DIY Electric Bicycle Conversion (Using a Brushless Motor)

We have recently converted our bicycle into an electric one, the conversion was a success but since we have used a geared brushed DC motor the top speed was miserable and this thing ended collecting dust in the corner. But not anymore as its time to give this thing a punch of power using a brushless motor and some crazy high discharge lithium pollymer batteries.

As we first converted the bicycle into an electric one our goal was just to get rolling so for the sake of simplicity we have used a geared Brushed DC motor but as we replaced it with a brushless motor, that third wire has brought some complexities to the whole project, so hang in there as we are going to address all those things as we get our hands dirty...

The list of tools and supplies for this project might go on and on but dont restrict yourself by the tools you have, the best one is right inside your head.

Materials:

• A bicycle (This one is a 20in BMX copy with fat tires to sustain those speeds)
• A Brushless motor from a hoverboard
• A sensorless BLDC motor speed controller from RC car
• Throttle assemble
• Two 11.1v 5200mAh lith
• Spocket
• Chain
• Footrests

Tools:

Drill machine and bitsBasic pliers Soldering tools A 3D printer (Just if you want to built customized 3D printed parts)

The first step is to rip off the previous setup, it was based arround a geared Brushed DC motor that are used in automatic gates. Lots of torque but not fast enough. Besides that those brushes make the efficiency of these motors down to less that 40% which is miserable. That combined with a gear box made things so horrible that we barely seems to cruise. Besides we have used a customized brushed DC motor controller that was powered by a customized lithium ion battery pack using recycled 18650 lithium ion cells from old hoverboard batteries.

The motor drives the rear wheel with a sprocket attached to the discbrake with the motor mounted on the bike rear frame on the left side.

Once we took everything apart, we were left with the stock bicycle except of the four mounting holes drilled in the rear frame and a spocket welded to the disc brake, well we dont need to worry about them as for this conversion they didn’t hurt us much.

For more details about the previous conversion do visit the instructables blogspot and the built video over our youtube channel.

To step up the game and give this bike a punch of power we have decided to replace the inefficient geared Brushed DC motor with a Brushless outrunner motor. Now that might sounds alot of customization, as those expensive motors and controller that can flush out your wallet but dont worry we are playing smart here. So we are going to use a brushless motor from an old hoverboard.

Now this beast is an outrunner brushless motor. Due to the design of these outrunner motors, the motor outer casing (having strong neodymium magnets) rotates, while the stator with the winding remains stationary. This means no brushes as the name says "Brushless" and thus all the loses related to that can be gained and the increase in torque eliminates the need of an additional grea box which gives these brushless motors efficiency of above 80%. This thing has proved to perform exceptionally well in most of our previous projects and has got massive amount of torque and hopefully thats going to get the job done.

Besides using this motor from an old hoverboard not only helped us to save money but compared to the hobby grade brushless motors they are larger in size and can handle more stress for a longer duration of time.

We opened the motor by undoing the six screws provided over the rear plate and removed the hoverboard rubber wheel. Now these hoverbaord motors comes with hall effect sensors that can work with a sensored speed controller for more accuraate motor movement. But to keep thing simple we are going to rip those sensors off as we are going to use a sensorless speed controller. Now thats where things started to twist a bit.

To drive the rear wheel we are going to use chain drive.With our experience from the previous conversion, we needed somewhere arround 16 N.m of torque to get the bike rolling. Now this brushless motor offered arround 13 N.m of torque and thats great as we only need very low amount of gear reduction to get the required speed thats arround 50 km/hr. Since this motor offers 16 KV means it rotates 16 times per volt of potential difference applied accross the motor. As we are going to use a sensorless RC car speed controller that can handle 22.2v max so it means that we can get arround 335 RPM.

Now with a 20in bicycle tyre, theoritically we are going to achieve a top speed of nearly 40km/hr with the current setup and a gear ratio close to 1:1 as we have got sufficient torque and later as we start rolling we can use the rear gear shif assembly to increase the speed of bicycle.

But there comes the twist, as we are using a sensorless speed controller from RC car, it fails to run the motor at such slow speed as we pull the throttle and suddenly the controller senses a stall and thus stops the motor from running.

For a while we have tried a bunch of thing to eliminate this problem, but nothing seems to help as the speed controller needs the motor to spin at a minimum RPM to complete a feedback loop and thus keeps the motor running.

So later on we settled for a 13 teeth spocket thats attached to the rear plate of the BLDC motor with the help of allen screws across a metal plate.

As we tried our luck with a bunch of gear ratios to achieve sufficient RPM to keep the motor spining, we settled for a whooping 52 teeth rear spocket thats modified to attach to the rear wheel using the disc brake adapter. Now that gear huge rear spocket increased the torque significantly but decreased the speed of the bicycle down to nearly 15km/hr.

Now with all those limitations we were stucked and the only way out was to rewind the motor. But this BLDC motor offers a preety compact stator and to increase the speed of rotation we have to decrease the number of turns over each pole while increasing the thickness of wire. That might sounds like a piece of cake but every single time we tried winding that motor we got the enamalled copper wire shorted to the stator body due to the limited space.

The only way to save this project from getting a piece of junk is to get our heads rolling.

Yup we can't rewind the stator but we can rewire it, Yes as its a three phase brushless motor with a total 27 poles stator. It means that each phase has nine poles through which the enamalled copper wire runs. Now the goal is to inrease the speed which can be achieved by decreasing the length of wire that makes up each phase. One way to achieve this is to rearrange the coils in parallel. Each phase has nine coils having three sets across the stator with each set having three coils. Currently all three sets are connected in series so it takes the current much longer to complete each cycle. But if we connect these sets in parallel we can get the current through each phase within one third of the original time.

Doing so we can get the motor spinning three times faster as compared to the stock winding patern. So its time to figure out a way to get this idea implemented.

To get the science implemented thats discribed in the previous step we have designed our own customized Printed Circuit Boards. All we needed to do is to spot the position if the coils that needs to be connected in parallel and finilized the layout of PCBs.

Here I would love to thank NextPCB for making this customization possible so flawlessly. As we designed the PCBs we headed to NextPCB website. The process to order the boards is preety straight forward. Just upload the gerber files and as you go through the option provided there you are all done.

They are providing a vast range of services for manufacturing your customized Printed Circuit Boards and their assembly throughout the globe at a very reasonable price. So be sure to check out their website for some great quality PCBs.

The PCBs arrived within just a week and as appears in the pictures the quality was really good. Schemati and Gerber Files can be foun here.

Now as we are going to use a RC car BLDC speed controller that can handle only 22.2v, so with our calculations we need to increase the speed of motor by almost three times to achieve a significant speed. One way to that is by rewinding the motor with a thicker gauge enamalled copper wire with lesser turns but not only does it takes alot of calculations but with such narrow slots its almost impossible to rewind this motor as everytime we tried to do that, we ended up scratching up the enamalled coating and thus shorting the motor winding.

The other way is to rewire the current winding. The stator has 27 poles, since its a three phase motor so each phase goes through three sets and each set has three poles. Currently all three sets of a phase are connected in series so it take more time for the current to complete a cycle and thus slowing the rotation of motor for a given voltage. Now we are going to rewire the stator with all three sets connected in parallel and hopefully we can get the current to get through each phase within 1/3rd of time achieving three time as much speed as of the stock winding. Ealier in our Drill Press conversion we have done that rewiring using a bunch of jumeper connection but this time we have to repack the motorso its time to improvise.

To mount the motor the best place was where the paddel assembly goes, so we removed the unwanted paddle assembly, machined a simple metal coupling for the motor shaft. The coupling offers two different mounting holes. A pair of which has 8mm gurb allen screws that holdes the motor shaft i.e the stator in place along with the coupling ans stops it from moving, while the other four threaded holes are for mounting the coupling within the paddle housing.

So we drilled four 8mm holes within the paddle assembly housing and mounted the motor along with the coupling in place.

To rest our foots we have made an assembly using a 12mm thick metal rod cut to the desired length. The metal rod is going to be held in place with a 10mm thick metal plate with threads in the centre. We have used an 8mm allen screw to hold everything in place as we count the footrests over the post right behind the motor mount.

One we painted and mounted the footrest assemble we then placed the rubber footrests over the metal shaft.

Due to the limited space we are going to use a pair of high discharge lithium pollymer battery packs. Each one of which is ratted for 11.1v and 5200 mAh with a peak discharge of 25C which means that these batteries can provide nearly 130 Amps of continious current draw and thats helpfull as we are going to need sufficient power to be delivered to the brushless motor. Now these two battery packs are connected in series to provide 22.2v to the speed controller and thus provided a 115 watts hour onboard capacity.

The next thing is the speed controller that we took from an old RC car. The speed controller is capable of delivering upto 150A current with a max volatge of 22.2v. We have used this speed controller as we can't get the high voltage Electric bike controllers (that might have solved alot of problems) due to the COVID-19 lockdown but that challange might now have taught us this much about the brushless motors and thier speed controllers. Anyways, we later connected the speed controller to the motor and the battery packs. To operate the speed controller we are going to use a servo tester. It offers a physical dial, dialing which we can increase the speed of rotation and vise versa.

Now with that being done we are ready to roll.

So reconverting the whole unit from a Brushled DC motor and a customized lihtium ion battery Pack worth that?

Yes for sure, we are able to get this thing crazy fast up to arround 40 km/hr. As due to the limitations of the speed controller we are unable to achieve a lower gearing ratio that could have enabled us to achieve a much higher speed but on the other hand the higher gearing ratio provided us with a whole lot of torque that can literally flip the bicycle over on demand. Now this whole thing is way more powerfull than a brushed geared DC motor and efficient too. We are able to get nearly 4-5km of range on a single charge, I know thats not enough but we are soo going to address that problem in our upcoming project so stay tuned for that.

Have a look at the upgarded version of this bicycle conversion that can pull 60km/hr of top speed with a 20km range on a single charge.

Dont forget to vote us in the Finish it Already Speed Challange.

Regards,

DIY King