Tom Stanton is an engineering student, and a few years ago he became interested in learning how to use CAD / CAM programs. As an exercise, he decided to design and build a light-weight and compact drive system using small RC components. I think he ended up with a great design, so let’s take a stroll down the path that produced several versions, with each one being a little better than the last.
The First Version, V1
Back in 2011, I designed and built a friction drive for my ebike, and it also used parts from the RC market (Radio Controlled), like model airplanes and quadcopters. I was able to achieve 1,000W of performance, but in spite of all the experimentation going on at the time, there was definitely room for improvement.
Today there is a much greater selection of parts, and the performance is much improved. As you will see below, the RC parts that were designed for small model aircraft do not work very well under ebike loads, which are much heavier, and accelerate much more slowly (leading to voltage ripple and voltage spikes).
However, the one area where RC components shine is in how small they are, so Tom kept upgrading the design to find out what it would take for RC parts to survive on an ebike. The controller that Tom ended up being happy with (after blowing up several cheaper models), was developed for a high-performance electric powered skateboards.
I was immediately impressed by Tom’s use of CAD (Computer Aided Design), and also the CAM (Computer Aided Manufacturing) which includes a 3D printer, and a CNC mill. You would draw a complex part on a computer screen, and then a machine makes it. Better still, once it is designed and tested, anyone can download a copy of the CAD file and have a local service print it out and mail it to you. Tom attends the University of Hertfordshire in the UK and he used a 3D printer to make the professional-looking parts that mounted the motor to the seat-stay, and to also make the custom pulley (The tiny drive pulley is a stock aluminum unit).
Tom’s initial 3D printer was a modest size, and he could not make the pulley for the rear wheel in one piece. He decided to produce it in four sections, and then glue them together.
Tom decided that for his first ebike, he would use a lightweight fixed-gear bicycle (a “fixie”). He felt that it would be easy to lift and carry indoors at the University, and sometimes up stairs. One of the options for a fixie is a flip-flop hub, which allows the rider to have two gears available without a derailleur, or an expensive IGH (Internally Geared Hub).
Another option when using a flip-flop hub is to have a fixed gear on one side (which means that if the wheel is moving, the pedals are also moving, and you can even pedal backwards), and it can have a single-speed freewheel on the other side. In that case, you have to stop the bike and flip the rear wheel over to change gears. Some riders have a fixed gear for riding on fairly flat land, and then a freewheel with a lower gear for climbing hills. They want the freewheel to have the lower gear, because if you go up a steep hill, you may want to freewheel down the other side of the hill (that way, the rear wheel is spinning very fast, but your pedals are not).
I added this explanation and pic to explain why the rear wheel on Tom’s V1 bike has a threaded mount on the left side, which he felt would be useful for mounting the pulley.
In the pic above, Tom is showing the 63mm diameter RC motor (2.5-inches). They are roughly the size of a small cup of coffee (but since he is in the UK, would it be like a cup of tea?). This style is an “outrunner”, which means the core is stationary, and the outer shell spins.
In the pic above, the tiny device shown is an “Electronic Speed Controller” (ESC). These are very different in size from standard ebike controllers. The black plastic mount was custom-designed and 3D-printed by Tom to fit this particular frame, in that exact location. The aluminum fins are the heat-sink which is designed to absorb and shed the heat that is generated by running the system. Unfortunately, this affordable model proved to be too small to survive.
For those who may want to try something similar, mount the ESC as close to the battery as possible, and the two wires to the battery (the red and black wires shown above) should be short and fat. The three wires that connect the ESC to the motor can be fairly long without causing any problems. Also, if you want to use RC components on a bicycle, search “adding voltage ripple capacitors to an ESC” to see how to do that. The additional capacitors should be rated for a voltage that is at least 10% higher than the voltage of the battery when it is fully charged, and the added capacitors should be “low ESR” (Equivalent Series Resistance).
Tom wanted the first system to be as compact and light as possible, so he used a LiPo pack (Lithium Polymer) that used six cells in series (6S) and 5000-mAh (milli Amp hours). If you operate the battery between our recommended 3.0V per cell to 4.1V, then a fully-charged 6S LiPo pack is 24.6V, with a nominal “average” voltage of 22.2V
I have used LiPo before, and I now prefer the high-current 18650-format cells that are available. I can’t help but to notice that Tom also switched to 18650 cells below.
Before running off to buy a bunch of RC components, be aware that Tom used an affordable off-the-shelf hand-throttle from an electric scooter, which sends an analog signal. The ESC’s normally use a digital input signal, so a friend of Tom’s put together an Arduino-based translator between the throttle and the RC ESC.
The Short-Lived V2
Tom was very excited at the performance of the first version, but after just a short while he also realized (as many do) that having more power would be nice. The first order of business was that it’s top speed was higher than necessary, and the acceleration up to top speed was too slow. The small pulley on the motor shaft was as small as it could be, so Tom decided to make the wheel-pulley larger. Doing that would decrease his top speed, but also increase the torque. Increasing the torque would not only help acceleration, it would reduce the amps needed to speed up, which would hopefully help his ESC survive the abuse.
Since he now had more reduction in the pulley-set, he also changed the kV of the motors (a winding designation that controls how many RPMs you get per applied volt). The motor for the V2 used a very low 192-Kv.
The belts he used were HTD5 pitch, and 15mm wide. His first youtube video brought him several sponsors, and he now had access to a very large 3D printer, so he was able to make the 184-tooth pulley from the V2 as one solid piece.
As you can see in the pic below, he also tried to improve the performance by adding a second motor, which used a separate belt to connect the two motors. For various reasons, he abandoned the dual-motor idea (discussed in the video below), rather than trying to improve it.
He also decided he wanted more volts and also more range, but a larger battery pack would not fit in the tiny seat-bag, so he put the new battery in a backpack that he would wear. In spite of the issues with the system, he recorded 3950W of power as a temporary peak, and he found that he REALLY liked that level of acceleration. By upgrading the battery pack from 22V to 28V (from 6S to 8S), his system can now achieve the same power (measured in Watts) with fewer amps, which will hopefully allow the ESC and motor to run a little cooler. He used a more advanced VESC-4, instead of the low-end 6S controllers that had been emitting smoke in order to indicate that it was time to buy a new one .
The flip-flop hub mount keeps the plastic pulley concentric (having the same centers) with the wheel. The actual rotational power is transmitted to the wheel through black zip-ties (shown in the pic) that are connecting the pulley to the spokes.
One of the features that is vital to a powerboard (but not to a model airplane or quadcopter) is that the VESC-4 has a very strong regenerative braking system (“regen”), which can act as a magnetic brake. Tom’s data-logging indicated that it returned only about 5% of the power he used, and put it back into the battery (which matches the experiments by many others).
However, many ebikers with a regen system end up liking it very much, simply as an additional brake which does not get hot. Tom has not used a freewheel in any of his motor systems so he can use the regen feature, which means that when he is on a downhill, the wheel drives the motor as a generator.
The high-performance V3
Tom’s brief experience with 4,000W of power had him hooked, but using two smaller motors was simply not working out, so he upgraded from the dual 63mm diameter motors to a single larger 80mm unit (3.1-inches diameter), which had a Kv of 160 (which provides 160-RPMs per volt).
The controller for the V3 was donated by one of his sponsors, a powered skateboard company called Trampa Boards. It is called a “VESC-6”, and it uses Field Oriented Control (FOC) which is a more sophisticated way to manage running the motor. The VESC-6 is a little larger, and much more robust compared to the ESC’s he had been using before. It also runs the motor MUCH more quietly, which was a pleasant surprise.
He also switched from the 8S LiPo, to a 12S pack made from the highly respected HB2 cells from LG, in the 18650 format. If using our recommended 3.0V to 4.1V range (per cell), a pack using 12 cells in series would run from 36.0V at the Low Voltage Cutoff (LVC) up to 49.2V when fully-charged, with a nominal “average” of 44.4V
He configured his pack as three sub-packs that are each 4S / 5P, and when they are connected in series, the end result is a 12S pack. The HG2 cell is rated as a 3000-mAh cell that is capable of 20A peaks, so five of them in parallel (5P) means this pack can easily produce 100A, and has 15-Ah of range.
The single-stage belt-pulleys are 184T / 12T, which results in a 15.3:1 ratio. If a 26-inch wheel is spinning at 414-RPM to achieve the 32-MPH he listed (51 km/h), the motor would be spinning (414 X 15.3 =) 6,333-RPMs.
Almost all RC motors and controllers do not have Hall sensors, which would constantly monitor the position of the rotor for the controller. RC components are typically “sensorless”, which uses sophisticated electronics to help the controller guess the position of the rotor, which it actually does fairly well. However, if you add Hall sensors to the motor, and then select a controller that has Hall sensor monitoring as an option, doing that can provide a variety of performance benefits.
In the video below, Tom does the walkaround and describes the system at the 12:30 mark.
The Final Version, V4
These RC motors from Hobby King and similar vendors are widely regarded as somewhat “affordable”. That is a kind way to describe the quality as being adequate, but not good. At only 400-miles, the V3 motor bearing failed, and Tom decided to tear it apart and replace both the bearings, along with upgrading the shaft to a hardened unit that also used a higher grade of steel. It has been speculated that the stock shafts in the Turnigy RC motors are made from a rather stiff alloy of cheese, sometimes called “Chinesium”.
Tom also decided that he wanted to upgrade the bicycle. As much as he enjoyed how lightweight the V1 had been, the performance that he was getting convinced him that he needed disc brakes, along with a suspension fork to manage the harsh potholes he was forced to ride over. He chose a stout downhill frame with 29-inch wheels. The larger diameter wheels and fatter tires made the potholes less of a shock, and the longer chainstays provided more space for the belt to clear the frame members.
Tom wanted the motor to be mounted inline with the frame, instead of sticking out to the side. He had experienced a few minor crashes that were not his fault, and he feared that eventually the protruding motor would be damaged. He made the initial prototype mount out of CNC’d plywood that placed the motor just behind the seat, but he then decided that he had enough clearance to mount it inside the frame triangle.
Below is the CAD graphic that he drew in order to determine the precise shape of the parts he would need. The two sideplates are aluminum, and the spaces between the sideplates and the frame-tubes are filled with 3D-printed plastic bushings.
If you want to have a lot of reduction with a single stage, you must have a tiny pulley driving a very large pulley. When Tom printed a large single pulley, it was stronger and much smoother-running, compared to making up several smaller sections and then glueing them together, but…he already had the largest pulley that could be printed in a single piece, and the previous 12T small drive-pulley only allowed the belt to engage 6 of the 12 pulley-teeth.
Tom decided to add a second stage through a jack-shaft with a 2:1 ratio. Doing that allowed him to get his desired amount of very high reduction (18:1) while allowing the small pulleys to have their tooth-counts increased to 20T (providing 9 teeth fully engaged with the belt), which reduced the occasional skipping and tooth-ratcheting that can damage the belt.
In the pic below, you can see that Tom upgraded the pulley-spokes to using aluminum. This way they are still very light, but twice as strong as the plastic spokes. Not only did the large diameter of the rear pulley provide a lot of RPM-reduction, it also helps the stock disc brake and caliper clear everything.
From the beginning of this long and winding road, Tom’s friend said that he could simply get a common hubmotor kit, which would be easy to install, and would also be much cheaper than some custom project. Tom agreed, but…the result ended up being much lighter than a 4000W hubmotor, with the weight being more centrally-located. The work he put into the design work also counted towards his engineering courses, and if any future potential employer asks if he has any “hands-on” experience, Tom has all of this documented for anyone who is interested.
In the video below, he does the walkaround and description of the V4 details at 7:37
If you want to see even more videos that cover Tom’s development of these RC ebikes, you can go to his youtube channel, which can be found here.
If you want to download all the CAD files that will allow you to print-out any of his 3D parts designs, just go to his Patreon page and sign up to donate $1 per month (found here). You can cancel at any time after that, and any money that is donated goes towards Tom designing and making cool shit, along with making the youtube videos that show how to do that.
If you like this configuration of RC build, here is a similar one from Patrick.
Plus Roy’s beautiful eCortina.
If you like custom builds of all types, here is a link to our index of dozens of them.
Tom may not have a degree yet, but engineers continue to learning over their entire life. Therefore…I wouldn’t say he is an engineering student anymore, he’s an engineer who is working towards a degree.
Written by Ron/spinningmagnets, September 2018