Springs instead of magnets for CF ball end arms

[bvandiepenbos]

Came across this idea on the Delta bots forum.
Use a long spring to hold effector arms together, CF tubes with ball ends but NO MAGNETS!
Looks like a really good solution. I am going to have to try it.
I have built mag-ball arms, they work ok, but I don't like how they can come apart easily when you move platform by hand.
Not had a fail when printing but at fast feeds they did clack apart a bit with quick direction changes.

Here's two threads that mention it, along with some photos:
https://groups.google.com/forum/#!topic ... ctySnsCPcw
https://groups.google.com/forum/#!topic ... 4Un0drhFoM

Photo:
https://lh6.googleusercontent.com/-t89f ... G_3776.JPG


thoughts?

[Polygonhell]

I saw that as well, I thought it was really clever.

[gestalt73]

What a small world...

I've been wondering if I could increase the grip strength of the magarms, so I'm actually upgrading them this weekend to use two magnets per side.

Just had to model the boomend and increase the length by 10mm.

So far so good...

[Generic Default]

That looks really nice.

The magnet ball system has no backlash, and it stays fairly rigid until it breaks, however the force required to break it is typically around a kilogram per joint. Torquing the platform causes a little bit of rotation because the magnetic joints apply a constant holding force.

This spring based system can be adjusted for a much higher amount of holding force. If you try to torque the platform, the string will apply more force to the joints, especially the one that is being pulled on. That's what makes the effector more rigid. It also has the benefit of evenly distributing the force between the two joints that each string pulls on.

Maybe the design could use a strong, short spring and a set screw or hook screw to allow adjustment easily.




I got a lathe recently and in the near future I'll be machining PTFE ball joint sockets for delta arms. A large preload and a low coefficient of friction would make delta printers more rigid. Make sure you choose thick rods for your arms!

[bvandiepenbos]

What a small world...

I've been wondering if I could increase the grip strength of the magarms, so I'm actually upgrading them this weekend to use two magnets per side.

Just had to model the boomend and increase the length by 10mm.

So far so good...

gestalt73,
by what I have read and tested myself more/longer magnets does not gain any more strength.
I could not tell a noticeable difference between 1/4 dia. x 3/8 long mags and 1/4 x 3/4 long

what about just adding a long spring to your existing mag-ball setup?
A screen door spring from the hardware should do the trick. would be easy and cheap to try out anyways.
just a thought.

[Captain Starfish]

Hopefully the forces involved mean that the offset of both spring end fixings from their associated arm axes of rotation and the resultant variance in force depending on arm angles doesn't come into play.

[bdjohns1]

gestalt73,
by what I have read and tested myself more/longer magnets does not gain any more strength.
I could not tell a noticeable difference between 1/4 dia. x 3/8 long mags and 1/4 x 3/4 long

I saw on one of the google discussions that you get into diminishing returns once the length of a cylindrical magnet exceeds the diameter. I use the 3/8" dia x 3/8" long magnets xnaron spec'd for his design.

I've infilled at 110-120 mm/s octagonal pattern on my setup without shaking the arms off, leaving acceleration at stock values.

[Max]

[disclaimer: I'm in the process of going magnetic, and trying a lot of different things. Some might not pan out.]


Why not use a cap?

It would snap on and off for maintenance, but should hold the ball in place in cases of high torque.

[Max]

Same disclaimer. I was considering rainbow loom bands, 600 for $2.50, or grab a few from the kids... add two or more for more force. Since I still need to print my connection plates I could add them.

[bvandiepenbos]

[disclaimer: I'm in the process of going magnetic, and trying a lot of different things. Some might not pan out.]


Why not use a cap?

cap_tube.JPG

It would snap on and off for maintenance, but should hold the ball in place in cases of high torque.

I was thinking about trying something like this also, but the cap would be held by spring pressure from a piece of rubber tubing slid over the CF rod.

[bvandiepenbos]

Same disclaimer. I was considering rainbow loom bands, 600 for $2.50, or grab a few from the kids... add two or more for more force. Since I still need to print my connection plates I could add them.

tube_bands.JPG

good idea, this one has interesting possibilities.
never heard of the bands.
How about rubber bands from dental braces? ...I think they come in various sizes/strengths ?

I think 3 per rod end, spaced 120 degrees apart might work better than 2 ?

[bdjohns1]

Same disclaimer. I was considering rainbow loom bands, 600 for $2.50, or grab a few from the kids... add two or more for more force. Since I still need to print my connection plates I could add them.

tube_bands.JPG

That's a friggin' clever idea. I think those bands are a silicone rubber, so they should hold up fairly well.

[Captain Starfish]

Same disclaimer. I was considering rainbow loom bands, 600 for $2.50, or grab a few from the kids... add two or more for more force. Since I still need to print my connection plates I could add them.

tube_bands.JPG

Problem with this is that your spring (or band) force is proportional to the amount it's stretched. Two things happen with this design. First is that when you bend one way, the shortening band loses pulling force whilst the tightening band increases pulling force - so the further the arm moves off centre, the greater the "centering" force on the arm. This is exacerbated by the short bands - you're doubling or halving the length and therefore doubling or halving the force as you move around, and that HAS to have an effect on the print job quality.

The long spring isn't so bad because it changes length only by a small proportion as it moves around - even though it's unbalanced and off centre.

A hollow ball with a pin in the centre and the band hooked from that pin, through a hole in the ball and a matching hole in the rod end, to a second pin in the axis of the arm would work better provided the holes could be made wide enough for the full range of motion. Or maybe you could do something with a magnet...

[teoman]

I just randomly found these...

http://www.banggood.com/50pcs-N50-D-103 ... 40059.html