Whilst out in America I was browsing a discount warehouse's catalog when I found a 24V 450W DC Motor for about $40, and with the thought of an electric bike, I impulse-bought it.
When it arrived I checked the motor was functional and then began designing the rest of the system. The motor had a roller clutch fixed to the axle with a mini-chain cog on the end. Not having a suitable bicycle for the moment, I set to work on the Electronic portion of the design first. For batteries I decided on a couple of small SLAs, these would provide 24V when in series and were rated up to 70A discharge current. These were chosen due to the low cost.
I will upload the full schematic once I have finally dug it out and scanned it in.
|I initially envisioned an ambitious plan to use a microcontroller as the controller, with features such as temperature compensation, current limit and diagnostic logging. However I was going to need so much A/D interfacing that it really wasn't practical.
So I redesigned the circuit around a couple of op-amps. One is configured as a triangle wave generator at around 25KHz, the other is a comparator, a variable voltage is supplied by the potentiometer to give a square-wave PWM output.
The DIL chip on the left is a low-current charge-pump inverter to produce a negative supply rail for the comparator. The TO-220 is a 5V regulator to feed the inverter.
|The PWM output then drives a couple of 25A MOSFETs in parallel, the nominal current of the motor is 22A, however peak currents can be many times this, and redundancy is always good!
The large resistors are used to sense current, the resulting voltage across these is then linked to the potentiometer controlling the camparator, producing the effect of automatically reducing the PWM duty-cycle when the current is above 25A.
The paperclipped heatsink is on a zener and reverse protection diode that protects the MOSFETs from flyback voltages produced by the inductance of the motor on the PWM edges.
|This photo shows the initial test setup of the full system, the battery borrowed from my car is supplying 12v and the motor can be seen alongside it.
You can see the Multimeter is reading 1A of current and the Oscilloscope shows 50% duty cycle from the controller.
This is with half the design voltage, half drive, and a completely unloaded motor, hence the low current.
(Please excuse the other junk on the desk, I stole my Dad's desk halfway through his GPS Timing Reference project.)
|I then built the prototype system on stripboard, and in late 2009 I bought a bicycle and began fitting parts for conversion to electric drive!
This is the system, minus batteries, sitting on the pannier of the target bicycle. The red and black wires will connnect to the batteries as these will be kept as low as possible for better handling of the bike.
The bicycle was then fitted with a custom wheel allowing the existing V-brakes on the frame to be used whilst providing a disc brake mount on which a cog could be fixed, a clutch was already present on the motor, so the rest of the drivetrain could be fixed directly.
I had purchased a suitable cog in the States with a gear ratio to the motor that when combined with a standard bicycle wheel would give 19.5 MPH at the motor's optimal power RPM. This may be a little conservative given the power available, however I wanted to assure the presence of sufficient torque for climbing hills at lower speeds.
Current State of the Project
The bicycle and custom wheel were unfortunately stolen from the Southampton Common in Late September 2010. This project is therefore on hold until I acquire another suitable bicycle.