Radial runout problem

[[Di...]]

Hi all,

I'm having a problem with radial runout inspection on a part that has primary datum on a face and secondary on a cone. Radial runout has to be inspected on a hole on the opposite side of part with respect to face and cone.

I have measured plane with two circular scans (it is circular plane) and then I have measured cone with 3 circular scans. Then I created a point recalled from cone and with that point and plane I created my alignment. Then I measure circle for runout and that is where problems start. I get runout 5 times bigger than it really is because we already measure runouts with a different method not involving cmm. Form and sigma of elements are okay, but the position of actual circle isn't. It is 0.005 mm offset from the center axis, which I suspect is the root of bad runout result.

My question is what causes measured circle to be offset like this when in reality it isn't?

Thanks!

[[Ri...]]

There are users on here infinitely wiser than I, but I'll venture a guess that you have an unstable base alignment.

[[Ri...]]

There are users on here infinitely wiser than I, but I'll venture a guess that you have an unstable base alignment. The part where you created a point based on your cone and then used it as an alignment feature is the most likely suspect.
You don't appear to be defining a clear axis along the length of the part.

[[Ja...]]

Y14.5-2009 9.3.1 States that the datum axis for a runout tolerance may be established by either one or two cylindrical datum features, or a cylindrical datum feature and a perpendicular face. If taken literally, I would assume that a conical secondary datum is invalid here and that the runout should be controlling the conical surface while taking the cylindrical hole to be a datum.

Also, are you measuring total or circular runout? What exactly does your alignment look like? How long is the cone?

[[Jo...]]

while we are on the subject. If the situation was reversed, the cylinder being the "datum" and the cone measured for runout.. Is the radial runout perpendicular to the cone surface, or the cylinder surface ? Radial or Axial or neither ?

[[SH...]]

Please sign in to view this quote.

I don't think 0.005 offset is a root cause of your problems, definitely it will affect the concentricity. From the first principle runout is the sum of concentricity and roundness. I don't think base alignment has a significant impact on run out. If you give primary and secondary datum to your runout template, calypso will automatically calculate datum reference frame for the runout. I think your datum reference frame of the runout is not according to designer intention.

[[Ja...]]

Please sign in to view this quote.

Axial - Perpendicular to the cone surface

[[Me...]]

You can't have radial runout to a plane. You want axial runout.
The plane is creating the spatial and the cone is the center origin. I would check the whole bore, not just a circle, and look at the perpendicularity as an indicator of what's going on. You can also use true position as a cross check.

[[Ma...]]

It is impossible to know what is wrong with your runout measurement without a lot more information. What CMM and probe are you using? How exactly are the measurements being taken? How many stylii are being used? Are parts being measured with or without a rotary table? How is the the datum structure implemented? What are the actual runout values observed by the two different methods?

[[SH...]]

Please sign in to view this quote.

Wow ,it looks like FBI interrogating 😀😀

[[Me...]]

Please sign in to view this quote.

Does your different method of checking runout hold and spatially orientate the part by the plane? The reason I ask is that checking runout between centers is completely different than what the datum structure you described is asking for.

[[An...]]

Please sign in to view this quote.

Not always.
See attached.

Contribution_07_02_2020_1.pdf

[[SH...]]

Please sign in to view this quote.

It's a wrong call, it should be perpendicular with respect to A.

IMG20200207161759.jpg

[[Di...]]

Please sign in to view this quote.

Thank you all for replies and taking time to help me with my issue.

I am using Micura and two R2.5 mm probes that are 50 mm long. Probes are opposite each other along X axis. Probes have been requalified before measurement. Parts are being measured without a rotary table. My base alignment is done by taking circular scan on the plane for spatial alignment and then one circular scan on the cone for Z and Y axis origin. Then I do one more scan of plane and 2 scans for the cone for my final alignment as I have said in my first post.

Second method that we use to measure this runout has the plane of workpiece sit on the gage on 3 points and the cone touches the gage in 2 points. Then we rotate the workpiece and inspect runout with dial indicator. This method gives a value of 0.002 to 0.003 mm of runout, which roughly corresponds with the form of circle that I inspect on the CMM. And the runout measured on the CMM is around 0.010 mm.

[[Ch...]]

Hi,

Have you : measured a feature to use for 'RT Axis' on the part itself? If not, have you indicated the part in prior to inspection?
Both of these should improve your results, and I strongly recommend using a cylindrical feature for RT Axis on the part. Calypso Manual Chapter 14.

Good luck and happy measuring.

[[Ma...]]

Please sign in to view this quote.

Here are some things that can cause discrepancies between your two methods:
1. Measuring only a circle on the cone for your base alignment is susceptible to error calculating the center coordinates of the circle because any off-axis scan path will be hyperbolic rather than circular.
2. If the plane is convex or concave, the deviation in your center point from item 1. can cause tilt error to the plane in your second measurement.
3. If the axis of the cone is perfectly perpendicular to the plane it doesn't matter which height along the cone axis the radial origins are set. If the two are not perfectly perpendicular then this height may become significant.

I would try measuring the cone by taking more than two circular scans. Then in order to duplicate the height of your fixtures radial contact points construct a theoretical plane parallel to the datum plane but at the same Z-height as your contact points. Construct an intersection between this plane and the axis of the cone. Then construct a perpendicular using the intersection point as the first element and the datum plane as the second element. You can then use this perpendicular as the primary and secondary datums for the runout evaluation. Or as primary, x, y, and z datums for a secondary alignment.

I would also recommend that you use gaussian fit for the cone and plane measurements to improve the repeatability.

[[An...]]

Runout and dimensioning.
See attached.

Contribution_10_02_2020_.pdf