Hi David, each bend is 2d planar but not in the same plane as each other. Here is another simple tube but you can see the relation ship between the two profiled ends is much more difficult to decipher. what I need is an Unbend tube function similar to how unroll works for surfaces.Laser tube sample 3.3dm (99.2 KB)
No thatās not it because the blue section is twisted throughout its bend where in reality the tube would be rotated by 20 degrees between the two bends
So what we need to be able to do is add up all the angles between consecutive planesā¦ correct? Which should be possible to get from the arcs. Thinkingā¦
-Pascal
Yes in reality the tube is only rotated not twisted but that is a good representation of the relation ship between the two bends. we just need to be able to measure it.
yes exactly
If you look at this file you can see how quickly it becomes a much bigger problem. we need to start with a datum on the first end then establish relation ship to the first bend then measure the twist or rotation between the two bends then the relationship from the last bend to the last intersection. Then the problem is solved.Laser tube sample 3.3dm (99.2 KB)
Am I correct that in the last example, the rotation we are looking for is 18.7 degrees?
-Pascal
I havenāt attempted to measure it yet. Its just random arcs and angles. If youre on the right track it occurs to me that the best place to measure that rotation angle would be half way down the straight length between the to bends because if we might have to re visit the model and change the bend radius which doesnāt affect the angle of rotation only the length of tube.
Yes I make it 18.72 although in reality the tube would be rotated in the bending machine 180-18.72 = 161.28
OK, yes, 18.72 ā¦ I think it should be possible to cook up some tool for figuring this out more automaticallyā¦ Iāll see if it seems reasonable.
-Pascal
Thatās what Iām hoping. Donāt forget the relationship for the end intersection profiles as thatās whatās needed for the laser. This is not easy to decipher in the industry without specific software. It seems like a surface CAD package should be able to solve it easier.
Thanks Pascal
Right, I see that getting the angle is only part of the battleā¦
-Pascal
We need a datum attached to the intersecting profile. I suggest we use a centreline of the cutting tube which I have grouped with the end profile.Laser tube sample 4.3dm (111.0 KB)
OKā¦ I think I see what to do - maybe -
I sent you a PM with an attempt at figuring it out, the long way firstā¦ Basically Iād say we do not need surfaces at all, the CL as arcs and straights, and the end profiles is enough.
-Pascal
Heya, I banged this out without thinking, excuse me if I oversimplified or missed something, but here was the thought that occurred to me when seeing the file:
You appear to be trying to compute the K-factor for a tube. Intrinsic to this will be the tube thickness vs the diameter. Therefore a simple unroll command will not yield accurate results, at least for OAL. In my experience because of an unknown K-factor it is often not possible to precisely determine the free end of a given tube wall/diameter combo any more easily than determining the precise length of the free end of a piece of sheet metal (beyond a certain tolerance).
Therefore I design this free end to be long and then trimmed to length in a subsequent op, which works in the custom environment I work in. Even on longer runs, determining free end length would be a matter of trial and error, but once the part had been run at a given spec/ final geometry a few times I would be able to then spec the correct length. W/ laser rube cutters, you will have control of the location of all bends on the way to end, just not the free end length (unless they have the right software). Without top end software, trying to hit the exact free end length via a formula is not something the foremen at the fabricators I work with would bother to try to do. A test article or two will do the trick however. Bending a few 1st articles will save all kinds of time killed on hypotheticals.
1 Like
Hi Cognitdiss, The effects of K factor are taken into consideration when the tooling is produced to give a true centre line radius which is easy to confirm with a 90 degree test bend. The centre line length is easy to calculate, Iām more concerned with keeping the angular misalignment relationship between the two intersection end profiles. I just need an easy way to recreate the unbent modeled tube.
Given each individual bend is 2D but the bends are not planar with each other, this problem can be reduced to two simpler problems.
One problem is the overall length of tube before bending. This can be determined from the length of the centerline of the bent tube, and then modifying this length to account for bend factors which will depend on the details of how the tubes are physically bent.
The second problem is the orientation of the ends of the tube to each other when straightened. This depends only on the angle of each bend and the orientation of the plane of the each bend. It can be reduced to a series of rotational transformations, and is independent of the radius of each bend and the length of straight tube between each bend.
Determine the angle and plane of each bend. Start at one end, and straighten each bend by rotating the portion of the tube beyond the bend being straightened. There is no need to preserve length during these operations because the length after straightening is calculated separately.
Laser tube sample 3 DC01.3dm (132.3 KB)
A schematic of the centerline and bend angles was created from the centerline information in the original file.
SrfPt was used to create a plane for each bend.
Rotate3D was used to unbend each the portion of the schematic beyond the first bend. The surface in the plane of the bend was used to define the axis of the bend (Rotate3D option)
The unbend process was repeated for the next bend.
After the tube is straightened the angle between the two datum lines is 76.806 degrees. Length of the centerline is 1589 mm.
Hi David, Using the tube centreline intersection point I agree the 1589mm. But I think you need to normalize the two datum lines to measure the angle at 71.070 viewed axially.
I assumed the datum lines were normal to the centerline which I now see was a bad assumption.
@jdelooze Was my post of any use to you?