Skip to main content

Magnets and Geometry

A bunch of us are in the middle of a short fad of fascination with magnets. I recently purchased about a thousand gold-plated spherical 6mm N35 neodymium-iron-boron magnets from Edwin Science. These are strong enough that swallowing two of them would probably require surgery. But I am not dumb enough to eat them, being far more interested in building things out of them, and thus I give you some photos of the nifty things I did when I got them, and some tips on how to assemble them. I regret not having the artistic skills or patience to illustrate all of the information I want to share here, but hopefully those interested can follow along.

The first building block is a loop of magnets, with the poles aligned circumferentially (that is, a string of aligned magnets joined end to end). In the photos below, anywhere you see two parallel rows of magnets arranged in a square configuration, that indicates the adjacent rows are aligned in opposite directions, and almost always part of separate loops. If the rows are arranged in a triangular configuration, that indicates they are aligned in the same direction. The loops can be treated as polygons (squares, pentagons, hexagons, octagons, etc) to create polyhedra. The strength of the magnets causes them to exert force in the direction of perfect alignment, which gives rigidity to what would otherwise be flexible circles.

Truncated OctahedronOctahedronOctahedronGreat RhombicuboctahedronCuboctahedronDodecahedronDodecahedronDodecahedronDodecahedronDodecahedron, truncated

Two or more loops of the appropriate relative size can be stacked to form a shape somewhere between a cone and a pyramid. 5-10-15-20 makes a pentagonal pyramid, 4-8-12-16 makes a square pyramid, 6-9-12-15 (a "loop" of 3 magnets is not stable) makes a triangular pyramid. For reasons related to the bulk of the individual spheres, the triangular pyramid is too tall and too circular to be of much use, which is unfortunate since a triangle is by far the most useful shape for building polyhedra. The square pyramid is better, but far less useful. This leaves the pentagonal pyramid as the building block of choice for a large number of assemblies. Each "pyramid" can be assembled from as few as two to as many as four concentric rings (a fifth ring results in most structures being too heavy to support their own weight... until I get some N52 spheres!). Each layer can be "up" or "down", so you may get a rippled surface instead of a cone shape, and the entire resulting shape can be installed into the final product pointing out or in. Also of note is that the pyramids have a "handedness". Where the rings can be simply flipped over if they are out of alignment, the pyramids have to be inverted if you need to flip them.

15-magnet Pentagon UnitDodecahedron, inverted pyramid faces, assemblyDodecahedron, inverted pyramid facesDodecahedron, offset pyramid facesTwelve Pentagon EggSix Pentagon CylinderDodecahedron, pyramid facesDodecahedron, pyramid facesDodecahedron, pyramid facesDodecahedron, pyramid faces, interior mountingTwelve Pentagon RingTetrahedronDodecahedron, pyramid faces, interior mountingRhombic Prism


  1. How much did a thousand of those cost?

  2. About $105 for 1100 magnets. I haven't seen cheaper.

  3. Complex shapes made from magnetic spheres:


Post a Comment

Popular posts from this blog

A Capacitive-Touch Janko Keyboard: What I Did at the 2017 Georgia Tech Moog Hackathon

Last weekend (February 10-12, 2017) I made a Janko-layout capacitive-touch keyboard for the Moog Werkstatt at the Georgia Tech Moog Hackathon. The day after (Monday the 13th), I made this short video of the keyboard being played: "Capacitive Touch Janko Keyboard for Moog Werkstatt" (Text from the video doobly doo) This is a Janko-layout touch keyboard I made at the 2017 Moog Hackathon at Georgia Tech, February 10-12. I'm playing a few classic bass and melody lines from popular and classic tunes. I only have one octave (13 notes) connected so far. The capacitive touch sensors use MPR121 capacitive-touch chips, on breakout boards from Adafruit (Moog Hackathon sponsor Sparkfun makes a similar board for the same chip). The example code from Adafruit was modified to read four boards (using the Adafruit library and making four sensor objects and initializing each to one of the four I2C addresses is remarkably easy for anyone with moderate familiarity with C++), and

Freesiders Hackers Collaborate in Medical / Surgical Research

Published in the May issue of the Journal of Foot and Ankle Surgery : " A Novel Combination of Printed 3-Dimensional Anatomic Templates and Computer-assisted Surgical Simulation for Virtual Preoperative Planning in Charcot Foot Reconstruction ." This collaboration of specialties represents an undertaking by members of Freeside Atlanta , Southern Arizona Limb Salvage Alliance , and The Podiatry Institute .  Charcot foot reconstruction remains on of the most challenging procedures in foot and ankle surgery.  These procedures are often lengthy procedures which can be riddled with complications. With the help of Freeside Atlanta Members, institutional researchers used open source Osirix Image viewer and 3D Software such as Newtek's Lightwave or Blender to create simulated surgical reductions as well as 3D printed templates.  Freeside Atlanta members assisted in providing 3D printing solutions and know-how to the project. Experimental test prints were done on a M

Onboard Firmware of the Human Brain

Freesiders are continually tinkering with robotics and other such machinery .  Many of these embedded processors and firmware are becoming open source and every-more diversified in the wake of the modern Maker movement . One notable boost to the hackerspace arsenal is the Arduino (an like platforms).  This offers designers an incredible power to devise not just individual devices but even the emergence of complex, integrated systems . This evolutionary pace of modern technological systems may be significantly faster the biologic system development, but there may be a few well learned tricks yet to be mastered.  It seems that studying how nature has managed to solve many development challenges will aid in designing robotics, where efficiently counts just as much. One  challenge, that is particularly interesting, is data processing.  Artificial intelligence is labored with processing data and producing a meaningful and useful output.  When considering the increase in sensory