hexagon logo

Midpoints in bends on a tube

Hello all. I was wondering if anyone could explain how to get some of these points in this drawing, and hopefully help in getting a dimension.

For this part, I would create a plane on the left end of the part, followed by cylinders for each straight. Then a plane on the right side. Next, I would construct a pierce point for plane 1 and cylinder 1 which would give me midpoint 1. Then construct an intersection point between cylinder 1 and 2, and that would be midpoint 2. I would continue this for the rest of the tube until i get to the last pierce point giving me midpoint 7. I have learned that i can create a cast point on cylinder 2, and it will give me midpoint 2.2; however, I have not yet been able to figure out a way of getting any of the x.1 midpoints.

That is where my question starts: How can i get these .1 midpoints? I feel like I've tried every combination of picking two cylinders and all the different types of points, and have had no luck in getting them. I have XYZ coordinates for these points, but more importantly, i have length-of-straights dimensions that are taken as a measurement between for example 1 to 2.1. This is a crucial measurement to the fit/form/function of our part, and as of now, i have absolutely no way of checking it.

My second question is this: Almost every part we have has a dimension for mid 1 to mid 7 across the Y-axis (as shown). But they will also dimension in this case, the top of cylinder 1 to the bottom of cylinder 7 (relative to the Y-axis) across the Y-axis. Every time i try to measure Cyl 1 to Cyl 2 across the Y-axis (with plus diameter getting me the closest), I never can seem to quite get a passable measurement.

Any help in either of these matters would be greatly appreciated.
  • Think of how the bending machine works, it -

    1 feed out the first straight of the tube,
    2 make the first bend
    3 feed out the next straight
    4 possibly rotate the tube
    5 make the next bend
    6 repeat from 3 until done

    Each point set of three consecutive intersection points defines the bending plane for the included bend, so we need to repeat: [construct plane, level to the plane, construct circle, construct pierce points] for every set of instersection points (1-2-3, 2-3-4, 3-4-5, 4-5-6, 5-6-7), startplane and endplane intersection points included.


    It would seem to me, that assuming i have an alignment already, it seems unnecessary to construct planes out of my midpoints (gotten from the intersection between the cylinders) to use for the circle. However, if it IS necessary, can you explain that in a little more detail please?


    As circles are 2D features, not 3D, we need a correct workplane when we measure or construct a circle. The first three intersection points form the bend plane (for the first bend), and so the bend radius circle must be constructed in this workplane. Unfortunately, in PC-DMIS circle construction can't reference a separate feature as working plane (as is possible when measuring a circle) , so we must do the level before construction.

    The lines you had labeled planes were achieved from the midpoint on either side of the cylinder (Or as a cast line based off that tube). Which would be the more accurate/reliable? Using the midpoints or the cast line to construct the plane?


    No lines were intentionally labeled planes, they are the cylinder center lines (or start point to end point, it's the same thing).

    Secondly, Considering i am creating a 2-d feature, would i be correct in thinking i need to be in a workplane looking DOWN at the circle, in a way similar to as if i was looking at your first picture?

    Exactly!

  • Thank you so much for your help Andersl! Hopefully, I only need to ask one more question:

    Let's say i probe out a part with a few bends in it.

    I then create my two cast lines, and intersecting point

    (Here's where I'm still confused)You said to level to the first line (or plane as you have been calling it), but that is only 1/3 of the alignment. Without at least locking down my rotation, I can't accurately choose a workplane to construct my tang-circle (because the axis i level to the cast line, wont be a suitable workplane, I dont think) However If i leveled to the first cast line, select the first line/or cylinder as my XYZ origin, and then rotated any axis (except my level axis) about my level axis towards the second cast line, I believe this would give me a suitable workplane to use.

    Would that part be correct? Sorry if it isn't, I am definitely still learning a lot to building my own alignments.
  • No! Each workplanes is constructed from three consecutive intersection points All you need as a workplane for the circle is LEVEL.
  • I think i may have just got it. I forgot all about using a feature as a workplane. So, I would just level to the first straight right before the bend, and use that same midpoint-midpoint plane as my work plane (Or do i need to use all three points; IE: midpoint for start of first straight, intersection points of both straights, midpoint for end of second straight? If that's the case, I didn't know i i could 'bend' a plane like that, which would explain some of the confusion)?

    Also, when I put in my theos, should i also put in the nominal XYZ for the midpoint of the bend when i key in my bend radius?
  • Threads like this should end up in a "PC-DMIS library" subforum or even as a standalone helpfile.

    There are massive amounts of good information and how-to's in this forum and should really be extracted and "published"/compiled somehow.
  • Or do i need to use all three points; IE: midpoint for start of first straight, intersection points of both straights, midpoint for end of second straight?


    That's it! But there's no "bending" involved in a three point plane - flat as the Nevada Desert.

    Also, when I put in my theos, should i also put in the nominal XYZ for the midpoint of the bend when i key in my bend radius?


    If you actually have those coordinates you can of course put them in, but it doesn't matter for the circle construction itself - there is only one circle with a given radius, lying in the workplane, touching the two lines.

  • Thats so awesome. Thank you so much Andersl! And sorry for not getting it sooner, I completed the PC-DMIS Arms 101 back in December, and clearly still have a ton to learn!

    THANK YOU!