Powabyke - electrical



General

The manufacturers literature indicate that a 250W motor is used on the front wheel hub of this particular bike. It is fed via a duty cycle controller from a 36v power source comprising three 12v 12a/h gell lead acid batteries, series connected. If required, the battery pack can be removed for charging away from the bike. A (very electrically noisy SMPU) charger is provided, and takes about 4 hours to fully charge.

Both the rear wheel and pedal crank rotation are monitored and fed into the motor controller, Two active positions on the ignition switch allow power to be fed to the motor either with just rear wheel rotation or both rear wheel and pedal crank rotation. Thus in all circumstances, power to the motor is inhibited whilst the bike is stationary. A twist grip control is used to allow some control of motor torque, and therefor speed, to be achieved. This twist grip uses a Hall effect device, so there is no requirement to make electrical contact with the rotating part of the grip.

An indication of battery condition is provided by a six section bar LED display.

Access to the controller electronics is fairly poor, mainly because the controller pcb is used also as a junction point for all the bike wiring, and the length of these prevent any meaningfull work being carried out on the unit in-situ.

No lighting is provided, which is a shame, given the battery power available.








Adding bike powered lights 



As long as the wiring is routed through the ignition switch (so that they cannot be left 'on' when the bike is not in use), there is no reason why the lighting should not be powered from the bike battery. In fact, it makes them more reliable, since you always have a good sense of how well charged the main battery is.

Using LEDs (light emitting diodes) is the most obvious route these days. Volt drop per LED is about 3 to 4v. So connecting several in series via a common series resistor gives a suitable and simple configuration. The rear light shown opposite has a single string of 15 LEDs.                                                            
 


rear light  rear lamp1

lampback
click on image to enlarge

         
                front_light

For the front light, two strings of eight white LEDs were used.

All LEDs are wide angle types,  mounted on replacement pcbs in a pair of cheap Tesco bike lamp housings.
lampfront
click on image to enlarge


Controller failure


After about four years use, the controller failed. The bike wasn't being stressed, and there was nothing else unusual that day. Knowing how difficult it was to work on this board (a friend had had a similar problem), it was decided that it was time to do a few experiments and measurements, prior to possibly scrapping the bike. The controller was replaced with a straightforward (power) push button, so that now the motor would either be fully on or fully off.

Results were suprising...

Although the motor had too much torque at near-zero speed, the maximum


Where to have the controller replaced in Cambridge

As far as I am aware, there is only one bike shop that will do major repairs on electric bikes, and that is:

http://www.electricbikesales.co.uk/

They are just off Mill Road, at:

Hope Street Yard
Hope Street
Cambridge CB1 3NA

01223 247410




speed was no differant than before. I had assumed that the controller set the maximum speed (to meet the 15 mph restriction required under UK law). An ammeter was then put in series with the motor. Remembering that the motor is rated at 250W, this power implies a current of 7 amps. At zero speed, current is well above 10A - applying the brake at standstill, then turning on the motor will blow the battery pack 16A slow-blow fuse in about half a second. Not a good thing to do.
Just a couple of mile-per-hour motion and the current drops below 10A, even up hill. By 5 mph, the current is below the 7A limit. At 10 mph, the current is barely 2A. These figures are for level gound, with no no or minimal headwind and no pedal assistance.

On this particular bike, the motor actually stops drawing current at 13.2 mph, limiting the unassisted speed on the flat to about 12 mph. Time for more thought...


 



Motor type and characteristic

Having found that the motor torque rapidly falls away with even modest speed, the implication would be that the unit uses a permanant magnet field, and that a back emf is generated as soon the motor rotates. To verify this, the motor was disconnected from the bike electrics, and a voltmeter placed across it instead. Sure enough, with any motion there is voltage appearing, and inevitably, at 13.2 mph the value reaches 36v. Incidentally, the motor drives through a ratchet, so although the motor itself could be used to supply power as a generator, in practice, the slightest load, will cause the motor to dissengage - so there is no point in trying to put power back into the battery when slowing down. On the other hand, neither can you make the the bike move backward in error!.

An interesting point with sort of motor is that when fed from 36v, is that maximum power is consummed not at maximum speed, but at relatively low speed, when accelerating from an even lower speed. If you peddle assist (if only out of boredom), and you live in Cambridge, where it is flat, the motor is only lightly loaded. At 10 mph, it consumes only about 50W. Living in one of the outlying villages, as I do, this is really not that useful - and remember, this applies even with the motor connected permantly across the battery. It was becoming obvious that more volts were going to have to be applied to the motor...

Initial tests with 48v applied to the motor

One of the recently replaced 12v batteries was wired in series with the existing three, and housed temporarily in the bike pannier bag. At very low speeds, the torque was a little too high, and the acceleration too aggressive. However, from about 5 mph upwards, the performance was very pleasant. Up to about 10 mph, there was a need to pulse the push button controller, in order to keep the speed at this value. On the flat, and without applying pedal power, the motor would now run the bike at just about 15 mph. By applying pedal power, or by going downhill, the motor was found to stop taking current at 17.2 mph. At these modest speeds (I would expect a 'tourist' biker to doing 20 mph and racer perhaps 25 mph average), the bike now felt much better.
[A motor resonance at about 13.5 mph was initially worrying, but found to be due to the gearing running dry - see 'mechanical' page]

Limiting the low speed torque

The extra 12v battery was very inconvenient, and the high low-speed torque needed calming down a bit, so what to do?

Changing the motor gearing was never going to be a realistic thing for me to do, but I did have lots of 36v dc-to-dc convertors of various isolated output voltages from 5 to 15v. The 12v variants were current limited at 4.3 amps, so it was expected that at least two would need to be paralled to provide an 8 amp capacability. In the event, the current limiting at 4.3 amps proved to be just right.

By connecting the battery to the dc-to-dc convertor, and  bootstrapping the battery output with the convertor output, a 48v supply was now available to supply the motor. Although this configuration is wired through the (two active position) ignition switch, so that position 1 supplies 36v and position 2 switches on the convertor to provide 48v, it has proved an unnecessary complication. Operating at 48v, the convertor simply current limits at low speeds, so the motor still sees only 36v, or very slightly higher. From 10 mph upwards (on flat ground), the convertor starts to come out of current limit, and at about 12 mph, the voltage is at it's full 48v value. It doesn't seem necessary to add a twist-grip controlled speed facility - just putting more or less effort into peddling (rather like a track Derny) is enough.
But that is just me, and for use in this flat, relatively even, terrain. In any case, these results may hopefully be of general interest - please let me know, if so.

controllercontr