Alignment Tolerance Crankshaft to Automatic Transmission

[AllorNothing]

When designing a transmission adaptor plate for an Automatic Transmission 4r70w or A40DE or e40d to another engine, assuming that the machined surfaces are parallel withing the oem's factory tolerance, how much misalignemnt shaft center to shaft center is practial, tolerable, and typical?

Are we aiming for .010, .050, or .002 here?

There certainly appears to be some "slop" at the torque converter to transmission input shaft. Flexplates don't run terribly true either.

 

[MikeHalloran]

The flexplate allows for some angular misalignment, but not radial misalignment. You are aiming for .002". That's what the dowels are for.



Mike Halloran
Pembroke Pines, FL, USA

 

[Fabrico]

In the case of transmissions to crankshafts, radial misalignment is angular misalignment. Both ends of the rotating component are always located radially, but each is free to wobble. Flexplates do nothing for alignment.

It might be difficult to find a factory spec on total misalignment, because no one questions transmissions.

Many truck component makers call for a maximum crank/bellhousing runout of .008".
Race transmission and clutch makers call for maximum of .005".

 

[PaulJackson]

I agree with Mike.

Ideally the point where the axis of the crankshaft is projected to pierce the plane of flex-plate attachment that is parallel to the rear face of the block ---likewise--- the point where the axis of the converter stator support is projected to pierce the plane of flex-plate attachment that is parallel to the front face of the transmission---must meet within some functional and manufacturably feasible tolerance, the-smaller-the-better.

Each dowel pattern (engine/transmission) establishes that axis-to-axis intersection (commonly at the plane where the mounting surfaces meet rather than at the plane where the converter attaches to the flexplate however small or large that difference may be). The dowel pattern forms a triangle whose spread is bigger or smaller than the specified spread but whose proportional center of the horizontal displacement and vertical displacement should be identical and coincident with the respective crank and stator axes. The mounting surfaces are never perfectly perpendicular to their respective component's axis but that is what the flex plate is for as Mike stated. The GD&T callout controlling the location of the dowel pattern position to the axis should always describe the axis as the primary datum feature rather than the mounting surface because the mounting surface perpendicularity (if declared as the primary) causes the intersection point to be projected from the midpoint of the axis rather than at the mounting plane. The spread can always be toleranced separately if a refinement can be achieved for dowel fit. Typically although, an aluminum bell housing can flex and tolerate some interference fit acknowledging that the spread varies some due to temperature. The design tolerance stacks would have to reflect that interference for spread but simultaneously reflect the proportional center of the horizontal displacement for the pattern to axis displacement.

In an automatic transmission the alignment and clearances of converter-hub-bushing to the stator or pump journal, reactor-to-impeller gap, reactor-to-turbine gap are all dependent on the intersection of those two axis.

In a manual transmission the input shaft bearings takes the brunt of the horizontal proportion or vertical misalignment.

 

[Fabrico]

I agree with Mike too. But, unless the pilot is worn, the torque converter cannot be radially misaligned with the crankshaft. Radial misalignment of the transmission equates to angular misalignment of the torque converter. Therefore, the flexplate does serve to address radial misalignment of the transmission.