Fly Electric!

CD-Rom motors

'CD-Rom' brushless motors are fantastic! They generate stunning power from very little weight. Combined with LiPoly cells they are unbeatable. I have made dozens of these motors for myself and friends, and more to experiment or understand the principles. This page probably tries to be more advanced than cover the basics. Other 'motor' pages on this site are as follows:
* Summary page (small and large motors)
* Machining tips (advanced techniques)
* Winding Density (advanced advice)

This page covers three topics: Some important principles, torquey motors and then high speed motors.


'Torque' is what makes 'outrunner' motors special. Torque on any motor can basically only come from combinations of:
1. strong magnets (eg: 'N' number, size and coverage)
2. efficient flux path (eg: rotor & stator size, materials, design & construction, air gaps)
3. greater diameter (more 'leverage')
4. many magnet poles
5. many turns
6. high current
7. gearing.
These motors are ideally suited to exploit all of these. However, the last two tend to have negative implications so are avoided in most outrunner designs.

Characteristics that increase RPM are the opposite of all the above (except current probably). In addition, RPM is roughly proportional to voltage so more volts = higher RPM.

Our controllers are three phase so the stator has to have multiples of three teeth. Magnets have two poles (N/S) so have to be used in even numbers. The number of magnets should be as close to the number of teeth as possible without being exactly the same (because then the motor will not run). However, there's no harm in trying greater or lesser pole numbers. Compromises usually have to be made. A really good motor is often the result of a lucky combination (or extensive design & testing).

The direction of winding and sequence in which teeth are wound is IMPORTANT and varies depending on the number of teeth and poles. Charts are available to illustrate valid combinations.


My single stator motors produce 12oz (340g) of thrust with an 8x4 drawing 7-8A from 3 lithiums. They weigh 0.8oz (23g) which is lighter than a GWS IPS motor and a quarter of the weight of a geared Speed 400. This power can easily result in a model that has over 2:1 power to weight ratio so flight performance is outstanding (see my Tiny Treasures). They produce similar power to a geared Speed 400 with much less current.

My motors are usually based on the bare 'gobrushless' 22.7mm nine tooth stator and matching rotor (can/bell). You can use any motor as the base but some cannot accommodate the magnets and wire most of us use. The GB components also help you replicate results exactly or make incremental changes on another motor. The 22.7mm stator is made from very thin stator sheets which is good for efficiency.

I almost always use twelve 5x5x1mm magnets (14 of 5x4x1mm are also good). Try to get N48 or stronger (the number indicates strength). For park fliers, 'Shock Flyers' and any 3D model these are orientated as 12 poles N-S-N-S... (where the '-' represents a gap). I have tried 10 and 8 poles but these motors tend to be noisy (with a star winding on nine teeth the magnetic forces are not balanced). I mostly use 0.45mm (26swg) wire in a Star configuration with 23 turns. However, I may be changing to 0.425mm wire to fit 24 turns. This reduces current by 0.5A which helps weaker Lipo batteries and controllers that struggle with bursts of power. As a principle thicker wire results in a more powerful and efficient motors. However, getting the right number of turns is more important than using the thickest possible wire.

Beginners will struggle to wind motors as described and 0.4mm wire will help. It is important to be VERY neat and concentrate VERY hard. I 'cheat' a little by making sure that I fit 10 turns on the first layer (see photos below). You must not normally wind more than three layers or it will not fit in the standard bell fully.

The teeth of a stator are rectangular but the wire takes an oval shape. On dense/tight windings you need to flatten the windings between teeth to create more space (particularly at the root). The side of a penknife blade can be useful but take care not to scratch the wire and damage the insulation. Hard wooden wedges are another useful tool (I trim bamboo skewers for this). NB: Remember to check all windings for shorts between each other and the stator before running the motor.

I turn an aluminium 'bearing holder' which is 9x23mm. This includes a 4x8mm section for the stator which is glued on with epoxy. I drill a 6.5mm hole through the centre and enlarge to 7mm at the 'inside' end and 8mm on the 'outside' for the 3x7mm and 3x8mm bearings. Smallerbearings are OK on both ends but are less tollerant of abuse. I make 3x40mm shafts from silver steel which are easy to thread at both ends with a 3mm die.

I mainly use APC 8x3.8 Slofly props outdoors. These props maintain their pitch and don't break often if you use a 'prop saver'/rubber band mounting. However, I prefer more flexible types indoors (eg: GWS 8x4.3) to reduce breakage (the APC are brittle and snap easily against hard surfaces). Many of the steps I follow are illustrated below. You will see from many of my newer models that the prop goes on the opposite end to the rotor. This is because the rotor is very soft and easily bent in a rough landing. If you do bend the rotor slightly, you should be able to remounted the shaft on the rotor with the cyano/sellotape technique (acetone disolves cyano). You need spacer washers ('straightened' spring/lock washers are ideal because they are smaller than the OD of the bearing) between the rotor and stator end bearing to ensure the windings do not rub against the rotor.

I either use a plastic saddle clamp across the bearing holder to mount the motor or turn a ali mounting ring. Both need a flat on the bearing holder to stop it rotating due to torque. The photos below show a 'low tech' approach to aligning the magnets evenly although these days I use gears on the lathe to serve as a dividing mechanism.

Bare components, stator and rotor

Score inside rotor at each 'dimple' (three times)

Glue three magnets at score marks
Use a flat surface to ensure flush with edge of rotor

Space out magnets with plastic spacers (~1.1mm)
Adjust thickness of spacers with selotape

I remove the central 'press fit' bush before
adding magnets; drill out to 3mm

'Squeeze' extra turns on first layer (10)
by 'overflowing' onto hammerhead

A little epoxy on windings inside rotor
prevents movement later

Wrap stator with sellotape to ensure exact alignment;
glue shaft to rotor/bell with epoxy or med.cyano/zap

Attach shaft with 3mm nuts with epoxy/med.cyano as filler
Don't tighten bolts too much; let epoxy/cyano hold position

A good way of mounting motors throught firewall;
protects bell from crashes!

I have been using a Tsunami 10 ESC on 'Mode 3'. I leave the timing at default ('auto') although have fiddled a few times with manual advance. The motors start and run equally well with Jeti Advance 08-3P, Castle Creations Phoenix and cheap Chinese ESCs.


I've tested a number of different combinations and ended up with two good Speed 400 (direct drive) replacement designs, both using Gunther props, 8xAR800 nicads, GB 22.7mm stator and standard bell in a fast glider (Sizzler). The first has more power than a 400 and a little longer flight time. The second is similar in power to a 400 but almost double the flight time.

1. More power/speed than a 400: 6 minutes at full throttle on 8xAR800's. 16 turns of 2 strands of 0.375mm wire. 14.5A static @ 15,600 RPM with RC2400's. Twelve 5x5x1mm magnets orientated in 6 poles (NN-SS-NN-SS-NN-SS). I have also had very similar results from 10 turns of 2 strands of 0.45mm wire and 12 magnets orientated in 12 poles.

2. Same power/speed as a 400: 7.5 minutes at full throttle on 8xAR800's with 19 turns of 0.45mm wire. 10A static at 15,000 RPM with RC2400's (8.5A with AR800's). Six magnets in six poles.

You will notice that there are 2 cans in the photos below. One is N-S-N-S-N-S (6 magnets) and the other is NN-SS-NN-SS-NN-SS (12 magnets). Both have six magnet poles. The best one is the one with more magnets (12). The 6 magnet version turns the same prop about 800 RPM less and draws 1.5A more (ie: the 6 magnet version is much less efficient than the 12 magnet one). I have also tried NS-SN-NS-SN-NS-SN orientation which drew even more current and yielded substantially less RPM.

Main components with the 2 cans

Closer view of magnet options

Mounted front view

Mounted rear view


Some test data from some of the many brushless motors I have made: Excel spreadsheet (63kb)

Top of page
Click on the links above for more info.
Click on the links below for other pages.

Home Art of the Possible Absence of Matter Links Updates Email