Help Finding Very Small Holes

[[Mi...]]

Looking for ideas on this:

- 1.5mm Diameter holes
- must be measured from the outside of the hub, not inside(obstructed)
- must use a 1mm ruby

Checking for size and position, pattern is allowed to rotate around that central bore... which is the issue. There is so little margin of error to get a 1mm ruby in a 1.5mm hole.

On many other parts we deal with this situation however the hole is either countersunk so I can take a self centering point with a larger probe and trust it will center in the cone, or I can access the ID and self center that way. Neither is true for this.

I had an idea to use a larger probe and Point Set scanning a grid, then somehow pulling the smallest radial distance, then self centering on that - but I couldn't get that working.

This program needs to work on a production shop floor with no manual interaction.

Hub Holes.PNG

[[No...]]

I achieved good results by scanning a circle segment over the hole with high point density (ruby larger than hole) and calculating the maximum coordinates. These should roughly represent the center of the hole. This will only work if the axial position of the hole pattern is fixed and you can make sure the circle segment hits at least one hole.

But I see another issue: If the pattern can rotate freely around the bore, how can you be sure your stylus has the correct angle to fit into the hole without shanking?

[[Mi...]]

Please sign in to view this quote.

GD&T Rules it can rotate freely - Thankfully our process is designed using to orient the holes in line with the large drilled in the photo you can see. Otherwise this would never work without a rotary table.

So on a perfect part, Oil hole 1 is perfectly in line with Pinion Hole 1, so all I'm needing to do is account for reasonable amounts of error. The position tolerance is 0.125mm, but in reality I need the program to be able to measure parts significantly more out from nominal than that, or reliably react to it if it's not (Missing bore + PCM handles that). PCM will also react to high form error indicating that the probe shanked.

I will try your method, I've never used the Maximum/Minimum coordinate on a circle - only for planes.

I did just get it to work once on a fairly good part I need to get some bad parts to make sure it still works:
-used the point set to scan an area that covers what I think is likely to cover any out of position holes.
-Set the Evaluation to Inner Tangential to hopefully grab the center of the hole
-used it in the Rotation of a secondary alignment for the First Oil hole.
-That first oil hole uses a Meas. Ref to find it with points then go back and scan the updated position
- then the other 3 holes are on yet another Secondary Alignment that use the first as it's rotation.

I dislike using too many secondary alignments, but for something like this it's unavoidable I think

[[Ri...]]

The min/max coordinate is great for finding things that aren't always where they should be. I have used it a lot to find gear teeth roots, ensuring I'm in the root of a thread, etc.

Is this being measured with an active sensor, so that you can utilize self-centering?

[[Mi...]]

Please sign in to view this quote.

Both Active on Contura and GageMax (XT Gold), and Passive on DuraMax and Eclipse (XXT TL3), but I'm not sure where the notion that XXT can't self center comes from? I've seen that before.

We self center with passive sensors all the time, it's much more limited than XT, and you usually have to lock at least 1 Axis for it to work properly where we wouldn't with XT, but it works.

All that said, I can't simply use a SS point here because there is the danger that the hole is so far out of position that the point will miss and slide around the hub, for that reason we usually measure from the ID, but that's not possible on this part.

[[Mi...]]

For anybody seeing this later, my current solution is a combination of my original idea and Norberts.

Min Coordinate on the Point Set is working, the results from that are within a few microns of what the actual scanned hole ends up being. I still have to experiment with very bad parts to see just how far out of location they can be before this method fails.

Thanks, and if anybody has something else they would try don't hesitate

[[Ri...]]

Please sign in to view this quote.

You absolutely can self-center with an XXT, but you are just limited in controlling the force, and the direction.

[[Mi...]]

Please sign in to view this quote.

Unfortunately I have to program for the least capable machine a given program will be on, it's made for some interesting solutions making sure a program runs both on a huge GageMax with XT and an Eclipse with XXT.

It means lots of PCM to take advantage of higher scanning speeds on the XT machines or hurting efficiency by programming for the passive head and it's limitations.