Randomly Rotated Hub.

[[Mi...]]

How else would you go about finding a randomly rotated spline. I need to scan the outside of the teeth. The hub on this part is potentially rotated in any position, independent of the rest of the part. This is powdered metal, the press punch is not keyed and every setup it will be in a different orientation, completely random.

What I'm currently doing is the Zeiss prescribed method, a pattern of two self centering points and then using minimum feature to find the one that actually centered, the minimum feature is then the Rotation of a secondary alignment that gets used for the measurement of this hub.

The problem with that method is on this part, it requires me to use a 20mm disc.

How else would you approach this? Any other idea I come up with isn't nearly as reliable, but I really don't want to use this disc if I can avoid it as we have more XXT machines than XT machines.

Thanks all!

Capture.PNG

[[Mi...]]

What I've sometimes done is make a secondary alignment that just uses a single feature (usually a symmetry point) set to use manual probing. The problem with this is that it requires operator input, and it can't be done until the base alignment has been completed, so someone has to hang around for a couple minutes to probe that one feature, and then it can go into CNC mode for the rest of the run.

I like your method, but if using the disc probe is the only issue with it it's hard to give specific advice without seeing the rest of the part.

[[Mi...]]

Please sign in to view this quote.

Yeah, manually probing a symmetry point would be a good solution but unfortunately here the skillset to do that just doesn't exist for our CMM operators. I was trying to come up with a way of having it automatically do what you suggest with a symmetry point.

The downsides to the disc are, an extra stylus system change, and then change back, on a program that otherwise would have none, adding a cycle time. And I wouldn't be able to do this on our machines equipped with XXT, so it means the only place this program could run in our Quality lab, on the opposite side of the plant from where the part is made and controlled.

[[Mi...]]

Could you clarify why disc probe necessary? is the spline oriented vertically or horizontally?

I'm just trying to picture it better in my head.

[[Mi...]]

Because it's the only thing we have that's large enough to self center on these teeth/castellation's.

The diameter of that hub at the smallest is about 90mm to put it in size perspective, the space between teeth is about 10.5mm.

I think a symmetry point would work perfectly, if I can first make sure it's actually hitting a tooth space.

Edit: Spline is oriented along the X axis. Our parts are mostly round gear carriers. We typically stand our parts up this way to access both sides and almost always avoid part flips.

Capture.PNG

[[Mi...]]

Could you:
- probe two place in the x-direction with a small probe (not self-centering)
- make an alignment clocked to the whichever point has the greater x-value
- make a symmetry point in the ±y direction using this alignment
- then make your final "spline alignment" clocked to the symmetry point
?

There might be an issue if one of the first points hits right on the corner of a spline, but maybe you could do 4 points instead of 2 or something.

Capture.PNG

[[La...]]

What about using a large probe to scan over a section of the teeth using a circle and then using Minimum Point to clock your alignment?
This works well with grooves or divots but I'm not sure if you are able to scan over these teeth without the probe hanging up.

[[Je...]]

.

I agree with Laura's strategy.

Also, I've had success with similar parts using 2D lines scanned into the material. I'd recommend looping your base alignment with something like the following. Also, if you happen to have curve, that works even better than 2d lines when incorporating min./max. features.

.
.


Jeff Frodermann
Meier Tool & Engineering
Anoka, Minnesota

.

[[Ma...]]

Is that hole on right side of image used to fixture part?

[[No...]]

Another thing you may want to try:
First do what you describe in your first post (2 self-centering points), only with a smaller probe. Then, based on the coordinates of the point that centered in the gap (use formulas), probe two more self-centered points in the corners of the gap. This may need some adjustment work to get it right, since you don't know the exact starting point. Make their point vectors point out of the corners by approx. 45 degrees (see image) and set the evaluation parameters to ball center. If you get it right you can calculate a symmetry point that is almost exactly in the center of the gap.

Capture.PNG

[[Mi...]]

Please sign in to view this quote.

I ended up doing exactly this, the shape and size of the teeth was preventing me from having much success with scanning lines unless I went to an 8mm ruby, but I really wanted to keep it with a 5mm as that's a very common size we have already available on most CMMs this may end up.

It's worked so far on current parts, time will tell if I need to adjust the vectors to steeper angles to make it work 100% of the time, or if I'll have any difficulty when I put it on an XXT machine, but any modifications should be minor.

Thanks all!

[[He...]]

Please sign in to view this quote.

I was about to suggest more or less exactly this with the difference that you should not "aim" for the corners with the 45° points but aim straight for the lowest of the first two points. If the lowest of your first two points is exactly in the corner there is a risk that the 45° point will miss the gap and search its way to the next gap. If you aim straight for the lowest point it will find its way into the corner no matter what. But other than that you solution will work perfectly.
Nice job! 🙂

[[Mi...]]

Please sign in to view this quote.

Yup, that's exactly what is happening.

The "Tooth Gap Points" just use formulas their coordinates taken from the Minimum feature - with opposite vectors.

I believe this is what Norbert intended anyway when he said "Then, based on the coordinates of the point that centered in the gap (use formulas)"

Capture2.PNGCapture.PNG

[[Ch...]]

Yes,
self centering and using minimum feature, etc. works well for gear/spline teeth.
I'd suggest adding a stop point to allow operator to align to probe and/or use this method to 'find' the clocking.

I think I got this attachment from someone at Zeiss a while back ..
good luck.

Clocking the Gap of a spline and creating a datum from the pitch circle.pdf

[[No...]]

Please sign in to view this quote.

I think you're right. I used this method a couple of times already, but to be honest, I never really "aimed" at a specific point, but adjusted the vectors and coordinates by gut feeling until it worked reliably.

[[Ow...]]

Please sign in to view this quote.

I randomly work with XXT heads but, I thought I was told that passive XXT heads don't have "self centering" capability.

[[Mi...]]

Please sign in to view this quote.

We use self-centering with XXT often, it works well. Although XT does it better, and is much more forgiving if the feature is not exactly where it's supposed to be. On rare occasion we've had parts that are so out of spec that won't run on XXT, but will run when brought into the lab and are put on an XT machine even though the programs use the same strategy/probe size

[[Da...]]

while the print may say that it is randomly located, chances are it is rotated to align with some other feature. ( if not, then ask the man making the part how he would relocate the part if he had to rework the depth of something) Ask the machinist. it's ok, it doesn't reveal "weakness" on your part. Once you know how he locates it do the same and use a symmetry point to pick it up.