how to evaluate axial runout and why the same result like it as parallelism

[Carlos Gonzalez]

Hello,
Someone knows what is the best way to evaluate these axial runouts, I already have a result but it attracts a lot of attention is because when evaluating a parallelism it shows exactly the same result

thank you

[JEFMAN]

In ISO 1101, there's no example like this, about axial runout...
They only say that an axial runout (here from A only) is the same than perpendicularity.
Here, I don't see why the value would be different than a parallelism, because the tolerance zone should be exactly the same...

[UKCMM]

For any runout the primary should be a datum axis about which the part can revolve 1 full revolution. In your example this is not possible as the primary is a plane and as Jefman stated the tolerance zone will be the same as parallelism.

[Jim Poehler]

Sorry, UK. I see datum A as being tied to a diameter call out of 85. It appears to be the center of a round part. As such circular run out would be appropriate call out. We are seeing only a sectional view of the print. Am I wrong? Forget that, did not see datum B so never mind,

[DungT]

Don't know about ISO, the callout is legit per ASME. In this case, perp parallel & circular runout should be the same

[NinjaBadger]

It's legit in ISO also.

ISO 1101 - 18.15.1 Circular run-out tolerance — Radial - gives an example (admittedly for Radial runout) with a Plane-Diameter datum system.

[louisd]

Good on the engineer for defining [B][A]. [A][B] would have been a nightmare!
This runout is effectively perpendicularity because the primary datum is a parallel plane.
What is happening is the B datum plane is set/established, then a vector is projected outwards from B plane, this vector is the axis line of a theoretical cylinder that is perfectly perpendicular to B.
.
Then the CMM is locating that axis line to whatever you are using to define -A- cylinder/circle.

Circular runout would be measured by revolving the part while level to B, and located center to A. Indicator will be dragged on the respective plane and output perpendicularity to B, while origin is centered on A.

[UKCMM]

While I agree with your logic of what the designer may be trying to control why not make life easy and just use parallelism.

[louisd]

it's probably a fast rotating component, like a flexplate (flywheel for an automatic transmission vehicle), where the conical/thickness error that would be picked up with parallelism, from stamping or other processes isn't significant, but the rotational 'wobble' is.

Parallelism would over engineer the material thickness, and if it's stamped or casted, the thickness will vary. this is an effective means of ensuring that the feature is rotationally balanced relative to B and A, without considering material thickness.

[NinjaBadger]

One other point which addresses UK CMM's point above - people have stated that the result should be the same as parallelism, but that's not necessarily the case.

From my understanding - that symbol is circular runout but in an axial direction, so the tolerance zone is between two circles (of the same diameter) separated by 0.1 parallel to the axis.

i.e. the face in question could be dished say 0.25 and the part still conforms to the 0.1 runout requirement. If the designer had used total runout then it would apply to the whole surface. (which would be the same as parallelism.

You'll note that the callout is applied in two locations (with a boxed dimension showing where the tolerance applies).

[vpt.se]

^^ This!

This axial runout is a single runout, meaning that it is one revolution on the surface - not moving the indicator. If you have measured the entire surface, the evaluation will become total runout = exactly as parallelism.

So, measure a circular plane on the surface at the position indicated by the drawing, then evaluate each axial runout.

[UKCMM]

Well just to muddy the waters a little more I have dug out a very old 70's British standard that shows this call out and it states runout to be measured parallel to datum at any point.

[NinjaBadger]

UKCMM I think that's the thing though isn't it.

At any point, not at all points.

Also unless it's drawn to that very old BS then it's a moot point.

[Carlos Gonzalez]

Definitely the part to evaluate its function is to rotate and have contact with this surface with another part ... my doubt is if the parallelism is a geometric orientation tolerance and the runout is a nod or swing tolerance I understand that the runout is more complete Or that you even consider errors of form then please help me to understand what sofwares do ... what data is taken to deliver a result? By default, most of the measurements are considered under the Gauss or Minimum Squares method for the parallelism or the planes, then the tangential maxims are taken? and that because of that it is the same result as the runout?