A Theoretical Approach for Selection of Magnets

dskids

Veteran
Joined
Dec 21, 2015
Messages
302
Location
Virginia
WCA ID
2017SKID01
Hey all

Magnetic cubes are getting pretty popular now, and for good reason - they are great.

I recently magnetized my Valk 3 with N35 4mm x 2mm magnets and I consider it a great success (credit: DMCubing youtube tutorial). The cube is even better than before and the improvement to stability and tactile feedback without the introduction of more friction is tremendous. I feel that this particular arrangement provides an ideal amount of magnetic force, i.e. any stronger or weaker and the effect wouldn't be nearly as pleasant.

This got me thinking: How can I find a systematic approach to selection of magnets for different cubes, so that I can replicate the same feeling as my Valk for cubes of different sizes, plastics, and space constraints, with as little trial and error as possible? After some number crunching I believe I have formed the basis for such a method.

Disclaimer: While the math all works out, this approach is still hypothetical and your mileage may vary. I will be "testing" this theory with a few upcoming magnetic cube projects, but proceed at your own risk.

Assumptions:
  • The magnetic force that we "feel" when turning the cube is the shear force of two magnets being drawn across each other. More specifically, the feeling is the torque due to this shear force acting at a certain distance from the axis of rotation.
  • Various plastics and sticker vinyl all have a relative magnetic permeability of approx 1.0
  • The shear force can be approximated as one half of the normal force between the two magnets (source: K&J Magnetics)

The Math:

The normal force between two magnets (the direct pull force that attracts the faces toward each other) can be approximated by the following:

[1] Fn = 0.01 * N * d * t * e^(-1.9 * r^0.7)

Where:
Fn = normal force in Newtons
N = Max Energy Product or "Grade" of magnet (i.e. 35 for N35, 52 for N52, etc)
d = diameter of magnet in mm
t = thickness of magnet in mm
r = the distance between the faces of the magnets (edge piece wall thickness + corner piece wall thickness, generally) in mm

Then the shear force and torque are approximated by:

[2] Fs = 0.5 * Fn
[3] Fst = Fs * n
and
[4] T = Fst * R

Where:
Fs = shear force in a pair of magnets in Newtons
Fst = total shear force for a layer in Newtons
n = number of magnets in the plane of rotation for a given layer (typically 4 for a 3x3)
T = magnetic torque for a layer in N-mm
R = distance from axis of rotation to the center of a magnet in the plane of rotation in mm

The Method:
  • Start with a magnetized cube you already enjoy with a feeling you wish to replicate
  • Use the formulas above to determine the magnetic layer torque value, T, for that cube
  • Then take measurements from the cube you wish to magnetize to determine values r and R
  • Then using the same value of T, but the new values of r, R, and n (if applicable) apply the formulas in reverse to determine the Fn required for the new cube
  • Plug in different values of N, d, and t to determine which size and strength of magnet gives the desired Fn
  • Note 1: Formula [1] is more accurate when the magnet diameter, d, is larger than the thickness, t. The typical case of d = 2 * t (as is the case for 4x2, 3x1.5, etc) gives good accuracy.
  • Note 2: I "reverse engineered" Formula [1] from K&J Magnetics pull force calculator. You could use their calculator instead of Formula [1], but note that even though we are talking about attraction between two magnets, the relationship between shear and normal force is based on the Case 1 - Magnet to Steel Plate result. Also note that their results display in pounds instead of Newtons, but that shouldn't matter as you are consistent with your units.
Additional Thoughts:
  • To use this method you will need a precise way to measure gap thickness, r, between a pair of magnets which requires a pair of dial calipers.
  • Pros: gives a starting "basis" for selecting magnets for a new project, should work for cubes such as 2x2 that require more magnets per layer
  • Cons/limitations: assumes both cubes have a similar starting turning "weight"; if one cube has substantially higher friction or weight, then the magnetic torque contribution will feel different, even if equal to the other cube
  • This theory is based on the above stated assumptions, but ultimately subjective "feel" of the resulting cube will determine whether a given magnet is a good fit for a particular cube. This method merely attempts to quantify that subjectivity for a more systematic approach.
  • This theory is untested and may break down in real world application. Other factors of a cube's design may interact with magnet feel in a way that this theory doesn't capture. In other words, always do what feels best to you.
So, what are your thoughts on this? Do you think something like this is useful or is it a waste of time for a trivial amount of info? Is word of mouth or tutorials by experienced cubers a more valuable approach in your opinion?
 
What would be the benefit over using the web calculator? I've been using that recently and have now ordered a load of 4mm x 3mm N35 magnets as I find 4mm x 2mm a but weak.
 
What would be the benefit over using the web calculator? I've been using that recently and have now ordered a load of 4mm x 3mm N35 magnets as I find 4mm x 2mm a but weak.
The web calculator only calculates normal force for a single pair of magnets. It doesn't account for location of magnet placement in the cube or for multiple magnet arrangements.

This approach assumes that torque due to shear force is the quantity of interest, and it provides a relation between torque, shear force, and normal force. It also generalizes the phenomenon so that it can be applied to cubes of different side length, or to cubes that require multiple magnets per interface (as is the case with 2x2's). Those would be the (theoretical) benefits.

The big question is whether the location of the magnet varies enough from cube to cube for it to really matter. Just a simple comparison of normal force strength in the online calculator may be sufficient for most 3x3 applications. My approach may be more helpful when trying to scale to larger or smaller cubes where the radius, R, or number of magnets per layer, n, is likely to be significantly different.
 
There are certainly a lot of factors to consider. I've wondered about the difference between magnets of different dimensions but with the same pull strength. For example, 5mm x 2mm and 4mm x 3mm N35s are almost the same strength but I'll bet the feel is quite different.
Another thing I was thinking about is bigger cubes. Inner slices have more magnets so I would have thought they would use weaker ones but I've not heard people mention that.
 
Interesting post! Pulled me right out of lurk mode!

You brought up some really good points. I think I would like to add a few. I'm writing this on my iPhone, so I will probably be limited to what I would normally have written had i been on the computer.

I think it's worth noting that there are other variables as well. For example, the location of the magnets (often times a compromise brought on by the very design of the cubie pieces- you're kind of limited to where you can mount magnets), specifically their location from the center of the cube and how torque comes into play with the perceived magnetic feel when the cube is put in motion. Secondly, the cube's plastic, which can be broken down into at least two factors: the color, and the slickness or porosity of the plastic. Regarding the former, it becomes obvious in some cubes, the Valk for example, that the magnetic pullforce is lessened in black cubes due to the ingredients (often iron oxides, ferrite, etc.) used in making the black pigment. And regarding the slickness and porosity of some types of plastics, these mixed with different types of lubes can give the cube a quicker action and change the perceived magnetic effect.
I like your thinking though and it's great using these calculations to make predictions. I personally tend to buy a ton of cubes and experiment. By the end of this week, I will have personally magnified 13 different Valks in various prototype forms. It's great seeing what the N38 and N42 4x2 magnets do for these cubes. (Kind of a shame so many folks relegate themselves to N35 or 50 for all the cubes out there when these "in between" grades can really help dial in a cube's performance).
 
There are certainly a lot of factors to consider. I've wondered about the difference between magnets of different dimensions but with the same pull strength. For example, 5mm x 2mm and 4mm x 3mm N35s are almost the same strength but I'll bet the feel is quite different.
Another thing I was thinking about is bigger cubes. Inner slices have more magnets so I would have thought they would use weaker ones but I've not heard people mention that.
Two great observations!

Addressing your first: The shear force between two magnets as they slide across each other is a function of the offset between their axes. So if you took two magnets of the same strength but with different diameters, yes I believe there is a lot of potential for different feel there. However, I believe the difference in diameters would have to be large to notice this effect. This is because as the magnet sits a ways from the axis of rotation, the magnets completely pass across one another with only a very small rotation of the face. So to notice a difference in what's going on during the short time they are passing each other, I would think the diameter would have to be suitably large.

And to your second point: At least one Youtuber has a tutorial up for magnetizing a 4x4 and they address this to some extent. It turns out the corners in particular benefit from using a single magnet of a higher strength than used for the edges.
 
Interesting post! Pulled me right out of lurk mode!
Hey, there he is! Your tutorials have been a big help; I will be referencing them for my upcoming projects.

Regarding the former, it becomes obvious in some cubes, the Valk for example, that the magnetic pullforce is lessened in black cubes due to the ingredients (often iron oxides, ferrite, etc.) used in making the black pigment.
Very interesting info. That actually kinda disproves one of my assumptions of roughly equal permeability across plastics and vinyls. If there were a way to quantify the difference I could augment formula [1] with a "plastic color factor".

And regarding the slickness and porosity of some types of plastics, these mixed with different types of lubes can give the cube a quicker action and change the perceived magnetic effect.
Absolutely true. One thing I noticed with the Valk is that after modding it, I had to completely change lube strategy and setup to get the most out of it. I tried to address this in my "Additional Thoughts" section by commenting that significant differences in weight or friction to a cube could completely change the feel relative to magnetic torque. It's a caveat to my approach for sure.
 
Interesting post! Pulled me right out of lurk mode!

You brought up some really good points. I think I would like to add a few. I'm writing this on my iPhone, so I will probably be limited to what I would normally have written had i been on the computer.

I think it's worth noting that there are other variables as well. For example, the location of the magnets (often times a compromise brought on by the very design of the cubie pieces- you're kind of limited to where you can mount magnets), specifically their location from the center of the cube and how torque comes into play with the perceived magnetic feel when the cube is put in motion. Secondly, the cube's plastic, which can be broken down into at least two factors: the color, and the slickness or porosity of the plastic. Regarding the former, it becomes obvious in some cubes, the Valk for example, that the magnetic pullforce is lessened in black cubes due to the ingredients (often iron oxides, ferrite, etc.) used in making the black pigment. And regarding the slickness and porosity of some types of plastics, these mixed with different types of lubes can give the cube a quicker action and change the perceived magnetic effect.
I like your thinking though and it's great using these calculations to make predictions. I personally tend to buy a ton of cubes and experiment. By the end of this week, I will have personally magnified 13 different Valks in various prototype forms. It's great seeing what the N38 and N42 4x2 magnets do for these cubes. (Kind of a shame so many folks relegate themselves to N35 or 50 for all the cubes out there when these "in between" grades can really help dial in a cube's performance).
I hadn't considered that the type of plastic affected the pull force. The only stickerless one I've done was a Guanlong and I really didn't like the effect but put it down to the lightness of the cube.
You've done an awful lot of cubes. How long does it take you now? I did a speed run and it still took me 35 minutes as I'm really slow at putting them back together.
Where do you source your magnets? I couldn't find N42 anywhere for a decent price which is why I've had to order N35 4mm x 3mm.
And to your second point: At least one Youtuber has a tutorial up for magnetizing a 4x4 and they address this to some extent. It turns out the corners in particular benefit from using a single magnet of a higher strength than used for the edges.
I used N50 magnets in the corners of one of my 3x3s and N35s in the edges to increase the pull strength. It worked well. (I could have done it the other way around!) I will tty a bigger cube at some point but I hate putting them back together.
 
I hadn't considered that the type of plastic affected the pull force. The only stickerless one I've done was a Guanlong and I really didn't like the effect but put it down to the lightness of the cube.
You've done an awful lot of cubes. How long does it take you now? I did a speed run and it still took me 35 minutes as I'm really slow at putting them back together.
Where do you source your magnets? I couldn't find N42 anywhere for a decent price which is why I've had to order N35 4mm x 3mm.

I used N50 magnets in the corners of one of my 3x3s and N35s in the edges to increase the pull strength. It worked well. (I could have done it the other way around!) I will tty a bigger cube at some point but I hate putting them back together.
I hadn't considered that the type of plastic affected the pull force. The only stickerless one I've done was a Guanlong and I really didn't like the effect but put it down to the lightness of the cube.
You've done an awful lot of cubes. How long does it take you now? I did a speed run and it still took me 35 minutes as I'm really slow at putting them back together.
Where do you source your magnets? I couldn't find N42 anywhere for a decent price which is why I've had to order N35 4mm x 3mm.

I used N50 magnets in the corners of one of my 3x3s and N35s in the edges to increase the pull strength. It worked well. (I could have done it the other way around!) I will tty a bigger cube at some point but I hate putting them back together.

Sorry if I mess up with the quotes.
It takes me exactly 1 1/2 hours to do the entire Valk to a point where I let it set overnight (to let the glue cure) before I set it up the next day. I'm not concerned about speed. I do it like a Zen thing. Like building ships in bottles, making trout flies, or something similar. It's really relaxing to me and I basically glue each magnet twice. I had a kid leave a comment on a video the other day saying he did his entire GTS in 25 minutes. Bless him. I'd hate to see his cube though! LOL! But anyway, I use a thicker type super glue and I like to give it time before I do too much reassembling. I think I've done close to 30 cubes now.

About the magnets. I really, I mean REALLY lucked out by having an observant subscriber tell me about gaussboys.com where I get the N38 4x2 and N42 4x2. I usually call to order and speak directly with Dave there. The N38 and N42 have changed the way I approach magnetizing cubes! (I had been searching for N42s since I started doing magcubes back in September of last year.) Most people resort to using really only two types of mags: weak and strong! There's a lot of room in between there!

Anyway, the N38 and 42 available from gaussboys are extra high quality (or, the highest quality magnet I've ever seen). The 42s are gold plated and are jewelry quality. I think they're affordable too. (I've learned, especially with N35s that you get what you pay for). Actually of all neodymium magnets, it's the N35s where you see the worse specs and constancy. None of them are truly 4mm x 2mm. (I've got a theory as to why but I'm still researching that). But I kind of stick to basically one type, the one sold by magnets365.com It's of average quality and I'm used to using it and its results.
For N48, I use Apex Magnets apexmagnets.com They have high shipping so you need to catch them on a weekend where they offer free shipping on orders over $15. They're excellent quality magnets, great plating and very consistent. For N52, I use Thackery thethackery.com Excellent costumer service and awesome magnets.

I don't really use N50, though I did just order several hundred from China to test out. Something interesting about N50. People have been buying them on ebay. I don't trust what they're ending up with. I had a kid tell me the other day he did his Valk in N50s purchased from ebay and the cube turned out great. (What???) Then the same day someone told me he bought N50s off of ebay and found them extremely weak. So I have my doubts as to how well these cheapo ebay magnets hold up to specs. Then just yesterday a guy said he made his Thunderclap V1 with 52s and couldn't feel them. His words, they were "ridiculously weak!" My test for N52s: if your finger and thumb tips aren't sore (from pulling single magnets off the stick) after you've done a cube, you probably don't have true N52s! I'm beginning to think there are a lot of "fake" magnets out there that are not up to true specs. I personally try to buy only from magnet companies and those that I can speak with someone on the phone.

I've got two 4x4s sitting here waiting for me to magnetize. I just recently ordered some 3x2 N35s for the middle layer. Yeah, I'm not a fan of reassembling 4x4s! I plan on doing it side at a time! LOL!

TTYL

Dave
 
Quote:"I'm beginning to think there are a lot of "fake" magnets out there that are not up to true specs"

I've seen this mentioned a few times on the interwebs, and it doesn't surprise me. They all look the same, that's the problem. It's similar to the rechargeable battery 'fraud' . There's a mass of poor quality / ex-equipment / old stock cells that get re-labelled with ficticious ( and excessive) capacity specs, and sold to unsuspecting buyers. They tend to be underweight, but otherwise look indistinguishable from 'quality' types.

It's a shame it's pretty tricky to measure accurately the BH strength of a magnet.
 
Sorry if I mess up with the quotes.
It takes me exactly 1 1/2 hours to do the entire Valk to a point where I let it set overnight (to let the glue cure) before I set it up the next day. I'm not concerned about speed. I do it like a Zen thing. Like building ships in bottles, making trout flies, or something similar. It's really relaxing to me and I basically glue each magnet twice. I had a kid leave a comment on a video the other day saying he did his entire GTS in 25 minutes. Bless him. I'd hate to see his cube though! LOL! But anyway, I use a thicker type super glue and I like to give it time before I do too much reassembling. I think I've done close to 30 cubes now.

About the magnets. I really, I mean REALLY lucked out by having an observant subscriber tell me about gaussboys.com where I get the N38 4x2 and N42 4x2. I usually call to order and speak directly with Dave there. The N38 and N42 have changed the way I approach magnetizing cubes! (I had been searching for N42s since I started doing magcubes back in September of last year.) Most people resort to using really only two types of mags: weak and strong! There's a lot of room in between there!

Anyway, the N38 and 42 available from gaussboys are extra high quality (or, the highest quality magnet I've ever seen). The 42s are gold plated and are jewelry quality. I think they're affordable too. (I've learned, especially with N35s that you get what you pay for). Actually of all neodymium magnets, it's the N35s where you see the worse specs and constancy. None of them are truly 4mm x 2mm. (I've got a theory as to why but I'm still researching that). But I kind of stick to basically one type, the one sold by magnets365.com It's of average quality and I'm used to using it and its results.
For N48, I use Apex Magnets apexmagnets.com They have high shipping so you need to catch them on a weekend where they offer free shipping on orders over $15. They're excellent quality magnets, great plating and very consistent. For N52, I use Thackery thethackery.com Excellent costumer service and awesome magnets.

I don't really use N50, though I did just order several hundred from China to test out. Something interesting about N50. People have been buying them on ebay. I don't trust what they're ending up with. I had a kid tell me the other day he did his Valk in N50s purchased from ebay and the cube turned out great. (What???) Then the same day someone told me he bought N50s off of ebay and found them extremely weak. So I have my doubts as to how well these cheapo ebay magnets hold up to specs. Then just yesterday a guy said he made his Thunderclap V1 with 52s and couldn't feel them. His words, they were "ridiculously weak!" My test for N52s: if your finger and thumb tips aren't sore (from pulling single magnets off the stick) after you've done a cube, you probably don't have true N52s! I'm beginning to think there are a lot of "fake" magnets out there that are not up to true specs. I personally try to buy only from magnet companies and those that I can speak with someone on the phone.

I've got two 4x4s sitting here waiting for me to magnetize. I just recently ordered some 3x2 N35s for the middle layer. Yeah, I'm not a fan of reassembling 4x4s! I plan on doing it side at a time! LOL!

TTYL

Dave
I've done 2 Thunderclaps with N52s and the magnets were too strong but I have got used to them to an extent. My order of N35 4mm x 3mm came early from China. They were due any time in March so I'm pleased (eBay seller shinequip2010). I've just done another Thunderclap with them and it's certainly a better feel. I found that the dimensions made it a bit trickier as they seem to be quite lively. Flatter magnets align themselves better too but it wasn't a big issue. I used a new glue too. It's the Gorilla superglue gel. Shawn Bouké had recommended it and I happened to see it on offer. It's a nice thick gel that dries quickly. The top isn't as pointy as the Loctite superglue gel I've been using so it wasn't quite as easy but again, no big issue. I'll probably do the rest of my Thunderclaps the same way for MBLD.
It's a shame the sites you linked are US based as I'm in the UK.
 
Back
Top