finding the center of a "triangle"

[JenWasch]

center of triangle.JPG
ALL of these surfaces are cast including Datum "A".
My understanding is that they would be looking to make sure the machining is centered on the casting. We are just in the beginning stages of this project. I am just looking forward towards the inspection process to ensure we are reporting the results correctly and looking for some ideas on how to achieve this
I am currently running PC-DMIS 2018R2
Any ideas would be awesome!

[Matthew D. Hoedeman]

4 LINES, then points at each intersection, then lines from opposite intersection points, point where those 2 lines cross.

[JenWasch]

The two angled "lines" on the sides are arced though. I see the intersection would work, that's a great idea.
any thoughts on how accurate do you suppose it would be measured as a line?

[Matthew D. Hoedeman]

CAST, how accurate is ANYTHING on a cast surface? Put the hits as close to the ends as you can and not have to worry about casting variation causing it to hit the radius. Who knows, the next casting might not have radii on those sides.

[Matthew D. Hoedeman]

OR do you mean the sides are supposed to be arcs? If so, measure an arc (circle) on those sides, then use line top & bottom, circle left & right, and still do the intersection points. You will probably have to construct 2 extra lines and 2 extra circles to get the intersection points where you want them, the direction of measure effects where the intersection point ends up being. Construct a line from a line simply makes a new line with the vector flipped, same for circle, and the combination of those 4 lines & 4 circles will get you the 4 corner points, one way or another.

[JEFMAN]

It's not a triangle, it's a trapeze...
If you search the centroid of the triangle constructed from sides, you have to construct each corner of the triangle, then create each median (line from a corner to the middle of the opposite side) and intersect them.
This intersection is at 1/3 of the length from the corner on each median.
If you search the centroid of the trapeze, you also have to construct the global triangle, then substract the little one :


hope this helps...

[Matthew D. Hoedeman]

oops, yeah, my bad, I just drew it up in cad, the 4 corner method I said won't work, but you CAN make it work. Up to the 4 corner points, THEN, construct mid-points from the 'top' 2 points, the 'bottom' 2 points, the 'left' 2 points, and the 'right' 2 points. Then construct lines from those mid-points. Where they cross should be the center.

[JenWasch]

I have not had to use them before, but would constructing tangent lines do anything?

[Ego Murphy]

center of triangle.JPG
ALL of these surfaces are cast including Datum "A".
My understanding is that they would be looking to make sure the machining is centered on the casting. We are just in the beginning stages of this project. I am just looking forward towards the inspection process to ensure we are reporting the results correctly and looking for some ideas on how to achieve this
I am currently running PC-DMIS 2018R2
Any ideas would be awesome!

I found the goal and highlighted it in bold.

You can play all day with imaginary prismatic geometry exercises (although JEFMAN's trapeze is pretty cool) and still have meaningless results due to the inescapable fact that casting surface variation always, always means no straight lines, no flat planes, and no circular round things - only lumpy surface. You can already see arcs on the triangle sides, right? That ruins the triangle before you even begin.

You need one thing: a Casting CAD model with it's origin set at the theorectical center of the casting, and (this is the hard part) the 1st op machining material removed. Hopefully you have a CAD wizard who can do this, no you can't borrow any of my wizards as they are all busy.

Take the model and put AutoPoints on it. Set them up for a rough 3-2-1 and build a rough Iterative alignment with massive tolerance and only run once. This is the manual alignment.
Note that these points MUST be the same ones that manufacturing uses for fixturing!
Then switch to DCC and make new versions of the 3-2-1 points, but with a fine Iterative alignment: tighter tolerances and looping 5 or so times.
Next, cover the CAD model in points all over the areas that make the triangle/trapeze. Points are free, use as many as you like.
Take all these plus the DCC 3-2-1 points and make a nice big 3D Best Fit Alignment, making sure to choose Vector Least Squared as the correct method.

Both the Iterative and Best Fit have the same goal: make a coordinate system where the target points' actuals are as close as possible to their CAD nominals. All these CAD nominals are based on an origin point in the theorectical center of the casting, so the end result alignment *should* get the casting centered as desired. Then you can inspect the machining and see how close it is to this center.

If you can't get the casting model with the 1st op machining cut into it, due to a shortage of competent wizards, then hopefully the print has basics from the casting target datums.

If you can't get any model at all, then do nothing until the powers that be step in and see to it you get one.

Good luck!