Tubing Alignment

[VONDY]

I have a tube that is sort of "L" shaped but of coarse it is not. It has a few bends of 10º & 4º and then 14º this way and that way. Anyway I am very new to tubing and I am having one he77 of a time leveling it. I have taken 4 manual circles. Two are on the only straight area on the part, the other 2 are on the 'L' section that are not square or parallel to anything. I have tried to use iterative and best fit, but niether have done me justice. The part seems to flip out into nowhere land. Can some one give me the complete ABC's of using the correct alignment. Maybe some suggestions on how to take my manual hits, should I be using circles or should I be taking points instead(?)

[Paul Sarrach]

can you scan the print, have a pic? maybe program, iges or .cad?

[cmmguy]

without a picture it is hard to say what the best way is to align.

Are all of the segments in the same plane?

[Matthew D. Hoedeman]

You can not use circles in an ITERATIVE alignment UNLESS they have a minimum of 3 surface sample hits. This could quite easily be the problem. Without the samples, it is not really sure exactly where the circles actually are along the vector of the circle.

[craiger_ny]

Like Matt said you can't use the circles. Perhaps after constructing lines between them you might have something useable (at least to get the thing started to go into DCC then realign) but I'd be skeptical depending on the tolerance. I'd try working with cylinders. A sketch or snip of the drawing would help.

[rebeldude]

Try to make a line between the 2 circles for each angle and then do an alignment

[VONDY]

[kbotta]

Are you trying to set this up to math data? You might want to consider using the intersection points - and best fitting it. Suggestion...
Kev

[VONDY]

I am using a cad model and will dimensioning it to the print the best I can.

How can I use points if the part is round? Surface points? Where?

[layout tech]

We measure tubular parts also. This could be aligned using a 3 rotation alignment and then angular axis rotations to duplicate the bends. It would take an adjustment of alignment (angle change) to measure each tube segment accurately. It would be very hard to explain in an e-mail, but once you understand the concept it is not to difficult.

I would be glad to try to explain e-mail me and I will give you a phone #

[kbotta]

Well, if you can find 3 intersection points (where a bend meets another bend) you could best fit those points. What I would do is: before any programming is done, find these points 'offline' by creating cylinder intersections, and writing the #'s down. Next for a manual alignment physicaly measure the part to create the intersection points. Adjust the theo's with the numbers you wrote down. Best fit for your alignment, then cad=part. Not the best way probably - but it should work. I have done a corvette hinge assy this way (looks a lot like your tube).
This should get you started, then you can hit the whole dcc thing...
Kev

[Don Ruggieri]

Alternatives

You can use features and induce multiple rotations to align it, if you can figure out where they belong.

I've worked with tubes in the past. I used a 3D best fit on the joint/intersection points. You have 3 major bends on that part, so there are three intersection points. You can measure adjecent cylinders (ignore the actual elbows for now) to get each intersection in 3D space. Also, there are the two endpoints, found by intersecting a cylinder with a plane, or by projecting a circle center onto a plane, your choice.

So, in total there are 5 "control" points on that part which are not too difficult to idenify and measure. You need a minimum of 3 for a unique alignment. You can use any 3, or 4, or all 5. Measure, intersect, construct, etc and then use 3D best fit.

Keep in mind, this is not the only way, and it may not be the best way, but it worked for me.