Im also curious to which orientation is best for the end magnets. When I was playing I copied the diagram.
Left .down .right .up .left
With horizontal polarity end magnets. And it didnt seem to work right. When I arranged the end magnets to both be facing north .
Up . Left . Down . right. up.
I seem to get a more satisfactory stick of advantageousness… this could be antidote. I’m sure future chimpo can check it out. And post bellow
Can’t sleep so asked chatgpt. And the results are counter intuitive… so ill defo have to test the concept
Gpt says :
Horizontal ends: poles face along the row (e.g., → ↑ ← ↓ →)
Vertical ends: poles face up/down relative to the row (e.g., ↑ → ↓ ← ↑)
If you have a small number of magnets (like 5), it’s generally better for the end magnets to be horizontal, i.e., pointing along the array axis.
That’s because:
Horizontal end magnets help the magnetic “rotation” pattern remain symmetric and directional — maximizing field concentration on the target side.
Vertical end magnets leave an open flux path at the array ends, which reduces field uniformity and weakens the focused side.
In short:
Horizontal ends: better lateral flux closure, stronger unidirectional field.
Vertical ends: flux leaks more upward/downward, weaker net field.
Practical tests show:Simulations and experimental arrays (even from Halbach’s original work) show roughly 10–20% stronger field on the active side when the first and last magnet magnetizations are tangential (horizontal to the array axis).
If you had many more magnets (say 9–12), the choice would matter less — but with 5, it’s significant.
Conclusion:
For a 5-magnet 2D Halbach array, it’s more advantageous to have the end magnets oriented horizontally (magnetization pointing along the array’s length).
That will:
Strengthen the one-sided field
Reduce flux leakage at the ends
Give a more uniform “strong side” field region
Results are in!!
( → ↑ ← ↓ →) wins
I reapeated it a few times… horizontal polarity ends have the most advantageousness
Thanks for playing
Spoke to tac0s will get an array made and sent to them for an official weigh in.
Subquest. I’ve just discovered triangle magnets… assuming I can find suitible polarity alignments… could we loop a 2d hallbach array into a circle to make super lifty ‘puck’ with all the advantageousness on the one side. Not to be confused with the 3d array that has all the adventageousness in the center…
Yeh.. i found it difficult find exact shape and size magnets needed for such an array.
I wish I had some sort of machine to precisely grind the magnets into the correct shape in respect to its northpole .. oh wait….
I have that machine left over that I used to make those diamonds. Surely that will shave away Neodymium cubes coldly slowly with precision.
Yes… I have already considered building a pulse magnitiser to alter polarity directions… hopefully I dont have to…
Let’s have a discussion before I start chopping magnets
So after a little chat with gpt. We have decided on a 6mm x 3mm octagon made from 8 triangle magnets cut from the 3mm cubes. The aray will be a twice looping hallbach aray at 90* intervals.
Gpt insists the polarity axis should 45 creating a single hallbach loop. But i dont have the means of cutting the 45* polarity axis from the cubes without making the whole aray smaller. And that’s a ballacke. So 90s will do.
Or add more wedges. And im not shaving down the wedges any smaller lol.
I may make a 5mm x 10mm version. Just coz its will be easier to see. To test the concept
Im now Back in the workshop (my bedroom). I’ve dug out the faceting machine and im currently printing a jig to hold the cube.
This is how the cubes will be shaped.
wish me luck.So the first one is slow and painstaking… as i have to set the machine as I go. But after that I can leave it cut em un attended… its going well.
Little more to go..
i plan to cut one side from each magnet. Using this jig. Then ill use a 2nd jig to cut the other sides
