Budget Magnetic Ball Joints

[MorbidSlowBurn]

I know a couple of people, myself included, have been drooling over the berrybot magnet joints. So I decided to attempt one on my own. I did impose some limits to make it more accessible to the masses and something I can produce without purchasing any tools.

I started with obtaining magnets and ball bearings. I selected .5dia chrome bearings for the ball portion of the joint. I also acquired 50 .25x.5 neodymium magnets. Half are grade N42 and the others are grade N52. I did the two magnet strengths, so I can eventually determine impacts of magnetic force over friction. Both magnets had comparable abilities with lifting the bearings. Each was able to lift 10 bearings as shown with the N52 almost able to lift an 11th.

New Toys

Magnet Strength

I went through a couple of design iterations and the last shows a lot of promise. Turns out, simpler is better. I also changed the orientation of my prints to take advantage of the printer's capabilities. Each one has been scaled to allow for shrinkage during printing.

Here is the evolution of the design.

Evolution

Evolution2

This last one was the simplest design but also has many pluses.
-Simple design can be made with a 3D printer or with simple tools.
-Easily modified for various types of rods to be inserted.
-Designed gap between the ball and magnet, Controls centering of the ball and reduces friction.
-Low friction, although I plan on using a candle to wax the interface.

Design

SolidWorks

To do
-Design a mount for the ball on the cheapskate bearings.
-Design a mount for the printer head.

[mhackney]

Nice work! I would suggest taking a close look at the drawings for some of the commercial units (and here) or patents. It looks like the socket is about 1/4 of the perimeter of the ball. This helps centering and keeping it centered. The magnet provides the force to hold the joint together, the ball and socket keeps the joint in the correct position and prevents misalignment. A small contact area won't do as good a job at keeping the socket/ball aligned properly with a sudden acceleration or deceleration. My experiments validated this. Something like this should be easy to print and keep the joint centered.

The commercial units typically claim 180° of motion, again, that's so they can make a larger socket to keep alignment.

[MorbidSlowBurn]

I am still working the kinks out. I did have a slight bevel in some of the designs but was having problems with my layer adhesion. Part of it was orientation when printing. The last one showed the most promise as it had some really good centering characteristics.

Thinking I might go with a thicker top and then use a counter sink to create the bevel. Target is to make something easy to make with standard tools. Not many people have a ball end mill and don't want to buy one for a single purpose.

Once I decide on how to tie it to the carriage and cheapskate bearing I will do some speed tests.

[mhackney]

I think your idea of printing and then finishing up with a countersink to make a clean bevel with simple tools is a good one, simple and effective.

[Polygonhell]

I would machine rather than print these, but I know that's not an option for most people.
So printing and then finishing with a 1/2 inch ball end mill sounds like a good option to me.
I think you really want the cup to match the ball, or you risk getting an effect where as the ball moves it levers very slightly out of the holder then is pulled back in by the magnet. My guess is this would happen during the acceleration at the start and end of moves.

[MorbidSlowBurn]

One of the thoughts I have on ensuring contact is to use wax from a candle and fill in the void between the ball and magnet. Maybe heat a ball to get a really good wax surface.

I have looked at the patents and although they use a ball mill type of interface I still think a ring interface will give similar results. As mhackney showed the larger ring will give a better control of ball alignment.

The magnets I am using have a fair amount of strength. I think there is going to be a trade off between contact patch size, and magnet strength.

It is going to take me little while to get to the point where I can adapt them into my Rostock but I thought I would put my ideas out there as they seem to have promise based on some initial tests. If anything maybe they will inspire others.

[mhackney]

The patent I linked to actually has a bevel rather than a hemispherical cut out. I think that would work great and minimizes contact surface for low friction. Not sure why you would want to fill it with wax? As long as the ball is supported in a plane (i.e. the bevel cut) it should be fine.

Magnetic force will depend on the force of the magnet and its diameter (I think that's what you mean by "contact patch size"?) and the distance it is from the ball (all else being the same). I don't anticipate the friction from even an entire hemispherical socket is going to be a factor, even less with the bevel. The "socket" is mostly to ensure alignment/centering of the ball.

I got a little larger size - 3/8"D x 3/8" long - N52 magnets from K&J with 3/8" steel balls. Might be overkill so it will be interesting to see how the 1/4" mags work. Your N52 1/4" x 1/2" have 4.79# of pull force in direct contact of the entire face. The 3/8" have 9.86# of pull force.

It is instructive to do a little back of the envelope calculations to ballpark the amount of pull force necessary. I plan on running two extruders so I'll have a little more mass on my delta head. The acceleration comes out of the firmware setting. This is greatly simplified but at least a starting point:

force = mass * acceleration

force = 400gm * 800mm/sec^2 = 3.2Newtons = .07 pound force

Remember, the 9.86# pull force for my magnets is if the entire end is attached to a steel plate. In this application, there will be no direct contact. A minimal gap will exist and the bearing is round so magnetic force drops off from the center of the magnet to the periphery. A gap of 1/32" results in a drop to 4.22#. I assumed that would be my center distance. (by the way, there is a neat magnetic calculator on the K&J site) I estimated a distance of .2" at the periphery of the magnet. That is .46# of force. That would be the worse case and is well above the .07# the rough f=ma calculation gave. And, there are two magnetic joints on each axis, so that in effect doubles the pull force to .92# per axis.

I should have mine machined up tomorrow. I did make 1 test joint and can attest that I am getting well over .46# of pull force as measured with a handheld fish scale.

[MorbidSlowBurn]

Sorry didn't look at the patent. Pretty much what I had in mind. And yes I kind of did some back of the envelope calcs. I went with .5" balls because I thought I would get a slightly better magnetic field as the ball has less curve.

The wax idea is a back end way of getting around a Teflon friction pad. It would pretty much stay put since it will be far enough away from any heat source. But that is going to be in my back pocket if I see sticking during motions.

If you look at my evolution of designs my fourth one had three towers around the magnet. This was to minimize contact patch. Problem was it was too detailed for my print setup and I wanted to incorporate a stop for the magnet so it wouldn't impact the ball. Required too much fitting after printing.

I do plan on having two magnets on each ball with the poles positioned to attrack each other. For example, the print carriage has the North Pole facing the ball while the mating arm has a south. This should improve forces between them. Initial (non-scientific) tests required a fair amount of side force to dislodge from the cup or even begin to come off the plane of contact.

I enjoy the feedback. If anything this may trigger a lightbulb for someone else.

[justnick]

What about using some axial ring magnets from the kjmagnetics website like: http://www.kjmagnetics.com/proddetail.asp?prod=R822CS-N that have a counter sunk hole for a screw to seat the steel ball.

Nickel plating should be durable and slippery enough. Anything to avoid an airgap.

And if you need more strength, add some disc magnets in back.

Didn't Werner use magnetic spheres? His entire design was an exercise in simplicity.

-Nick