Kossel magnetic joins problem

Hi, I would like to ask you if you ever experienced problems with magnetic joins on delta printer. Currently I build my own Kossel Mini with magnetic joins - 10mm ball bearings and neodym magnet Ø 10/7-3,5 mm, height 3 mm, the magnetic force stated should be 1,6 kg. Print is running fine but after some time the join disconnect. Last time I did one hour print and then it randomly disconnect. I also tested printing without extrusion and have the same problem. Till now it never happen in the center of plate but mostly further from center. Do you think it can be caused by some inaccuracies on rod lengths or printed parts? It looks it is not influenced by the speed of printing as I also tried manual moving of the arms and sometimes got the same problem.
Any help will be appreciated.

Post a picture of your Effector arangment

Here are the pictures of effector and linear part.

Is the carriage crooked in the third picture? Does one rod start closer to the effector than the other? Everything needs to be perfectly symmetric.

I did not notice that before. For me it looks more crooked on the picture then in real. But you could be right. I can try to print the carriage again when I get new magnets for it. Thanks for telling me that.

Your magnet look pretty weak , the usual size is 3/8X3/8 cylinder (about 9.5X9.5mm) near 7.5-9 lbs (3.5-4kg) and the best configuration is the other way around, 10mm balls on effector and magnet in the arm with a 45 degree angle at the effector and at the carrier and a very minimal gap between magnet and balls.

Everyone that I saw run the other configuration like your have problem with disconnect and finally switch it around or go back to RC joint saying magnetic suck.

BTW if you do switch them around there will be a good opportunity to get carbon arm, way better then aluminum one

The problem with magnetic joints like this is that the forces do not line up correctly.

Take the effector, the magnets are pulling the balls down vertically. There is only a small dimple formed by the diameter of the hole in the magnet to handle any horizontal forces. And those are the strongest. So depending on the arm position the forces are more or less handled by the magnet.

If you had cylindrical magnets in the arms and balls on the effector, the forces are always aligned with the arm direction, so the magnet is always attracting at maximum force.


One way to increase the magnets attraction force is to place a central iron core in the hole of the magnet, while NOT touching the magnet directly. so a 2 mm diameter rod, if the hole is 3 mm. Also make sure the rod just misses touching the ball. This gives "focus" to the magnet and increases the effective force.

Lykle

Yep I had k800 style magnets and they disconnect too easily. Magnets in the rods works, make your own or buy some from Haydn, I can print at very high speeds/accels/jerk without disconnections now.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Thanks guys for your comments. I was thinking also about the RC traxxas joins but when I build them there is always a small volition in the traxxas. But I guess that it can be solve by connection of both arms, right? Anyway when I build the magnetic joins I was surprised by the nice looking prints. So I wanted to use them. I will think about your proposal to switch the magnet and ball. About the carbon ones, I also plan to use them as better choice. You are right.

Currently I solved the problem by adding a string rubber (later I plan to use spring instead) between the effector and the carriage. Till now it is working well. It does not looks so nice but it is working and it is currently more important for me. What do you think about this solution? For me it looks like a good combination. There can be even completely removed the magnets from this solution but maybe the combination is better.

Sounds reasonable, any images?

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Here are some images. It does not look good but it is fast and working solution. Sorry for bad quality of the images.

Should help. If you can keep the magnets connected at reasonable speeds this system has a lot to offer, my machine was incredibly precise with the metal version of the k800 system.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

In the end I switch to carbon rods plus Traxxas 5347 and now I am happy with that solution. No more worries about disconnection of rods and the print quality with the Hiwin linear rails looks perfect for me