Electric Motors – part 4
          Brian Muldermm


This month we look at the stator / magnet air gap and how to position magnets accurately.

Firstly, let’s look at how the “air gap” affects the motor.

The air gap is the distance between the rotating magnets and stator.  The motor builder’s aim is to get the smallest gap possible without the magnets touching any part of the stator.  The closer the magnets get to the windings, the more effective the magnets’ field becomes.  As we have already discussed in previous articles, a stronger magnetic field will result in a lower RPM/volt figure and can have an impact on the motor efficiency.  Increasing the air gap will obviously have the opposite effect, something we do not want.

When deciding on what magnets to use, something worth considering is how the air gap is affected.   For example, a CD rom motor is to be built using 12 magnets.   After a few calculations, and looking at what magnets can be purchased, it would appear that 6mm wide by 2mm thick magnets are ideal.  If you look at fig 1 carefully, you will notice that the air gap is not consistent and the closest point will be at the centre of the magnet and then the air gap grows significantly towards the magnet edges.
Not ideal, but the motor will still work effectively.

One way of improving this is to use 2 magnets for each magnet pole as shown in fig 2.  Now the air gap is more uniform and one can expect better performance.  There has not been a lot of talk regards better use of magnets for the air gap, but my experience has shown that the more uniform air gap does give a significant improvement in performance.

Using two magnets for each pole does require additional work though,  and depending on the application or your target goal, it may or may not be worth the effort. 
That one is up to you to decide.  Normally though, this decision is based on what magnets you have at your disposal.

Lastly, there is still another option.  You can purchase curved magnet segments.  They are available, but only in certain sizes.

There is also an added bonus when using these magnets and that is bonding coverage.  You in effect have a greater area to bond / glue your magnets to the bell, which in turn provides better heatsinking for the magnets.

All of this however comes at a slight increase in cost over conventional rectangular magnets.

Magnets and Heat
Seeing that I have just mentioned the heatsinking ability of curved magnets, a quick comment about Neodymium magnets and temperature.  For our motors, you typically get 80°C and the higher 120°C magnets. The stronger the magnetic material gets, the lower is the cutoff operating temperature.  Exceed the specified temperature, your magnets lose their magnetic properties and result in a motor that is pretty much useless.  During my own learning curve, I had wondered if the 80°C specification was a little too low.  In reality, a motor that gets so hot as to warm the magnets anywhere close to this temperature has a problem.  Yes, some motors are designed to run hot, but for our outrunner motors, this is not an issue.

Positioning Magnets
If you thought winding a stator took time, well adjusting the magnets uniformly around the inside of the motor bell can be even more time consuming.

When first attempting to do this, one would typically place a magnet in the bell and then try to position another close to it.  The aim would be to work your way around the bell in this fashion.  Problem is . . . it will not work very well and frustrate you.  The moment you insert another magnet anywhere close to an existing magnet, the two will attract each other and snap together making it difficult for you to get them apart.  And if you are lucky enough to get one in a good position without it jumping up next to some other magnet, don’t worry . . . just give it a little time and it will.

There are a few tricks that help when playing with the magnets.  The first thing is, get an estimate of the gap that will be evident between the magnets.  Now find some thin spacer material slightly thinner than the calculated gap (balsa offcuts work fine) and position it between magnets as you insert them one at a time.  This prevents the magnets snapping together and makes them easier to move into position.

Once all of the magnets are in, they will be much easier to work with, as each magnet is now being pulled from both sides.

Now, to position the magnets accurately, the easiest way would be to have spacers of the correct thickness between them.  This sounds easy enough, but let me assure you that it is not.  I went to the trouble of machining extremely accurate spacers only to find that the last magnet to be positioned had either too small or too big a gap.  Bear in mind that magnets also have tolerances on their dimensions.   So if you are fortunate to get magnets spaced nicely with some form of a spacer, consider yourself very lucky.

So how do we do it then?

Well, prior to inserting the magnets, a jig needs to be made.

I start by printing out a sheet of paper with a series of lines spread like the spokes on a bicycle wheel.

Use whatever software package suits you for this.  For 14 magnets, my sheet of paper will have 14 lines drawn in a circular fashion as per photograph.  Basically, a circle divided up into equal segments

The sheet of paper is then taped down to some scrap wood and a hole drilled through the centre.  The diameter of the hole should be the same as the motor shaft thickness, as the motor shaft will be inserted into this hole.  Ideally, you would want to use a drill press to drill the hole as it is important to get the hole perpendicular to the base, as the motor shaft must stand straight up when inserted.

The bell assembly, without magnets, can now be mounted on the shaft.

Next step is to create a positioning jig.  For this, I use any scrap plywood I can find.  One end has a hole drilled in it using the same drill bit as the shaft diameter.  On one end I have made up a plywood wedge and the other end some scrap balsa.  The following photograph shows how the jig is going to work.

The jig must be trimmed to the correct dimensions so that it can move around the bell without hindrance, and it is helpful to mark one side of the “inner vane” to remind you of which side will locate the magnet.  The outer balsa should be long enough so that it just touches the paper.  This makes it easier to position the jig accurately against each line on the paper.

Once the jig has been completed, you can insert the magnets into the bell as described earlier.

Now extract one of the temporary balsa spacers and slide the jig between the magnets.  With the bell firmly positioned, use the marked face of the jig to place a magnet in a straight up position and adjust the bell or magnet so that the outer vane is aligned with a line on the paper.  Once done, drop a dab of cyno onto the magnet to secure it.   Make sure the cyno is just enough for the magnet — you do not want the cyno getting to the balsa spacers!

Now slide out the next spacer and move the jig to that position by sliding it up the shaft, over the magnet and down.  Repeat the procedure for all the magnets.

A few pointers regarding this procedure.  The bell can land up rotating in the board.  You should try to get a tight fit to prevent this from happening.  If the bell does move though, simply position the jig up against a ‘set’ magnet and rotate the bell so that the jig is re-aligned with the paper.  I tend to do this a lot as the bell always moves from time to time.  I also place a tiny piece of masking tape to mark the last magnet set.  This helps to keep track of where you are when you re-align the bell.

And there you have it.  Accurately spaced magnets within the bell!

Machining a Bell Housing
Whilst we have been working with the bell, one of the most common questions asked about motor building is what should be the thickness of the bell and what material to use.

Well, if you are modifying an existing CD rom motor, you will already have a bell.  They are pressed steel cans that are quite thin.  Possibly too thin though.  Anyway,  the function of the bell is basically to concentrate all of the magnetic field lines inside the bell.  Any magnetic field that penetrates outside the bell could be considered a wasted resource.

Now based on what I have just said, it should be rather obvious that the bell material must be magnetic.  Raw iron is typically a good magnetic material, but we do not really get it too easily in this form.  In South Africa, we use a lot of mild steel, which at the end of the day is made up of iron and a tiny amount of carbon.  Local scrap yards and metal suppliers have an abundance of mild steel pipe offcuts which work just fine.

There are other better materials to be used (I think silicon steel) but the mild steel seems to be okay.  When searching through a scrap yard, you often find chunks of solid steel bar lying around.  Sometimes you can be unsure of the actual material — my rule is . . . if it has a good deal of rust on it, it is good to use.

When machining a bell, the wall must be thick enough not to allow any of the magnetic field to escape.  My own trial and error has resulted in my rather machining a wall that is too thick than too thin.  Experience has shown that between about 0,5 and maybe a bit over 1mm should be fine, depending on the magnets being used.  For 1mm thick magnets in a CD rom motor, you could get away with 0,5mm thick wall . . .  Maybe even thinner.  For thicker 2mm magnets, you will need to up the wall thickness.

There is a simple rule that has been used by motor builders on the net.  It simply calls for the paper clip test.  Position a few magnets in the bell and then see if you can get a paper clip to hang on the outside.  If the magnetic field is strong enough to hold the clip, the wall is too thin.  A simple test that seems to work okay.  As a side note, how many of you remember seeing those nasty old speed 600 (or 400) sized motors with the additional  metal ring on the outside.  I think they were called flux rings.  Well now you know why they were put there!

Lastly, people tend to machine the bell as thin as they can go, in an attempt to build a very light motor.  Sometimes, a motor may be replacing an oily in which case, the electric conversion  may result in the plane being tail heavy.  Now you have to add more nose weight.  Well why not do it properly and rather machine a thicker bell.  A win-win situation.

To finish off this issue’s article,  lets have a look at Emanuele’s CD rom motor, which he built recently.


The stator is a stock standard 24mm diameter one from a CD rom drive and will be in the order of  4,5 mm thick.  The windings have been very well laid out using 0,4 mm thick wire.  To neaten the assembly, he has used some printed circuit board turned into a suitable sized disc on a lathe.  The copper has then been removed leaving islands to which the ends of the windings have been soldered.

The bell has been machined on a lathe, in this case from solid steel.  Normally only the ring is steel and the front end that holds the pin, aluminium.  This is done for the weight advantage, but I doubt it makes that much of a difference at all.  Aluminium is simply more easy to machine on a lathe.   The only thing that still needs to be done though, is to drill some holes in the front to allow for air cooling of the coils.

Magnets used here are 12 off  6 mm square and 1,4 mm thick.

Emanuele reports that tests revealed that, using a 9x3,8 prop,

7-cell 800 mAh NiCd gave  270 g thrust at 3,6 A
2-cell 1000 mAh LiPo gave 225 g thrust at 3,2 A
3-cell 1200 mAh LiPo gave 380 g thrust at 5,2 A

*  *  *

Next issue —
Determining/measuring Motor Performance

- - - o o O o o - - -