Finding the center of a very small clocking notch

[[Ry...]]

Good afternoon, I have a part with a large diameter and a very small alignment notch. We are talking about >1ft diameter cylindrical part with a notch around .050" wide and barely any depth to it. This is a clocking notch to set up orientation.

Rotation in space is set as the top plane along with the Z origin. X and Y are zeroed on the cylinder.

I have been trying many methods to get the position of the small notch on the OD and clock the rest of the geometry to that.

Using a 0.7mm ball and performing a short trace over the notch I get something that looks like
I am trying to create a point at the lowest radial spot of the curve to clock the geometry to it. So far I am not having luck with the "circle in contour best fit" tool, I think because the groove is so shallow. Hard to see here, but you can see my 2D curve trace (Pink) and the circle in contour in blue.
I have tried a whole bunch of diameters for the circle best fit with min contact angle set at 2°. Every time I get it to show a blue line and tell me nominal values for X,Y it is drawn to inside of the part - not to the notch.

Is there a better way I can clock to the center of this barely perceivable feature?

[[Ju...]]

Is the notch big enough to use a self-centering point to probe it?

[[Ma...]]

Sometimes it's better to use manual approach - self-centering point would be nice with combination of manual touch.

[[Ry...]]

I have not heard of the self centering point, will check it out as an option! Not sure where to find it but I will look in the pdf manual. Will report back as to whether or not it works.

[[Ke...]]

Circle in contour is hit or miss... mostly miss, for me.
If a self centering point does not work (I probably wouldn't even try it, given the shallow depth of the notch)
If you have Curve, you can try doing a Min Point, but that may be dependant on how you position the part.
You could also try using Kink Point on each side of the notch, and then creating a symmetry between the two results.
If the clocking isn't overly critical, I would lean toward using a manual point, using a Midpoint.

[[Jo...]]

I've used this method to locate a scribed mark:
1. scan a short high density line crossing the groove
2. get the minimum coordinate of that line
3 use that coordinate as your location

You can probably use a circle with a constrained location and get minimum point from radius

[[Ju...]]

Self-centering is an option under the strategy of a point.

SelfCenteringPoint.jpg

[[Je...]]

Please sign in to view this quote.

.
How did the self-centering option work, Ryan?

I've had to do a similar task when measuring electric motors. The axial slot on the OD was usually the clocking feature.

Both of the strategies that John and Justin described have worked for me. Sometimes, I've had to use the OD of the cylinder to constrain spatial for the base alignment, but then for evaluation to the print, I would make a secondary alignment using a face on the end of the cylinder.

Here's a document that floats around the forums that examples a type of self-centering strategy. I believe Brad Miner's name is attached to it (Senior Applications Engineer at Zeiss).
.
.

[[Ry...]]

It looks like the self-centering will work. I played using it on some counterbores, then on the fixtures set at some arbitrary angles. It did what it needed to do there but not on my test part. The actual manufactured part will have just as wide of a feature but will be deeper than what was provided to me. The machinists are modifying my test part to put 2 index marks of representative size and depth of the single index mark on the final part. That way I can truly test that it worked by knowing they were machined to "X" angle and measured at "X" angle. This was definitely helpful feedback and suggestions!