Patrick is a university engineering student from the Netherlands. Seven years ago (in 2010) he began posting on endless-sphere, asking questions, and building up custom ebikes…with each being better than the last. After several versions, here is what his experience has led him to build as his latest creation…
The Early Days, 2010
Patrick was only 17 when he began this journey. Since gasoline is very expensive in the Netherlands, he knew he would likely be riding a bicycle as much as possible when he attended university. In fact, the Netherlands is one of the most ‘bicycle friendly” countries in the world. He was already familiar with RC car models, so…it was a short leap to wondering if he could use his expertise with electrical components to motorise his bicycle.
He knew it would be wise to benefit from the experience if others, so he began reading and asking questions at an electric bike forum. He decided he wanted a mid drive, to give the motor the use of the bikes gears, and he also wanted to use the biggest RC motor that was available at the time. He purchased a Turnigy 80-100 outrunner (the stator is 80mm in diameter, 100mm long / 3.1 inches by 4.0).
Outrunners are different than the motors that most people are familiar with. In order to increase the torque per a given size (RC motors are typically pretty small), The stationary “stator” is in the center, and the spinning permanent magnets are inside a shell that spins. In the pic above, the large gold-anodized aluminum baseplate at the bottom is stationary. The chrome shell, the shaft, and the thin gold end-plate at the top all spin when the motor is energized.
The “130” number refers to its Kv, meaning how many RPMs per volt that this winding will provide. The same size and model of motor can be produced with any one of several Kv’s. 130 is lower than average, and that is very helpful for an electric bike, where the motor RPMs must be reduced down to a bicycles wheel-speed. A common 26-inch wheel at 28-MPH is turning roughly 360-RPMs (a 130-Kv motor at 37V is turning 4800-RPMs)…
The primary reduction needed to be as large as possible for reasons I will discuss farther below, and Patrick chose a 14T drive sprocket and 72T sprocket, using a #25 chain which has a tiny 1/4-inch pitch. High-RPM drive sprockets that have fewer than 11T are very loud due to a phenomenon called the polygonal effect, so this 14T sprocket shown should be as quiet as a chain-primary could be.
The jackshaft is a 1/2-inch diameter that has an aluminum adapter at the other end, which allows a smooth shaft to accept a threaded freewheel. This freewheel is a heavy-duty White Industries ENO 16T unit, and locating it there allows the rider to pedal the bicycle without the pedals driving an un-powered motor.
In the pic above, Patrick purposefully chose a frame that had a cylindrical downtube, to make it easier to design and fabricate aluminum billet clamps for attaching the drive. The secondary reduction chain is tensioned by sliding the entire drive unit up the downtube.
The secondary reduction uses common 3/32 bicycle chain and sprockets using 16T and 44T, which provides a 2.7:1 ratio. The two ratio’s together end up at a (5.1 X 2.7 =) 13.8:1 ratio. This means that for every 13.8 turns the motor makes, the pedal-sprocket will spin once. However, Patrick also chose to use a slightly smaller output chainring on the bottom bracket, using 30T (44T/30T = 1.4:1), so…the final output reduction ended up at (13.8 X 1.4 =) a 19.3:1 ratio between the motor RPM’s and the chain to the rear wheels’ gears.
Patrick actually did fry two of the expensive RC controllers, so..based on advice, he added low-ESR capacitors for voltage ripple-smoothing, which fixed that problem. You can see in the pic above, the fat red and black wires that go into the tiny ESC (at the top left).
This unit came from the factory with a single black cylindrical ripple-smoothing capacitor located right where those two wires enter the ESC. Just above that place you can see the four low-ESR capacitors (Equivalent Series Resistance) that he added. They are the same part number that the factory used, and they are only $2 each, so it is cheap insurance to buy four, instead of just adding one extra and crossing your fingers…
I have seen many builders fry these tiny HV-160 controllers, and here’s what Patrick did right. The maximum voltage they are rated for is 12 lithium cells in series (12S), which is generally referred to as 44V nominal, and fully charged at 12S X 4.1V per cell = 49.2V. Rather than run this controller right at the ragged edge by using 44V, he only used a 10S battery pack, which is 37V nominal, and is fully charged at 10S X 4.1V = 41.0V.
The constant on-and-off current that is drawn from the battery causes voltage ripple, and that produces very short voltage spikes that are actually higher than the battery voltage…and anything over 50V can damage the expensive controller. To reduce voltage ripple, make the red-and-black power wires from the battery to the controller as short and fat as is practical, run a system voltage that is not too close to the maximum that the ESC is capable of, and add several ripple-smoothing low-resistance capacitors (as close to the ESC as possible). These capacitors have two leads, so you simply hook up the positives to the red wire, and the negatives to the black wire (google image: “ESC capacitors” to see examples).
Once you add-in wind resistance and the weight of the rider and bicycle, Patrick achieved 44 km/h (27-MPH) with this combination. That may not sound very fast, but it runs very efficiently due to the high magnet speed of the motor, and it also provides an incredible amount of wheel-torque due to the very high reduction. Of course, one of the benefits of a non-hubmotor system is that you can trade off some of the torque to get more top-speed with a simple and cheap chanring swap.
The Second Version, 2012
As happy as Patrick was with the first version of his mid drive street commuter, the Netherlands is very flat, and he noticed that he usually only used one gear. Running the motors power through the bikes gears allows the rider to downshift the motor when you want to climb very steep hills, but it also adds wear onto the pedal-driveline. Patrick also noticed that the rate of wear on the driveline was near the max he was willing to tolerate when he was using 41V X 60A = 2500W.
He wanted to use even more power on occasion (who doesn’t?), so…he didn’t want to run a higher power level through the bikes gears, especially since his flat terrain made him realize that he now only needed one gear for the motor.
The easy route would have been to adapt a simple rear hubmotor to the next ebike project, but hubmotors spin at the lower RPMs of the wheel, so you need a larger and heavier motor to get the same performance. Not only is a mid-mounted motor (with a high reduction) lighter and more efficient for a given power level, it also provides a more balanced feel to the entire bike, since the weight is near the center.
“…This will be my second E-Bike. The main points of this bike are: lightweight and durable but still have a good amount of power. I bought this bike specifically for E-conversion, it weighs 14.9kg’s stock. The motor that will be used is a brushless Turnigy 80-100 130kV in combination with a Castle Creations 160HV running on 10S / 2P LiPo. It will be a left-hand driven bike….the chain is #25 from sdp-si.com…I got the shaft-to-freewheel adapter from Recumpence (Matt) for about $90 or so. The first test ride, the maximum power I used was about 2500W (approx 41V X 60A). The ESC is a HV160 and the motor is a 130-Kv…The battery box is made of white acrylglas and aluminium. The acrylglas is only used for the outside of the battery box, while the aluminium is for the strength of it…”
In the pic above, you can see the three heavy-duty “T-bolt” tubing clamps that he used to hold the motor mounting bracket onto the cylindrical downtube. He also used those to hold the jackshaft bracket onto the cylindrical seat-tube. By replacing the stock bolt on this style of clamp with a section of all-threaded rod, this style of clamp can have threads extending out on both sides of it.
Having a sprocket right next to a brake disc is something that has to be carefully considered. Chains require lubrication to achieve the maximum possible life-cycle, and any lube that gets onto the brake disc would cause a loss of braking ability. It is possible to use a wax-lube that does not drip or fling, but since Patrick decided to use a #219 chain and sprocket set, he could specify a sealed O-ring chain. This is where each pin in the chain has tiny rubber seals at its ends, to hold the factory lube inside the link. It has worked well for Patrick so far.
If Patrick had not already begun planning the next version, he was contemplating a swap to a large 203mm diameter disc in order to move the caliper much farther away from the fairly small sprocket.
There can be an added RPM-reduction benefit as a result of keeping a small brake disc, and then adding a large diameter sprocket to create more physical separation (as opposed to using a large disc with a small sprocket), but…most bicycle frames have flared chain-stays (closer together at the front, and wider at the rear). This means a very large driven-sprocket would place the moving chain very close to the frame. When it comes to design?… every choice seems to be a compromise between a benefit and a drawback, and each designer is forced to “choose their poison”.
The Final 2017 version, 11,000W
Patrick had learned new things on each previous version, and the saddle-style battery box was a compromise that he accepted at the time to get that version running. But the more he used it, the more he wanted the battery to be located somewhere else. Not only was it awkward to have that bulk located there, the weight being mounted so high made the handling less desirable. There would be several changes on the next version, but the major design choice was to mount the motor and jackshaft inside the frame triangle, and then to put the weight and bulk of the battery box in-front of the downtube, nice and low.
At this stage in his university engineering education, Patrick had begun learning computer aided design (CAD) drawing, and the renderings he posted really knocked me out. Back in 2014, we wrote about the very useful CAD and also CAM methods like water-jetting and laser-cutting. This allows anyone to draw a custom part on a computer screen in their home, and a shop that’s somewhere else makes it and ships it to your mailbox.
The last version only had about 2500W, but…for the next one, he wanted more…MUCH MORE! Patrick knew he would be needing a stronger mounting bracket for the motor and jackshaft, so this one has three mounting points that are attached to both the seat-tube and the top-tube of the frame. This prevented the systems’ torque from twisting the mounts, and easily kept them aligned with the rear wheel.
Patrick still used the Castle Creations HV-160 ESC that he had dialed-in, but this time he chose a very similar motor that was from Alien Power Systems, the 80-100-130. The battery is a LiPo 37V / 16-Ah pack (590-Watt Hours).
He had gotten the previous motor too hot when he started raising the amps, in order to see how far he could push it. Unfortunately, he had gone too far, and then had to order a new motor. Other builders had enjoyed great success by adding a forced-air cooling fan, which would allow him to dramatically raise the temporary peak amps, and his data-logging ESC showed that he had drawn occasional peaks of 297A! (37V X 297A = 11,000W).
The primary reduction still uses a #25 chain, but now the sprockets are 14T / 54T for a 3.8:1 reduction, which provides a higher top speed. The secondary reduction from the jackshaft to the rear wheel uses a very tough Wipperman Connex 1G8 1/8-inch chain, going to a chainring-to-diskmount “top hat”adapter from DaVinci drives.
The bicycle is an aluminum 15 kg (33-lb) Bergamont Kiez Dirt, and the total weight with the drive system and batteries added is 25kg (55.1-lbs). It is currently geared to provide a top-speed of 70-km/h (43-MPH).
“…I run the motor sensorless and I am quite happy with the start up. However, I always start by pedaling to get it going, and after that I apply the throttle. (If you don’t do this, you might blow the ESC)…“
Here is a one-minute video walkaround:
“…I did some runs with the bicycle as it was perfect weather last weekend. The bicycle handles so well, is amazingly lightweight and it is so much fun driving it! Its very fast and wheelies easily. Top speed measured by GPS was 67 km/h (41 mph) which is very fast on a bicycle…“
Written by Ron/spinningmagnets, May 2017