View Full Version : magnetic field question

08-20-2011, 11:08 PM
Here's a quick drawing of a magnetic field assembly. There's a flux ring, six magnets fastened to it, and an inner core of steel. Six air gaps are created, through which the magnetic flux flows. Loops of wire will pass through the air gaps and will be forced to move in a rotary direction when current is passed through them. This is a basic coreless electric PM motor.

When current is applied to the windings at the proper time, it pushes against the magnetic field in the air gaps, causing the rotor (which is not shown for clarity) to turn. Pushing against the magnetic field lines causes them to be displaced in the opposite direction to some extent.


Here's the question- what will keep the lines of force in place better, a solid steel slug in the center, or a slotted one like my drawing shows?

Another technique I've read about is to make a large number of narrower and shallower slots around the core piece. Supposedly, the lines of flux will tend to 'pin' at all the high spots around the core. If one were to machine a large number of shallow slots, the average air gap distance would increase, lessening the total number of flux lines between the magnets and the pole piece. The desirable situation is to have the maximum flux in the air gaps, and the maximum 'immovability' of the flux lines as well. It might make sense that there would be some kind of an optimum in flux/air gap architecture.

There are two ways such a motor can work. One is the field assembly, including the core, rotates as a unit, and the windings are stationary. The other way is the magnetic field assembly is stationary, and the windings rotate. If the windings rotate, there has to be some kind of commutation so that current flows in the appropriate coils at the appropriate times to make the motor turn. If the magnetic assembly rotates, the connections to the windings are fixed, but some electronics are required to feed current to the windings in proper timing.

Looking at ways to increase the flux strength in the air gaps, I can envision another set of magnets fastened to the core, opposite the existing magnets. It would seem that you would be increasing the field strength, and probably enhancing the ability of the field lines to remain in place when acted against by the current flow in the windings. I don't know if this is really true or not- on one hand you'd have double the thickness of magnet, and it might be true that a twice as thick magnet inside the flux ring would produce that same increase in field strength in the air gap.

One problem with magnets inside the flux ring and opposing magnets stuck to the core is that if you elect to have the magnetic field assembly rotating, you could easily reach a point where the magnets on the core fly off. But- you would not need brushes to ride on a spinning commutator. There is one more option, and that is to have the wired tube rotating, but equip it with slip rings to pass the current, and use electronic commutation as well. Having the wired tube being the rotor means the inertia is far less, and the rpms can be controlled much more quickly, rising or falling. This is often very desirable.

Because the brushes don't switch between segments, there would basically be no arcing, and no current flowing from commutator segment to segment through the brushes, so there should be much less brush heating and wear- the slip rings could be comparatively small and friction would be lower as well. The lifetime of the brushes assembly should be something pretty good- look at how long alternators can last between rebuilds requiring brush changes.

So, there you have some of my musings on the topic of motors. All comments welcome.

08-21-2011, 12:13 AM
Where are the poles on the magnets?

The Artful Bodger
08-21-2011, 12:40 AM
I will make an assumption that your magnets have poles on the inner face and against the 'flux ring', the logical arrangement I think if flux is to lie in the flux ring.

Adding magnets to the core which is to move in relation to the magnets on the flux ring cannot add any advantage that I can see, on the contary they would induce additional load.

08-21-2011, 12:49 AM
Poles are on the flat sides of the magnets. I have drawn six magnets, but it could be any even number that suits. It would likely be built as a 3 phase, and there would also be an appropriate ratio between the width of the magnet and the spacing between them. I'm sure the width of the air gap would be part of this equation. Where there's a steel rotor within a magnetic field, the gap would be smaller and the magnets could be closer together, but where there's a larger gap to accommodate the tubular rotor, the magnet spacing would have to be further apart.


The core never rotates independently of the outer ring in this type of motor. Where there would be magnets attached to the core, they would sit opposite the magnets in the ring, and maintain that relationship.

I've read quite a lot in the r/c forums, on the LRK site, etc. I've seen homebuilt motors where the magnets are pretty much touching, and others where they're at least a magnet width apart, sometimes more. I know there are a lot of factors that interact, and it's almost rocket science to get all the parameters right if you're a hobbiest building your own motor. Most of these homebuilt motors will run, and quite powerfully, but I wonder if any more than a few are set up fairly optimally- most are built for power anyway, so there's no real need for efficiency at lower speeds and lighter loading.

Personally, I like the idea of the coreless motor, which is what I've talked about in my first post. I'd like to build one myself, but I also want to get the physical parameters pretty close to optimum in the design stage, before actually cutting metal and mixing epoxy.

Paul Alciatore
08-21-2011, 01:06 AM
Assuming that the core piece does not rotate relative to the outer magnet assembly (or rotates in perfect sync with it), I would think that the slotted core piece you show in your first drawing would help to keep the magnetic field in place when the coils are rotating through the gaps. What I do not know is if it is actually desirable to do so for improved performance in the motor. It may just make the power pulses more pronounced and add a greater modulation of the speed of the motor. Said more plainly, I don't know which situation would produce the smoothest torque curve vs rotational angle. It may be possible to smooth this torque variation out by using a different number of poles on the rotating windings as opposed to the stationary magnetic poles - for instance, four windings or eight poles on the windings vs the six magnetic poles you show. The windings would have to be energized in a sequence as they pass a set of magnets respectively, instead of all at once.

Adding opposing magnets on the core piece would increase the magnetic field and assumedly the power/torque of the motor. It would probably increase it by close to two times assuming that the second set of magnets are the same strength as the first. Some experimenting would be needed to determine the actual amount and an optimal configuration. You would need to be sure that the inner core piece as well as the outer ring that the magnets are fastened to are capable of holding the total field you are producing with the magnets without reaching the saturation point. If they saturate, then any further increases in the number of magnets would be just wasted and may even become detremental as the excess field starts to spill out of the metal and into the surrounding air spaces. In short, you can't just keep adding more magnets and expect the magnetic field to increase in a linear manner for each such addition. When the magnetic circuit reaches saturation, each additional magnet will add little or nothing to the field strength.

Using a greater current in the windings would also increase the power/torque and may be easier to do.

08-21-2011, 01:19 AM
That's a good point, Paul. There could well be more modulation in the rotation if the flux lines were pinned. It could be that a solid round core would have the motor running smoother.

Having a different number of poles in the windings than in the magnets- hm, I'll have to think about that-

08-21-2011, 03:55 AM
You definitely want a different number of rotating poles vs stationary poles. If they are the same cogging will be very strong. Also, to compress the field lines at the poles make the centre core hollow so there aren't any magnetic circuits that cross the centre. The shorter the circuit the higher the flux density until it saturates. To prevent eddy current losses the core should be laminated and preferably the flux ring too. That can make a 50% difference in motor efficiency.

Even though the core is energized with a static magnetic field it isn't static when it is running because it interacts with the windings. Eddy currents will result in back emf that effectively cancels part of the field. Also, using silicon iron will provide much greater efficiency than plain steel or iron. You could probably find an old DC motor armature to do the job. Also, the gap must be kept to an absolute minimum. Close tolerances are very important to efficiency.

08-21-2011, 08:33 AM
This may be more than you are interested in getting involved with, but in case you have not come across it before there is a quite capable magnetic modeling program that you can download and use for free called FEMM (Finite Element Method Magnetics).

You can get it here:


There is a learning curve to be sure, but it is very capable bit of software, especially considering the price.

Just an aside and FYI that might be of interest.

Weston Bye
08-21-2011, 10:15 AM
To prevent eddy current losses the core should be laminated and preferably the flux ring too.

Not done in practice.

Laminated cores make sense for wound armatures and AC field coils where the magnetic field varies widely to the point of full magnetic reversal. However, the permanent magnet field does not change, or changes very little when disturbed by the passing rotor windings. Indeed, if the permanent magnetic field is varying much at all, there is something wrong with the design.

A laminated core would have to be more bulky to achieve the same magnetic conduction; the laminations have to be built with a thin electrical insulation on the surfaces to inhibit electrical eddy current circuits from lamination to lamination. The insulation layer is thin, but constitutes some percentage of the core that is not conducting the magnetic field.

Also, magnetism enters and leaves ferrous materials at right angles. Once in free air or within the ferrous material, the fields may curve. A laminated core in contact with the face of a permanent magnet will not achieve 100% contact with ferrous material due to the thin layers of electrical insulation and will suffer some saturation at the very corners of each lamination, reducing efficiency of the interface.

Hence, the solid core and flux ring gives superior performance for the amount of material involved.

Some devices make use of powdered metal for cores and flux rings due to the complex shapes involved - they are easier to manufacture. Powdered metal, though acceptable for many applications, does not perform as well as the equivalent machined-from-solid part in a DC circuit. I know this from experience.

However, for pulse or AC circuits powdered metal or more special types of sintered metal such as Somaloy may be superior to laminated cores.

Factors affecting performance for motors, generators, solenoids, etc:
Magnetic field strength
coil turns
coil amps
velocity (for generators)

All other factors are used to improve the above:
Air gaps
magnetic conduction (lack of saturation)
coil resistance

J Tiers
08-21-2011, 11:21 AM
What Weston says is dependent on this having the windings on a separate thin rotor.

Obviously if the windings were on the center core, and it moved as in a DC motor, you would use laminations, since the commutator would change the flux in the core. DC PM motors have solid shells, and laminated rotors.

The salient pole vs non-salient pole is probably LESS IMPORTANT here, if windings are in a thin "pipe" form between the inner and outer portions....

if the salient pole part turned, you would have severe cogging, but I understand you to be NOT turning that part, just a shell of windings. With just the shell of windings turning (or just the main structure) the sharp poles will be much less of an issue.

Weston Bye
08-21-2011, 11:52 AM
What Weston says is dependent on this having the windings on a separate thin rotor.

This is my understanding of Darryl's original post.

Here's the question- what will keep the lines of force in place better, a solid steel slug in the center, or a slotted one like my drawing shows?

There is the conundrum. All the center core is doing is providing a return path for the magnetic circuit - not that this is unimportant. Darryl's question is specifically about focusing or not focusing the magnetic field. The slotted core will indeed focus the field, but not that much more than a cylinder.

I think, more important in slotting the core is that it relieves the potential for any back EMF in the armature coils to induce drag producing eddy currents in the surface of central core as the coils pass close to the core between the desired poles.

This last effect can be easily demonstrated by dropping a magnet through a brass aluminum or copper tube. It will travel slower than an equal weight of unmagnetized steel, due to induced eddy currents causing localized magnetic fields in the metal tube that then interact with the movment of the magnet.

Weston Bye
08-21-2011, 12:28 PM
I think, more important in slotting the core is that it relieves the potential for any back EMF in the armature coils to induce drag producing eddy currents in the surface of central core as the coils pass close to the core between the desired poles.

This effect may in theory support Evan's suggestion concerning a laminated core. Indeed, good theory, but in practice it it easier and less costly to just increase the air gap where the magnetic field is not needed.

08-21-2011, 08:53 PM
Well, then it might be good to design the core as a thick-walled tube, and make it thinner between magnets, but not so thin that it can't fully conduct the full magnetic flux. Somewhere I read that steel needs to be 1-1/2 times as thick as a nib magnet to fully contain the flux. If that's true, the flux ring would have to be about .2 inches thick for a .125 thick magnet. The same formula would apply for the thinner parts of the core. But because there would be an air gap, the flux won't be as intense and the thickness of the materials could be less.

With an air gap thick enough for a tubular armature to fit in, it might be a workable rule of thumb to make the metal about as thick as the magnets would be. You could always add a flux ring around the OD if there was any magnetism detectable on the outside, and you could do the same with the ID of the core. Better to make it thick enough to start with, but at least this is something you are able to add after the fact.

This does of course bring up another aspect, that of the 'steel' used. Ideally, it would be an alloy that has the highest permeability to magnetic flux, in the direction that the flux would flow. Steel pipe would probably have the best flux conduction along the length, which is the opposite of what we'd want. Maybe that's part of the reason why some outer motor housings are made from a rolled section of steel finger jointed together- or maybe not. Could be that's just a way to form the housing cheaply.

May also be that the grain direction in the steel is not of much significance. Magnetic flux has to flow across the grain direction in every E part of a transformer core. Quite possibly the only thing of significance is to have enough of a cross section of core or flux ring that no magnetic field is detectable on the outside. If that requires a higher volume of material, the penalty is weight and size- that may or may not be an issue.

08-21-2011, 08:57 PM
Whatever type of iron you use for the flux capacitors make sure it is fully annealed. Any hardening of the metal disrupts the ability to conduct magnetic flux.

08-22-2011, 12:56 AM
Whatever type of iron you use for the flux capacitors make sure it is fully annealed. Any hardening of the metal disrupts the ability to conduct magnetic flux.

And make sure you never take it over 88mph without the time circuits enabled :D

Forrest Addy
08-22-2011, 01:28 AM
I thought thesse were called "shell motors" a bare cylindrical winding shaped like an opened tin can extending into an annular magnetc gap. Super low inertia but not very compact. The gap is the killer for power density and rigidity for right hand rule forces. Last I heard, a thin shell tends to scallop under the magnetic forces.

The rotor has to be commutated somehow to get power to the windings. It's a great design with some practical limitations. Real long small dia cored armatures work pretty well. They are inherently stiff and can have bearings at both ends. I've seen versions of the shell motor in trade pubs where the iron core is mounted on the shell shaft in bearings. The the shell windings can be supported on both ends but no there's rotary connection between windings and core.

The RC people use a rotating shell type motor where the three phase stator is in the center and a shell of Neo-whatever magnets rotate around it. They go like hell, maybe 100,000 rpm and have high power density.

Look at this one for example http://www.ebay.com/itm/New-RC-Timer-BC1712-13-2000KV-Outrunner-Brushless-Motor-/120501242307?pt=Radio_Control_Parts_Accessories&hash=item1c0e6f09c3 and imagine it scaled up to machine motor size. 2 million RPM. 120 watts input for 32 grams (a bit less than one oz) motor weight. If things scaled exactly a 2 HP motor could weigh 1 pound and run 120,000 RPM. Flat leather belt and cone pulley, here we come.

08-22-2011, 02:38 AM
The largest one of these model motors I've seen is about the size of a soup can and is rated at 15 horsepower. It would fit on my Unimat. I could probably get a spindle speed of about 60,000 rpm, and have enough torque to turn impossibilium with tibornite insert.

Some of this technology amazes me. It's been many years now since the 1 horse per pound figure has been surpassed with electric motors. A lot of this has been driven by NASA and by R/C hobbiests.

I was at our local air show today and watched the models fly. The jets were awesome and the electric models were certainly no slouches. One of them had multiple electric motors and amazing performance- had a really cool sound also.

I have built a few electric motors using nib magnets. One of them uses the same pole count as the magnet count. The detent torque is huge, but once running it's pretty smooth. It will reach full rpm for whatever voltage you feed it in about a millisecond (seems like it anyway). Another one is a converted shaver motor. Adding magnetic field strength slows a motor in rpm, so you have to add voltage to compensate, but the torque increases a lot. If you could equate the difference in the motor to the speed of shaving, you could have your facial lawn mowed in about 5 seconds!