class... so I'm pretty lost.

Objectives:
1. The mating faces of the flanges should be flat/parallel enough to each other so that the o-ring between them is compressed and seals properly. The groove for this o-ring has a height tolerance of ±.002'' (pic attached). So I think the control I add to achieve this needs to fall within that spec. I understand flatness will not ensure the two faces mate properly, because there is no datum for that control.

2. Once the two pieces are partially assembled via a hand press fit on the centering lip, I would like for the 6XM6 clearance holes on each flange to line up so that the bolts slip through and a nut can go on the other side of the flanges. I figure the centerline of each bolt hole in the pattern needs to be parallel to the respective half's centerline. I'm not sure this is the right way to go.

3. The shaft is around 1' long. So when both halves are assembled and being installed, I need one end to insert in an automotive transmission and the other end into the engine's crankshaft. I'm concerned if I don't put position tolerances on the shaft centers as the diameter varies, we may have trouble mating the engine to the transmission if the shaft centers are misaligned enough.

I can update my drawings to have GD&T but before I learn it, I want to confirm that it's the right tool for the job and that I'm thinking about it correctly first. So drawings without GD&T and just basic dimensioning are attached. Please excuse my notes, this is a draft. This part is a one-off part. Potentially to be produced in higher volumes later. If I'm over engineering by using GD&T at this phase, there's a subset of what I've said here that's useful or there's something I'm not considering let me know! Thanks.

 

[New to this]

Update regarding the location of the bolt holes to ensure the two parts bolt together properly:

I added datums and a positional tolerance to the bolt holes using the Floating Fastener Formula in addition to the standard drilled hole tolerances for M6 bolts. Please ignore the positional tolerances and flatness on the left side view, I haven't specified those yet.

Hole MMC - Fastener MMC = Clearance
(.252" - .001") - (.262") = .011"

The same control was applied to the mating part. This should ensure that all 6 bolts are able to go through without having to waller them out, right? And since the A datum is the centerline of the centering lip, these two features (bolt pattern and centering lip) should work together to align the part rather than interfere and cause one to work and not the other? Meaning if the centering lip is pressed into the mating part, the bolts should be able to go through all 6 holes, right?

 

[3DDave]

Typically the flat face is the primary to orient the datum reference frame and the pilot is secondary to refine the location of the axis. Specifically, using [B|A] which also matches their definition.

 

[Burunduk]

Since this is a press-fit, once the centering lip and the bore are mated, the mutual location and orientation (excluding clocking) is set between the parts. For a press fit, the centering lip (and the bore on the mating part) being the primary datum feature A and the face being the secondary datum feature B for the positional tolerance of the holes makes sense. Datum feature B which is used as secondary should have a perpendicularity tolerance relative to datum feature A, and a flatness tolerance can be considered for it only as a refinement of form, if the flatness error should be smaller than the perpendicularity tolerance (the perp. tol. limits both the form and orientation). Since the diameters constrain location and orientation, the perpendicularity (and flatness) allowance on the flat faces of the flange will influence the sealing quality; there may be a gap resulting between the mating faces, at least at free state. But the only way to eliminate that influence would be to sufficiently enlarge the clearance between the centering lip and the bore, and thus allow the two faces to always mate flush with each other. This would also make them the proper candidates for the primary datum features, but it would come at the expense of the coaxial location alignment between the two components of the shaft. There is always a trade-off between sealing and location in such cases. Possibly, the clamping forces exerted by the six M6 bolts will eliminate the gap, and ensure sealing, but the tolerances of perpendicularity (and possibly the flatness refinement, if needed) applied to the flat faces should guarantee that the gap will be closeable or sufficiently minimized by the clamping forces (in other words, the tolerances should ensure that the gap is small enough to begin with).

 

[New to this]

Thanks for the replies guys. I learned some things! I have read over your responses a couple times and it seems I was at least headed in the right direction.. It seems I could improve what I had above to be more appropriate and consider some of the tradeoffs Burunduk was mentioning.

I spoke to the shop that will make these parts and they don't require GD&T so for this one-off I will just call out my assembly requirements in notes. However, if this part makes it to higher production volumes, I'd like to get GD&T training to make sure the parts fit how we expect them to.

 

[3DDave]

Anyone who has ever pushed an untapered dowel pin into a hole with only 10% of the diameter engaged will know that the dowel pin won't be guided straight by the hole and at that point the interference doesn't control the orientation of the pin.

If the fit was long enough to provide positive orientation (50%+ of the diameter) there would be no need for a second datum for the holes to be perpendicular to. They could simply be parallel to the primary datum feature. However, as is obvious, determining the orientation from such a short feature of such large diameter is very difficult. That same applies to trying to make a large surface perpendicular to a similar short feature.

Imagine grabbing that narrow feature with a wrench as if it's the edge of a washer and try to apply a bending force to it. The leverage will be tiny and the wrench would slip off. It won't be establishing orientation.

Which is why they don't show the pilot as primary as the example in the standard when combining a stub pilot diameter and a flange.

By making the flange the primary the pilot can be very short so it cannot control orientation; only it's location matters. The features on the part are then oriented to the flange and located to the pilot.

Make that pilot 2 inches long and Burunduk would still be wrong in the positioning of the holes - the pilot would be sufficient alone.

Suppose that an ideal pilot - perfectly cylindrical, and an ideal flange - perfectly flat were made a sloppy 1 degree off.

At .197 inches extension, the sin(1 degree) would make a change of .003 inches apparent diameter. That's on a 2 inch feature, or a 50ksi contact load if it is steel. The drawing already allows twice that. But what do typical machining centers hold? Typically around 0.1 degree; about 5ksi would result. The best part - the same is true if one assumes the pilot controls orientation.

Depending on the material the interference will either yield a little or just flex. In either case the pilot of this part will not control orientation, the flange will.

 

[Burunduk]

3DDave has a good point.
I wasn't paying enough attention to the length of the pilot diameter.
As a primary cylindrical datum feature, the interaction between it and its datum feature simulator would be required to constrain 4 degrees of freedom: two translations in directions normal to the axis and two rotations about the two translational directions that are constrained. Because of the short length of the pilot, it would be problematic to constrain the two rotational degrees of freedom by using this datum feature. The flange face as a primary datum feature would constrain these two rotational DOF without problem.

Possibly the fullest possible face contact can be obtained in the functional assembly due to the ability of the flange to flex as a result of the clamping forces - this is reflected in the drawing by selecting the flange face as the primary datum feature. The pilot would need to have a perpendicularity tolerance relative to the flange face - as shown in the second post (although the diameter symbol is missing for the perpendicularity tolerance, and the order in the feature control frame for the positional tolerance should be reversed).

It is worth noting that unlike it is usually described, the outer boundary resulting from the pilot's MMC size and the perpendicularity tolerance will not act in this case as the worst case diameter the mating part "sees" for fit, because the pilot is mated first and only after the face clamping is done the mutual orientation is set. What the perpendicularity tolerance limitation should insure is the ability to flex the part into the fullest possible face contact (so that the flange face becomes the orienting feature) without stressing the parts more than allowable.