Intersecting Two Radii

[[WE...]]

Hello,

I'm trying to write a program that has a 2.500" radius into a 0.250" radius. There is a 1.5638 diameter called out where these two radii intersect.

How do I intersect these two radii? I do have Curve, but it won't let me intersect when I create circles or 2D curve. Is there a way to intersect two torus features?

Any suggestions? Any help is greatly appreciated!

Thank you,

Wendi Dowler

[[Cl...]]

Kink point might work for this.

[[Mi...]]

Make a 3d line between each center. Will it let you intersect the line with one of the arcs? This should be at the point the arcs meet.

If this doesn't work there's probably a way to work it out with formulas.

[[WE...]]

Thank you! I'll try both ways and see what I can come up!

[[Cl...]]

Can you post a drawing?

[[WE...]]

I've attached the drawing. The large radius is 2.500" and the 1.500 diameter should be 1.5638 (customer revision).

Drawing - Need to Get 1.500 diameter.png

[[Me...]]

If you have a CAD model I would section the model and create a curve from both radii and check it with profile. Trying to report the radius of those 2 features is going to be difficult with such a small section of the diameter. You'll probably need to constrain the location to get a more accurate radius measurement if you decide to go that route.

If you're trying to do this with no model then you will need the center locations of the radii relative to each other and/or some common features to accurately program their nominal location.

[[Ro...]]

you can intersect the 2 circles, it can be done.
Create the 2 circles (which im sure youve already done 10x over and over), update your nominals so they are exactly in line with each other, set your Strategy to POINTS, 5 points each.
The final thing is to have the last point of each circle be VERY CLOSE to one another, im talking a couple thou of an inch.
and that's it, it'll now find its intersection every time.

[[Mi...]]

Please sign in to view this quote.

What I said with the 3d line won't work. I was thinking the part's curvature reversed, like going from a fillet to an outside round.

[[WE...]]

I will try these methods out and will let you know how it works.

Thank you again for everyone's help! 😃

[[To...]]

Excellent opportunity for engineer to incorporate Line Profile.....

[[Da...]]

The sad truth is that these two radii only intersect each other in CAD land. They have to be perfect, or you'll get no intersection, or two intersections. Profile would be a better way of defining the shape that you're looking for.

[[WE...]]

Hello,

So I tried the different options mentioned on this post, and I'm still not getting the two radii to intersect both manually and in CAD. I've had several models made, and I can't get any of them to intersect. Unfortunately, the customer wants the diameter read out and not a profile read out on the report. 🙁

[[Mi...]]

Please sign in to view this quote.

This is pretty cheesy, but it might get you the dimension you need. Note that I'm assuming the part is oriented with the hole and countersink pointing up: the drawing's left is +Z, top is +Y, and the view is looking -X.

1) Make a theoretical plane oriented the way the drawing you posted is, so the plane's normal is defined as +X.
2) Scan both circles. Try to set it up so the scan paths lie in the plane you just created.
3) Project both circles to your theoretical plane. (This will just project their center points)
4) Make two theoretical circles. Make their normals +X, the same as the theoretical plane. Use formulas to tie the X, Y and Z locations to the projections you just made, and their diameters to the circles you just scanned.
5) Intersect these two theoretical circles.
6) Repeat with more theoretical planes, each representing a different "slice" of the feature in question. Then make a circle recalling all the intersections.

You could optimize this a bit with patterns and groups, but this seemed like the simplest way to explain it.

[[Me...]]

This may or may not work...

Scan each side as a curve that includes both radii. Recall the curves into a cone and use Cone Calculation to find the diameter at the depth from the top. You will need to know that depth where the radii are intersected.

[[Ke...]]

Please sign in to view this quote.

Building on Mr Scott's point: Do you have any equipment at your disposal to perform linear traces? (like a Surfcom, MahrSurf or TallySurf) It would be helpful if you could verify that the part is formed properly, and that you can actually see the point of intersection. if it is blended, and you don't have a clearly defined intersection, you could be chasing this ghost forever.

[[WE...]]

Please sign in to view this quote.

I traced on our Surfcom and our Mitutoyo, and the rad is blended with no visible transition from one to the next. They require a mirror finish and smooth transition. Both radii are reading in tolerance; however, I cannot get a clear intersection point on the trace on either machine.

I did try Mr. Scott's method, and I get an error that says "No results: Error with references" Not enough points to calculate feature. I also tried the cone calculation method and had the same error that there were not enough points to calculate feature. I think Keith is right about chasing a ghost on this one.

I think I'll just have to provide a separate report and measure using the comparator. Measure down the .095 length and get the diameter at the position. Similar to the cone calculation method. I have a large enough tolerance that I think I can get away with that. They just wanted everything measured on the Zeiss.

Thank you all so much for your help and input on this subject! 😃

Thank you,

Wendi Dowler

[[Ke...]]

Please sign in to view this quote.

You can try circle on a cone, at the 0.95 measurement point (verify correlation with the optical comparator), so it can be measured on the Zeiss. That way you can keep data on roundness, if it's not already being reported in some way.

If you need to go the optical comparator route for the measurement, you can enter it using the Result Element. At least that way it is all on the same report.

[[Is...]]

The intersection of the two radius is a tangential position of the geometry. You could generate a bestfit alignment using the complete curve, then you can intersect based on the displacement:



The main advantage of this is that can analyze measurement on the line profile:



You can increase the accuracy using only segments of the curve near to the intersection point:

curve.7z

[[Da...]]

It's not an answer, because the radii are such small sections, but, if you could probe each arc as a circle and create a line from the centers of both circles. that line would be at the actual intersection . If both radii were nearly perfect. If you intersect the line you just created with each circle respectively , that would create a point above and below the theoretical "blend point" use the mean of the two points and your length would be pretty darn close to that.