Swan TAFE
Electric Vehicle Challenge
Electric Vehicle Challenge
 
2009 Event Information
eV Video Competition
Rules
Venue
Sponsors
Contact
Teams Registered
Entry Forms
Useful Information
Yuasa  eVolution
Past Events

Building an E-bike

by Phillip Calais

Printed in Renew 71, April - June 2000

During the past few years alternatives have been sought to supplement or replace motor vehicles. One old alternative, the human-powered bicycle, is all too frequently regarded as simply a form of exercise and not as a means of transport. This is due partially to Australia's often hot climate that results in the rider arriving at their destination dripping with sweat and asking where the shower is. While this may be fine as exercise, as a mode of transport it leaves much to be desired. Electric bicycles and power-assisted bicycles (PABs) are one type of bicycle in which this is not a problem.

I first saw an 'e-bike' in Germany in 1994 while I was working at a electricity utility. As well as several electric VW Golfs, one of which I frequently drove, the company also had a Hercules e-bike. I decided that if I ever had the money, an electric car would be high on the list.

That time is yet to come, so instead I decided to build an e-bike and after arriving back in Australia I started investigating the matter more thoroughly. The German e-bike was fitted with a 180 watt motor and a large battery pack that sat in the handlebar carry basket and although it was somewhat awkward, unbalanced and underpowered, it worked. But I decided that it wouldn't be too hard to improve upon the design. Thus started my odyssey

Why?

On first sight, one may think that they are a ridiculous item - what is the point of having a powered bicycle? I am often asked things like "An electric bike? Kinda defeats the purpose, doesn't it?!" or "What? Are you some kind of cripple or something? You've got legs haven't you?" This first question was asked by a body-builder who happily drives his V-8 HSV commodore to a health club to ride an exercise bike. The second was asked by a T-shirt and shorts wearing bike dealer in Berlin where 28° C is a hot day and the only time he has had to wear a suit was walking down the isle and the next time will be when he is carried down it.

Obviously there are times when I do want to pedal my way for exercise or relaxation. But these times I am not wearing a suite and running late for work!

Hey, that's me on the right with my new bike with the hub-mounted motor! I haven't got a good photo of the bike I made and which is described in this story as some mongrel pinched it... The other guy is my friend Sunny Miller. He's on a eBike we got from Canning Bridge Cycles. (click here for full size image)

The motor

The most efficient way of converting the power stored in the battery into motion is by using a hub-mounted motor. In other words, the shaft of the motor is the axle of the wheel and the spokes are connected to the motor casing. Hub motors are potentially lighter and with lower transmission losses than other types of motors have the potential of being more efficient. Brushless motors are also usually more efficient than brushed motors. Unfortunate the availability of these types is rather limited and the best I could do was to get a reply from a custom motor builder who simply asked "How much money do you have?"

At the other end of the scale are the friction drive systems. The few friction drive bikes that I've seen weren't very impressive apart from the phenomenally high rate at which the tyre wore out and their inefficiency. But more about these later.

In the end I came to a compromise. I used a 12 volt DC brushed motor from a Black and Decker Stealth lawnmower (available from your favourite B&D person - and I don't mean that sort of B&D!!) which I mounted on a carry rack along with the batteries. Good DC motors can sometimes also be obtained from old mainframe computers - check out your local scrap yard.

One of the main problems with fitting any electric motor to a bike is that most electric motors spin at about 3600 rpm or more while the bike wheel only needs about 300 rpm. This means that the motor needs to be geared down by a factor of about 12 but even using the smallest sized sprocket possible on the motor and the biggest possible on the wheel still gives a mismatch of about two or three to one. The bike can be run like this but it means that the motor is only being used at its very lowest speed range and efficiency suffers.

Here's the motor mounted with the end of an angle-grinder as the right-angle gear. You can also see the 80 'C' size NiCd batteries. (click here for full size image)

Transmission and stuff

The other problem was that the motor I chose is quite long and mounting it so the shaft sprocket lined up with the wheel sprocket unbalanced the bike. The obvious answer was to mount the motor laterally and use a right-angle reduction drive - except the cost of a 2:1 right angle gear was more than the rest of the bike put together. This was solved when a friend suggested using an angle grinder. So an old Makita angle grinder was hacksawed in half, cleaned up, threaded and joined onto a Black and Decker lawnmower motor. The total reduction with the 14 and 54 teeth sprockets was now 9.6:1, giving a wheel speed of 373 rpm at a motor speed of 3600 rpm. This equates to a top speed of 46 km/h. In practice, the drop in motor speed under load results in a maximum speed of about 40 km/h.

I didn't want to connect the motor to the existing chain and sprocket as that would mean that the pedals would go around when electric powered and the motor spins when I pedal. In the end I fitted a free-wheeling sprocket to the left hand side of the wheel and this was perhaps the biggest problem I had to solve.

If a standard freewheeling sprocket is fitted to the left hand side, it unwinds as soon as power is applied. The solution seemed to be to weld the sucker on but this would then would have deformed the ball bearing race and if anything broke or wore it was not repairable. In the end I made a hub with the standard gear cluster on the right and a freewheeling hub on the left. I did this by getting a BMX 'flip-flop' hub which has threads on both sides. A flip-flop is designed so that a quick ratio change can be accomplished by removing the wheel and putting it on reversed.

I overcame the sprocket unwinding problem by getting a track-racing hub which uses a simple sub-diameter reverse-thread locking ring to fix the normal threaded sprocket onto the hub. This hub was hacksawed in half and then machined with an internal thread so that it could be screwed onto the flip-flop. Assisted with copious quantities of Type-620 Locktite and locking pins, this was fixed onto the flip-flop. The freewheeling sprocket could then be screwed on and then locked with the locking ring.

The only other problem was that the biggest freewheeling sprocket I could get at a reasonable price was a 17 tooth - far too small so I had to make an adaptor so that I could fix a 54 tooth sprocket over the 17 tooth freewheeling sprocket.

Sound complicated? Well it was!

click here for full size image

Total Control (almost)

Finally the controller was built. At first just to get the thing going I used a simple relay and switch. Crude, but reasonably effective. I called it my digital control system - it was either on or off. After pedalling to get the bike moving, the button was hit and with 600 watts the acceleration was quite amazing.

However this isn't really a satisfactory solution, due to the sudden stress on the components as well as its dubious legality (more about that later).

What was needed was a controller which would vary the speed from zero up to the maximum. With small DC motors, speed control can be achieved by controlling the voltage and the easiest way to do this is with a 'pulse-width modulation' or PWM controller. This sort of electronic circuit chops the DC into pulses of varying width - hence the name.

To cut a long story short (maybe more about this stuff in another article?) the pulses, which go right up to the battery voltage, of say, 12 volts, are separated by periods when the voltage drops to zero. This may occur a few hundred to a few thousand times a second but the motor only 'sees' the average of these 0 or 12 volt pulses. If the pulses are on for half the time and off the other half, then the motor sees only 6 volts and so goes at half speed. If the pulses are on for ¾ of the time, then it sees 9 volts and so on. With this method of control, when the pulse is on, the current is only limited by the motor and full torque is available.

click here for full size image

Voltage control controls the motor speed, but current control is also possible and this controls the torque or turning force of the motor which indirectly controls the speed. Current control has a number of advantages including better efficiency but the design of the controller is complex and I'll leave this for another time.

In the end I experimented with a few different PWM circuits. The first I tried used a NE555 IC and was based on the diagram given in the October 1994 edition of Electronics Australia but I had problems with this. (See what Lance Turner had to say about this type of circuit on page 72 of Renew #68.)

While a great little chip, I have since decided that controllers that use the 555 chip leave much to be desired. I found that they were quite unstable, unreliable, often affected by 'noise' and were themselves very noisy. By noise I mean electromagnetic interference - when passing some other noisy appliance the controller would sometimes start misbehaving - like suddenly decide to switch on full or stop dead.

A much better circuit is given by Lance on page 68 of Renew #66, but improvements can still be made and so I started working on microprocessor-based controller. While more complex, these (can be!) even more reliable and versatile. As well as it being possible to have either voltage or current control or both, this type can also be used to monitor battery state-of-charge, motor temperature and other parameters.

Unfortunately while I was in the process of testing this, my bike was stolen and so this avenue of research has been postponed for the time being. However I intend to take it up again sometime soon and hopefully will have an article in a later issue of Renew about controllers.

Before I forget to mention this, many DC motors have an internal diode which can cause problems. These diodes help reduce electromagnetic interference and prevent the motor spinning in reverse. The rapid voltage switching from a PWM controller can make the motor windings and diode act like a voltage 'boost-buck' converter and the cheap and nasty diode will probably explode. The diode needs to be removed and either replaced with a better diode or some other type of suitable noise filter.

Other Stuff

I previously mentioned my dim view of friction roller-type systems. There are a few of these available commercially and they are also the easiest and cheapest system to make. A rubber roller is simply stuck on the motor shaft and then pressed against the tyre. A 50 - 60 mm diameter roller on a 3600 rpm motor will give the right speed without any further mucking around.

While very simple the few that I had seen never worked very well. The roller slipped when wet or going up an incline and they often wore the tyre out very quickly. However my opinion of these changed when I saw an electric bike in a local cycle shop which used this principle. More of an electric moped - pedalling was really only a last resort - it was well made, balanced and ran very well. The motor, roller, battery and controller are mounted in a neat pannier on the back and the performance was very good.

How they achieved this was by using special fat tyres - about 50 mm wide - with a tread that 'cogs' into the toothed roller. This seems to eliminate all slip (if the tyre isn't bald) and also helps keep the price down as compared to systems that use fancy gearing and transmissions.

With a bit of research into suitable tyre types and rollers, it should be possible to make a simple, but very effective e-bike in this manner.

Another issue that needs to be addressed is the power limit. According to Australian law, vehicles with less that 200 watts need not be licensed - a bike with 198 watts is still a bicycle but one with 202 watts is classed as a motorbike and needs to be licensed and can only be ridden on roads and not on cycle paths.

This has caused a lot of problems for both the back-yard tinkerers as well as for the commercial manufacturers of e-bikes. In most civilised countries of the world, the limits are either higher - usually 250 to 500 watts - or they simply impose a speed limit on the e-bikes. In the United States for example a 1000 watt motor is allowed - but the bike must be limited to 20 mph (36 kmh).

Some people have tried importing e-bikes from the US or Japan, only to find that the bikes are over the power limit and can't be sold or used and it's virtually impossible to have one licensed as a motorbike.

Different Australian states have different interpretations of this rather ambiguous law which was originally made up by some techno-bureaucrats in the 1970s for petrol-powered motorised bikes which have completely different characteristics to electric bikes. Petrol powered motors when tested under certain standard temperature and other conditions will give a particular result. Not so with electric motors. What affects these more is how you feed them - what sort of batteries, what sort of controller and so on.

In some states the power limit is taken to mean the power input to the motor and taking into account the losses in the motor, maybe only a 170 watt motor is legal. Elsewhere the 200 W limit is taken to mean the motor's nominal output or the power at the wheel - the brake-horsepower so to speak. I asked someone in the Northern Territory about the NT interpretation and they just said something like "What limit? Hey is the beer cold enough yet and throw another Barbie to the croc."

I contacted the WA transport authority and they told me that they take the law to mean 200 watts at the wheel. So with a system such as mine with a 80% efficient brushed DC motor driving through a 80% efficient gearbox and then through a 90% efficient chain and sprocket, the motor output power allowed is then 200 ¸ (0.8 x 0.9) = 280 watts and the DC input power allowed is 280 ¸ 0.8 = 350 watts.

With this interpretation of the law the motor rating is irrelevant and legal power can easily be achieved by electronically controlling the motor so that the maximum output power at the wheel never exceeds the limit.

It seems strange to me that all other vehicles are speed limited - cars to 110 or 100 (depending on the State), buses to 100, electric wheelchairs to 10 km/h etc. Only bikes are power limited.

After the theft of my e-bike I bought a new bike with the insurance money (which didn't pay for any of the 100+ hours of work I put into it) and fitted on a Taiwanese hub-mounted brushed DC motor. This came complete with a motor-bike style twist-grip and controller.

Taking only an hour or so to fit on, this system works very well and even cruising along at 30 kmh without pedalling, its surprising how few people notice that it has a motor.

With an equivalent fuel consumption of 0.15 litres per 100 km (or 1875 miles per gallon!) its rather hard to beat. Even taking into account the losses converting coal into electricity and then sending it around the countryside and charging batteries, it still works out at about 0.75 litres per 100 km (400 mpg). However a single solar panel will easily charge the batteries, meaning that no fossil fuels at all be used.

My next project is likely to be something a bit more adventurous - maybe an electric motorbike that has all one would except from a motorbike - speed and acceleration. Later still, maybe an electric car. For the time being though, I'll just keep experimenting.


Types of motors and drives

Many types of motors are available and new types are constantly being developed. The table below lists the main types of DC motors but within each group there are also various sub-types - axial flux, radial flux and so on. AC motors can also be used and brushless 'DC' motors are really a type of 'uni-polar multi-phase' AC motor and this is why they require a complex controller - essentially a small inverter - to operate them. However their efficiency is much better than other types of motors and most of the racing electric and solar cars and bikes use them.

Motor

Advantages

Disadvantages

brushed DC(field-coil type) cheap, readily available, reliable, simple speed control low efficiency (60 - 80%), requires periodic maintenance (new brushes)
brushed DC (permanent magnet) cheap, readily available, reliable, simple speed control medium efficiency (80 - 90%), requires periodic maintenance (new brushes)
brushless DC(normally permanent magnet) high efficiency (90 - 96%), needs little or no maintenance generally expensive, hard to obtain, speed controller complex

The other important part of the mechanical construction is the transmission between the motor and wheel. The most simple type is a rubber roller on the motor shaft that rubs against the tyre while the most elegant solution is to make the hub of the wheel the motor itself - the shaft of the motor becomes the axle of the wheel and the motor casing becomes the wheel hub. As most electric motors spin at 3000 rpm or more - ten times faster than is needed on a bike, some sort of speed reduction or gearing is also needed unless a special low-speed motor is used.

Drive / Transmission

Advantages

Disadvantages

scrubber or roller on tyre simple, cheap, generally reliable, can mount on front or rear wheel, acts as its own reduction can wear tyre out quickly, tends to slip in wet weather, requires some maintenance
belt-drive to wheel fairly simple, cheap, generally reliable, can mount on front or rear wheel tends to slip unless kept taunt, needs reduction
sprocket and chain to existing bike chain and sprocket fairly simple, cheap, reliable, good efficiency pedals go around when operating electric, pedalling will also drive motor unless a free-wheeling device is fitted to motor, needs reduction
chain and free-wheeling sprocket to left-side of rear wheel good efficiency, reliable, doesn't cause pedals to rotate, when pedalling, motor isn't driven rather complex - needs specially made rear hub, needs reduction
hub motor very elegant - high efficiency, can be quickly fitted or removed, no modifications needed to bike unless motor is special low-speed type, requires internal reduction gearing for best efficiency

 



Swan TAFE logo

Disclaimer  |  Report website issues to web@swantafe.wa.edu.au  | Last updated 18 July 2006