Wire forming machine design issue

[Jim Lewis]

I'm designing a desktop CNC wire forming machine and running into a problem. As the wire advances it is rotating several degrees and not even in a predictable direction. The wire is pushed thru by a V groove in a bushing mounted on a stepper and a free turning bearing with a V groove that is spring loaded to press against the stepper. Having seen many other machines that show several pulleys I tried adding a second pair of spring loaded bearings with V grooves but that did not help. The V groove bushings are aluminum due to its higher coefficient of friction. The wire I'm testing with is music wire in straight (not coil) form. Any ideas would be appreciated.

 

[EdStainless]

My first thought is that this has to do with the way that the music wire was straightened.
There is helictical pattern to this and it results in a slight twist of the wire.
You could start by taking a few lengths and stress relieving them at 500F for 30 min.
There might be a slight change in properties, but I have seen this done to finished parts many times.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Jim Lewis]

Excellent point but I can't ask my customers to do that extra step and the wire does not always rotate in the same direction so it seems something else is going on.

 

[EdStainless]

You need to figure out if this is what the issue is.
Until you know the root cause you have no hope of solution.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[3DDave]

If it isn't the wire then a tiny lack of parallelism between the drive roller and the wire may also impart a twist.

Change the rollers to allow for the axes to tilt fore and aft slightly with a fine resolution adjustment to see if that changes the twist.

You might also add a brake that grabs the wire when advancing that has a linear guide to restrain against turning; after the advance, release the grip on the drive wheels to release any twist that has developed, then regrip on the drive wheels and return the brake, perhaps by simply using a spring, to get ready for the next advance.

 

[Jim Lewis]

I was also thinking that I might need to improve the precision of the prototype. And I like the brake idea though it adds some complexity. Thanks. Will continue plugging away.

 

[Pud]

On a bigger scale, I have seen a cnc pipe bender in operation.
The pipe was held in a collet chuck and indexed forward to a former and a hydraulic ram put bend #1. Pipe rotated and indexed forward for bend #2 and so on. There was also a saw station to cut the tube at whatever angle required.
This was making hospital beds.

Politicians like to panic, they need activity. It is their substitute for achievement.

 

[Metalfixer]

Maybe look at the drive on a GMAW (MIG) welder.

 

[Stick.]

If you haven't already, I'd try running the same piece of wire back and forth a few times and monitoring the rotation at each end of travel, mainly to see if the rotation is consistent over multiple advance/retract cycles, or if there's some hysteresis and/or randomness to it.

 

[JohnRBaker]

Is the final formed item flat? That is, are all the bends are on the same plane? If so, perhaps what you need to do is create a sacrificial feature, like a right-angle bent section, which is then retained (prevented from turning) during the remainder of the bends, and is then sheared off at the end. Showing us what the final formed shape looks like might help.

John R. Baker, P.E. (ret)
Irvine, CA
Siemens PLM:

The secret of life is not finding someone to live with
It's finding someone you can't live without

 

[Jim Lewis]

Stick - it's a little erratic I think partly because it's not easy to get a good grip on smooth music wire. I tried a rubber roller but that did not work well. I also tried a sort of knurled roller.

JRB - Yes it's a 2D bender. I thought about a right angle bend but don't like the disadvantages. The final form could be any shape - it's programmable.

Does anyone know if the many in-feed rollers on existing machines are free wheeling or are they all driven together? My latest version has one wheel connected to the stepper and the rest free spinning. Tying them together adds a lot of components.

 

[Jim Lewis]

I measured about 5 deg rotation per inch of travel. It rotates the same direction regardless of the direction of advance. And the rotation is not smooth - it's a little erratic. The more I think about it, there is so little contact between the stepper bushing on the wire - just two points on a smooth surface - that it should not be surprising I'm having this trouble. I'm thinking I need to drive all the rollers to distribute and average the small biases - thought that's a major redesign.

 

[EdStainless]

Wire form machines that I have worked with grip the wire in a collet like device and then index it into the form section.
The wire can't rotate.
Often this collet is mounted on a rack to give precise control over the amount of wire fed each time.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[swimfar]

I don't think you need to change your design significantly to fix the problem. Here's an example build that doesn't have super tight tolerances and it still works fine. They are just using a bearing and a copper tube for the feeding mechanism. Maybe the v-grooves are the problem. Try swapping them out for simple rollers.
Arduino 3D Wire Bending Machine:

https://www.youtube.com/watch?v=HPQbKTJPsU4