magnetic coupling torque calculations 7

[dnleng]

Although this is my first post here, I have spend many hours during the last years following the forum and getting information from others, but this time it seems that I am in a dead end.

As far as magnetism goes I am a complete newbie ... So take it easy with the answers !

I am looking to construct a magnetic coupling with N52 magnets. Imagine a disc with a circular array of 9 magnets 10mm in diameter, 5mm thick. North pole facing out. Another disc (the driven one) with the same array but this time South pole facing out. I did some preliminary testing and it works, but I need to get the calculations done so instead of manufacturing every time new things we simulate and then we validate what suits us best. At work, we have ANSYS but it seems that only electromagnets are there so I did not have any luck with that...

The other thing is the polarity of the magnets. Is it better to have each array with a North-South-North configuration or the North only South only will work ?

The problem is that we are working in a very small space so we have to keep everything as small as possible and the air gap should be as large as possible. Therefore I think a simulation, or a method of calculating the max torque would be really helpful.

Many thanks in advance !
D.

 

[MagMike]

It'll work much more efficiently if each array has a North-South-North arrangement. This will also require an even number of magnets on each disc, since you'll lose the N-S-N pattern with 9 magnets on a disk.

 

[EdStainless]

You alternate poles and use enough steel to give good return path.
I'll have to look in my notes but you can reduce the calculation to a simple relationship.

You do need to be careful with Neo, when they slip they will get very hot and you can get serious demag if you aren't craful.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[dnleng]

Appreciate the replies,

MagMike, although my knowledge is limited, it will work more efficient because of the magnetic field shapes produced due to this arrangement no ? I can go with ten 10mm N52 but then again they will be very close to each other. I think that does not matter but I am not sure.

EdStainless. What do you mean by enough steel? The discs that the magnets are placed? It has to be from aluminium due to weight restrictions ...

Imagine that we are talking for a very low rpm coupling here. Low rpm and very low rpm motor. 2Nm at its best and 72rpm. The first set of pulleys is connected to the motor via o-ring belts. Then the set transmits the torque through an air gap to the other set of magnetic discs. The trick is to find the correct air gap so the magnets slip before the o-ring slips to protect the motor and the o-rings. I do not think they will overheat at this speed, and we would go with Samarium magnets but their strength is much smaller.

D.

 

[IRstuff]

One way to minimize the need for back iron is to use a Halbach array:

http://en.wikipedia.org/wiki/Halbach_array

The alternating poles are mandatory; using the same poles on each plate gives you attraction, but no significant torque. Just imagine each pole as part of a rotor on a stepper motor. You need to have some semblance of a lateral counterforce to provide the coupling.

TTFN
faq731-376

 

[RyreInc]

You could probably come out with a lower total weight by using a back iron, since there will be much less leakage. Probably substantially cheaper too.

I did a quick 2D FEA sim to demonstrate this, see pics below.

Case 1 uses 45MGOe magnets and no back irons, and results in 3.9 ft*in torque.
Case 2 uses 35MGOe magnets and steel back irons, and results in 6.6 ft*in. (Don't pay attention to the units, just the relative change)

This is with zero optimization--the airgap is huge, and no dimensions were carefully chosen.

Instead of lower energy magnets you could reduce their size. You should be able to add less weight in steel than you remove from the magnets.

Case 1

Case 2

 

[MagBen]

How does your configuration look like? looks to me you are using plane air gap, other than cylindrical air gap as demonstrated by RyreInc, since you mentioned the dics of the driver and follower had a same size.

The arrange of magnet polarizations can be either radial, or tangentia, or axial, with regard to the two discs. which one will give your the highest torque depends on the gap, the diameter of disc, the number of magnets used etc.

The magnets with an even number should be set with alternating poles on both discs. An iron back sitting right on the bottow of magnets will help to increase usable flux density, so higher torque.

 

[dnleng]

Again, many thanks for your time,

I understand the concept of pulling (i.e. one disc North, one disc South). You essentially 'drag' the magnet opposite to you because of this attraction. Then I have seen Halbach arrays in the internet but then again I do not understand the concept even by reading it in wiki so I decided to stay clear of it. You have the same magnets all over just the North Pole is facing elsewhere, correct?

The second thing that I do not understand is the back iron. Is is like a ring outside that the magnets touch ? The coupling will be exposed to dirt and moisture and I cannot risk rust issues. Maybe coat it ?
The idea om making it smaller helps but it is not critical at the moment as the size in these things is not the limiting factor.

Since circular pockets were easier to manufacture (see pic below) we have this arrangement at the moment which is bolted onto a shaft and transmits the torque.
I was thinking of testing the North-South-North arrangement (circular discs) with ten magnets and find the maximum air gap. The more the better for us because the part with one of the discs has to be removable, so you pull it out and then you put it back in if anything goes wrong, wanna service etc.

 

[dnleng]

Hi Ben,

Many thanks for your reply,

Here it is how it looks

The disc on the right is pure aluminium. The disc on the left is a pulley, with a circular slot and another disc (exactly the same as the one on the right) is bolted inside.

D.

 

[dnleng]

Forgot to mention that the gap is exaggerated in the picture... we are looking for 4-5mm.
I know that the more magnets the better, the higher the radius the better, the stonger magnets the better. I have to find a way of putting an iron plate at the back but then again corrosion issues may arise and we do not want that as I said. It would be helpful to have something at hand where you can calculate the torque based on these parameters were now I am going experimental by manufacturing things...

 

[RyreInc]

You're using an axial gap instead of the radial gap that I demonstrated, but that doesn't really change anything--they're topologically the same. My point is that you can lower the total weight by using back irons. You should be able to approximately halve the amount of magnet material you're using, and add considerably less weight than that in steel, for a given torque.

Learn about magnetic circuits, that will explain the purpose of the back iron.

As far as corrosion, I understand there is magnetic stainless steel out there, but someone else will have to chime in on that subject as I know very little. Magnets will corrode too if not coated!

(Also, the units I used should obviously not have been ft*in, but lbf*in! Doesn't matter anyways.)

 

[MagBen]

3xx ferrite stainless steels, but not austenitic stainless steel are good for the backing iron. Even low carbon steels has a better corrosion resistance than Neo magnets

 

[dgallup]

The 300 series SS are austenitic, not ferritic. You need to use something in the 400 series. 430F is very common because it is free machining. If you need higher hardness then 425 or 440 as these can be hardened. You probably need to nickel plate the magnets if you need corrosion resistance.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[MagBen]

Whoops! thanks for correction, dgallup! cold worked, unannealed condition of 430F gives even a better machinability.

 

[IRstuff]

How much torque are you trying to produce? By your image, I'd say that your probability of success is low. Your design looks similar to an auto A/C clutch, which typically depends on contact and friction to transmit torque.

TTFN
faq731-376

 

[dnleng]

The maximum torque should be less than 1.2 Nm which is the motor stall torque.
Actually I had success with 8mm magnets using 2mm air gap. The problem is that every time you want to experiment you have to manufacture things and buy magnets etc so a computer simulation about the torque transmitted would save the day. My goal was to use Design Of Experiments (DoE) to optimise the simulation runs, then run the simulations so I get the system behaviour and I find the optimum.

Indeed magnetic ss should be available, alhtough the design would need some minor changes but it should be ok. As long as the magnets get thinner because of this advantage that is something that I may have to do. The thinner the better. There is actually a limit because the smaller it gets the more difficult it is to manufacture. Anyways, this is a nice thing to consider.

All the magnets will be coated, either with a NiCuNi coating or black Nickel coating or epoxy. Aluminium parts will be anodised.

The torque should not be a problem. Imagine that the Pitch Circle Diameter of the magnets is 32mm (magnets as stated earlier d10mm by 5mm thick) so it is a small coupling but N52 magnets of this size are vey capable from what I have seen. The coupling will be used to roll paper, so you are getting the idea. Low speed, low torque... Of course when they touch it is very difficult to detach axialy. From what I have read that is one of the reasons that the N-S-N or the other configurations with the North pole changing directions. The axial load goes down so you do not put an enormous load on the bearings (correct me if I am wrong)

 

[dgallup]

If you use a radial gap like RyreInc showed I think you can have near zero axial force and fewer alignment problems.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[dnleng]

Indeed, radial gap would be betterm but the the removable concept goes bang. Imagine the the right pulley one the image is connected to a mechanism and you remove the whole thing by miving it upwards. The radial gap would create more design problems, at least for this application.

 

[IRstuff]

Why must there be a noncontact solution? A contacting electromagnetic clutch could probably be < 1/2 the size you've shown here.


TTFN
faq731-376

 

[dnleng]

Because the right side is attached to a geometry which is removable. The whole thing is placed in a box and the left side pulley that you see in the image is attached permanently to the box. The right side is attached to another geometry which is removable. The electromagnetic clutches that I have seen are both wider than this (in the image the distance is exaggerated for clarity) and the are not removable...