Electric Motion EM 2015 resurrection – The Motor.

I recently purchased a 2015 EM. the bike had minor battery problems and a click in the transmission, but the bike was in excellent condition. It had had an easy life. It had already had two BMS upgrades and would cut out occasionally. I had trained as an engineer in London and had built three electric bicycles including triangular battery packs to fit in the bicycle frame. I set about building a spare battery pack and resuscitating this six year old battery. I shall document each item rather than give a chronology. My purpose is to bring new life to all the other aging EMs around the world. Please email me on my errors and omissions. a n d y (at) c h a l k l e y . a u


This is a very nice Golden Motor 5kw, commonly described as an “axial flux motor”. This ‘axial flux motor’ consists of a metal disk of about six inches diameter with strong flat magnets stuck on one face. The magnets are like the very strong magnets used in computer hard-drives. These magnets are so strong that you have no chance of pulling them apart without some form of puller. Facing the magnets is a stator with numerous coils. There are three fat power wires from the stator down which three-phase alternating current is supplied from the ‘controller’. The controller is supplied with forty-eight volts from the battery pack. The controller is given a signal from the accelerator that ranges from 0 volts to 5v. The stator is held to the fan side of the casing next to your right knee by three 8mm Allen bolts. The the fan is missing. You will need that white thermal paste from an electronics shop before assembly. When disassembled, you will not be able to get the disk of magnets from the stator because the magnetism is stronger than you or me. If you can pull a Gas Gas to pieces and get it back together and still have a functioning motorcycle, you will manage the servicing of this Golden Motor. Do not assume that it is simple. The only serviceable parts are the bearings with the usual trials problem of dirt ingress on the drive side. Axial flux motors are great. They are narrow, cheap, and reliable. The main issues are:

  • getting heat out of them.
  • bearings.
  • the air gap between the rotating magnets and the stator.
  • total sealing of the unit.
  • condensation.


In trials, we tend to have short bursts of high output and prolonged periods of moderate speed. The same model of motor comes with water cooling, but it appears unnecessary. Similar motors from other manufacturers have air cooling holes, much like the casing of an automotive alternator. On some ebike hub motors, enthusiasts cut casing holes, whilst others use ‘fero fluid’. The fero fluid appears to have very fine iron particles in a tiny quantity of a light oil. It clings to the stator in the air gap allowing greater cooling. It may also give better flux in the ‘air gap’. I don’t think the trials engine warrants this and the gap on the axial motor is adjustable. You may wish to take a temperature reading and let me know what you get.

Heat is generated in the windings due to the resistance of the copper wire. So the grade and diameter of the copper is of significance. The controller appears to have a rating of 150 amps peak and the fuse is 300A. I have yet to determine what current we use at peak current. What is fascinating about these motors is that maximum current is at low revolutions. The forty-eight volt pack works against the winding resistance of the copper coils along with the ‘back emf’ that is generated as the motor turns. Maximum current is determined by ohms law and occurs when you open the throttle to full at zero rpm. Current drops off as the motor spins and is low when the motor get close to maximum rpm. Don’t expect to learn that immediately. So heat is generated when it is most difficult to get rid of heat. However, we use short bursts of power. We all know how to keep our car engines from boiling like a kettle. We have a new learning curve for electric motors.

The motor casing has ribs to assist convection cooling. As one might expect, the interior has no finning. Heat is a waste product of the motor and comes almost entirely from winding resistance. The heat path would be from copper, through the varnish coating, through more copper wires and varnish, through any insulating layer into the iron laminated stator, to the fan side cover of the motor. Some will find its way to the main body of the motor at the outer edge of the fan cover. Some heat will travel from the copper windings to the air in the motor, whence it should pass to the casing of the motor, then to the passing air. Under a dramatic increase in load for a prolonged time, heat will be generated and pass as as in the above list. If the heat generation is prolonged, all items will become hot and cooling will be impeded. Thus, there is a log between high load and over heating. The time scale is not known. Heat generation will be greatest at high amperage which tends to be a full throttle and low engine revolution speed. The characteristic of these engines is that they take less current as the rotational speed increases.


The motor uses regular ‘ball journal bearings’ with double seals like trials wheel bearings. Some axial motors, use an ‘axial ball journal bearing’ on the non-drive-side as that bearing takes the thrust of the magnetic attraction between stator and rotor. There is a suitable, bearing of the same size, but it does not have internal seals. There is no provision for alternative sealing. On some similar motors, a roller bearing is used on the drive side. I think the ball journal bearings are quite adequate.

I believe that the grease inside the bearing is suitable for the high speed operation of the motor. I did not pack them with water-proof grease as I do with wheel bearings. I will for warn you, the non-drive-side bearing has to be adjusted to give the correct ‘air gap’ between rotor and stator. I shall call the non-drive-side bearing the fan side bearing, although there is no fan. The fan side bearing has a threaded adjuster the diameter of the bearing. To lock it in place, it has a couple of small countersunk screws that clamp it in place by means of a split that is not visible. Stage one is to put two centre punch marks or drill marks on the adjusting ring and the case. Two marks will tell you that it was the original location. Subsequent marks can be with single marks. The adjuster ring has two holes to engage a double prong tool similar to the centre of an angle grinder. Using these is not a wise move. The force is likely to damage the threads in the aluminum case. Better is to make a puller. Six millimetre threaded rod is more than sufficient. Next time I pull the motor out, I will tap two 6mm holes in the thick aluminum plate of the transmission to take a home made puller. This can be a piece of 20mm square tube with a bolt and nut in the middle. Even a piece of wood would probably suffice. Warning: you must use a puller.

Air Gap

This is contentious and appears to be a guarded secret. After many hours of search over many days, I can find no definitive measurement for the air gap. The adjusting thread on the retainer is 1mm. I estimate that the air gap was 0.8mm or 1.0mm before disassembly. I believe I now have mine set to 0.5mm and have completed one trial with this air gap. If the air gap is too small, catastrophe could occur. If the air gap is large, there is less magnetic flux and less torque, although the motor may spin a little more freely. I don’t know my original setting as I marked the location with a felt pen! I still feel a fool for making this rooky mistake! The lines subsequently disappeared when I cleaned the items! The lesson: Centre punch the original setting with two sets of two marks so that it is obviously the original setting.


Besides major items such as rewinds, the only modifications that one might look at are:

  • Weight. The casing is not strong enough to start removing metal. The bearing retainer for the drive side bearing appears to be superfluous as we are not using it in a helicopter or aircraft. The drive side retaining ring could be dispensed with in my opinion. I have no memory of what I did. I am seventy two and happy to continue the occasional ride with my failing memory! The fan fins could be trimmed.
  • Air gap. Feel free to experiment. With a rudimentary puller, this could be done in the field.
  • Increased battery voltage.
  • Some experts talk about the phase difference between the hall sensors and the magnetic field from the stator.


With magnetism, a north pole attracts a south pole, a south pole attracts a north pole, but south repels south and north repels north. The rotor is a flat disk with powerful magnets stuck to the surface such that the surface has north south north south north south and so on. The stator has a similar number of poles each with copper coils around them. The controller supplies current such that the poles are magnetised south north south north south and so on in a manner such that a north on the stator is slightly ahead of the corresponding south magnet on the rotor. With all the souths slightly advanced on the norths and all the norths slightly advanced on the souths, the rotor is pulled round to get norths to meet the souths. However, the clever controller, moves all the poles on the stator round to the next poles to give continuous motion to the rotor. This is why the controller can cost as much as the motor and plays an important role in motor characteristics and performance. One might also see that the ‘hall effect sensors’ detect the position of the rotor and thus determine the ‘advancement’ of the magnetic field in the stator. For a tuner, the behaviour of the hall effect sensors is of interest and alternative arrangements may be possible. The electrical engineers amongst us would handle this better tan myself.

Even more complex is the following. When the armature with its magnets is rotated, it produces a voltage in the windings. It thus acts as a generator like the alternator in your car. The voltage tends to be dependent on the speed of rotation. This generated voltage is commonly called ‘back emf’. In most situations, this back emf is not great enough to exceed the battery voltage. So when the throttle is closed, the motor tends to freewheel with just minor losses, giving minimal retardation. However, clever engineers played some tricks with the controller circuitry. If you press the ‘regen’ button, the motor acts as a generator giving significant braking power. Some controllers have on/off regen and some controllers have a progressive function like a reverse throttle. To make this work, the controller collects the available voltage and boosts it so that it charges the battery, hence performing work and adding drag and braking force. One distraction is that the regen braking tends to drop out below a certain revs. Controllers can give you control over this and many other settings. I have modified the location of the regen switch on my handlebars so that my hand does not jump off the handlebars. I have the regen button under the bar and operated by my left thumb. My ‘regen’ braking drops off suddenly when the revs drop too low. However, if i repress the button, regen reactivates. It is something I have to live with. Regen puts energy back into the battery so I use it on all downhill situations. I will make a left side ‘regen’ brake lever for my left foot to accompany my thumb operated ‘regen’ button.

The Strip Down

The main casing:
It is of lighter construction than it appears before you take it apart. It may benefit from an extra spacer between the shell and the two large aluminium plates that form the transmission to aid stiffness. The bearing location shown is the drive side bearing. Cooling is important for an electric motor. A trials bike gives short bursts of high power. There is only air cooling inside. It needs to be consisdered but it does not seem to be a problem. the face between this outer-casing and the front face palate with the stator should surely need to pass heat. The joint surely needs to be sealed but also pass heat.

Of interest, there is no breather. If submerge your warm motor in cool water, there will be a contraction of air with the possibility of sucking in air through any crevice. I did not fit a breather on my motor. I would consider a breather and a small computer fan to ventilate and stop condensation. All parts were covered in silicone spray before assembly.

The face plate:

This carries the internals. It is not fitted with a fan although there is physical space for a fan. The bearing takes special consideration. Take care! it carries an axial load which is created by extremely powerful magnets. The face magnet has exceedingly strong magnets. Mine were attracted to the face stator with great force. The adjuster ring shown has a fine thread of one millimetre. Use a pair of centre punch marks to show its alignment before you do anything. I used felt pen which disappeared in the solvent. I disgraced my mechanical prowess! One turn is obviously one millimetre. My memory of its position is t hat a one millimetre gap is standard. These seems to correspond with similar motors. I reduced mine to half milimetre. I did get improved power and torque but I also changed the throttle and mapping settings at the same time. You need to reduce the load on this adjuster before you can undo the adjuster ring. That is the reason for the puller shown in the next image.

I found no maintenance or rebuild information anywhere and rebuild information is scant for other similar motors. Specifications for the air gap may be crucial but I could find no practical or scholarly articles on the subject of air gap. I based my guessed gap on my assessed rigidity of the assembly and the likely affects of wear.

The puller:

I happened to have this puller in my workshop. You may have to make something. The puller is only pushing against the magnets, but that is significant. The tension needs to be taken to avoid damage to the fine threads. When tensioned, the adjuster ring will move and has normal thread. I happened to have this ‘one-size-fits-all’ tool for angle grinders. I had to bend it slightly to clear the body. This is not something for a novice as reasonable care is needed. You are not going to find spares if you stuff up!

The angle-grinder tool:

My friendly local bearing shop had the required bearings. I probably left the outer seal off the thrust bearing and packed the bering with my own grease. In trials, we have slow speeds, so greater quantities of grease in bearings helps to stop problems.

The fine threads:

The terminals. Ensure they are adequately protected and acceptably clean. I had to pare some plastic covering away to ensure appropriate contact. Take a picture of the connection positions for reassembly or you motor may run in reverse when you reconnect it!


The bearings on the drive shaft appear to be inadequately protected from dirt ingress. Mine had a previously kind life but still had dirt issues. Do not expect them to last. These drive parts may become difficult to source so look after them. They are a bit of a nightmare to replace as it requires a major disassembly including swing arm and motor removal. Nothing in the transmission is well sealed.

Here is the belt. I would order a new one. It appears to be a standard belt that should be commonly available.

Here are the controller numbers. My reading of the available information is as follows. Mine had a serious setup error.
The controller in standard form has a trigger voltage from 0 volt to 5 volt which is supplied as the signal from the throttle. 0 volt is no power. 5 volt is full power. To the best of my knowledge, all common throttles in use on motorcycles and bicycles supply a 5v for full power. All seem to have a linear transition from 0 to 5 volt, which to me is inappropriate.

Now for the anomaly in the setup of this model of EM. Unless I am corrected, for use with the EM mapping unit, the controller was set to give full power at 3 volt. This is unusual. The mapping unit contains diodes. Diodes allow electricity to flow in one direction, which is not of issue here. Diodes are used because they give a voltage drop of 0.6v. Switching between different diode configurations allowed the maximum voltage to be altered to give different mapping settings. That is fine, but, a big but, the controller was set to a standard setting of 5v for full power not 3v for full power!

There is some way of hooking this controller to a computer to alter its settings. I may have the appropriate cables.

This issue can be overcome by simply removing the mapping unit as the mapping unit is in series in the wiring setup. Removing the mapping unit allows a full 5v from the accelerator to the controller. Bingo, much more power.

I thought I took a few more pictures of this rebuild. If I did, I have yet to find them. Here is my controller connection diagram:

The swinging-arm has shims that were inappropriate. They were too thin and they were if inadequate diameter. I made my own. The increased diameter helps to restrict the ingress of dirt and water. There are no grease nipples! I got a suitable hole expander an ground away the excess. I clamped thin aluminium or whatever between plywood and drilled through the layers.  This gives me a clean cut.

The modified cutter:

The rough and ready full-diameter shim-spacer:

Good luck!
I am half way through creating a new battery. I have made a new aluminium box. I will shortly spot weld the cells together. I will document it in another article.

Back to bearings. The drive bearing on this motor is a standard ‘ball journal bearing’ similar to a standard wheel bearing. Some similar motors use roller bearings as sometimes found on crankshafts. This motor has an ‘axial thrust bearing’ on the stator side. Be aware that some similar motors have standard ‘ball journal bearings’. I will stick with ‘axial thrust bearing’ on the stator side. Some consideration must be given to breathing of the motor. The motor is sealed which may cause water and dirt to be sucked in when rushing into deep water. In future, I may run a breather to a hidden high location. I would probably run two breather tubes so I could blow warm air through the motor.
Andy Chalkley

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *