Go Bike Home Page
About us:
Join Go Bike!
Contact Go Bike!
Achievements
Meetings
Cycle Rides
News:
News & Consultations
Newsletters
Connect2
Core Paths
Cycle map:
Glasgow Cycle Map
Out on your bike:
Cycle Routes
Features
Reporting Faults
Donating Bikes
Cycling Links
Go Bike! Strathclyde Cycle Campaign
Features
Cycle lighting system
by Arthur Moody
Introduction
This short article describes a cycle lighting system which uses a dynamo and battery set. Normally a battery lighting set would require replacement batteries or frequent recharges. A dynamo lighting set avoids these problems but the lights go out when you stop (or when climbing a steep hill). By combining the two systems I hope to present a solution which avoids all of these problems.
If you know a little about electronics than you should be able to understand the circuit description. If you wish to make a system for yourself then you will also need some tools and a fair degree of manual dexterity.
Requirements
The design of such a system present a number of challenges.
- The dynamo is not actually a dynamo, it is an alternator with an AC output so a rectifier is required. I'll call it a generator from now on.
- One side of the generator is connected to the frame of the bike. You could try to insulate it but an electronic solution is better.
- The generator is inefficient, this helps to reduce the number of blown bulbs. We want to charge a battery and run lights at the same time so we need to increase the efficiency.
- If for some reason the battery becomes discharged we still want the lights to work while we're cycling.
Rectifier
To obtain maximum power output from the generator we need to use a full wave rectifier. This means that both the positive and negative voltages produced by the generator are used. We have an additional complication that one side of the generator is earthed to the bicycle frame so this calls for a voltage doubling rectifier. When this is used a nominal 6VAC from the generator will end up as about 17VDC. The actual voltage available will be slightly less than this because we will lose about 0.7V across each of the rectifier diodes. If we use schottky rectifiers then we will lose only about 0.4V across each diode. This gives us about 16.2VDC. We will have to find a way to efficiently convert 16.2VDC to about 6VDC.
Efficiency
While researching this design on the internet I found information which suggested that the generator was a current output device and that this allowed the lamps to regulate the output power. I investigated and found that the output impedance of the generator was surprisingly high. This also suggested a method for extracting more then the measly 3W, which is the rating of the generator. Normally they will deliver 0.5 A into a 12 ohm load at 6V. If I could increase the load impedance then I might get something like 0.5A into 18 ohms at 9V, this would be 4.5W, a 50% increase in available power.
Voltage conversion
Most people with a bit of physics behind them will know that a transformer can be used to convert an alternating electric current (AC) from one voltage to another with very little loss of power. They are generally very efficient. However they are also bulky and heavy, not ideal for a bicycle. There is an alternative which can convert DC from one voltage to another using only electronic components, a DC to DC converter or switching regulator. It is very efficient but also small and light. This sounds ideal but will it interface with a bicycle generator? Unfortunately not. As I described earlier to increase the efficiency of the generator and extract more power you need to feed it in to something which has a higher input impedance than the lamps it would usually feed. However the DC to DC converter has a negative input impedance. (Ref 1.) I was trying to figure out how to get round this when I saw a circuit idea in Electronics World submitted by P W Fry, G4SBF of Southampton. The circuit idea uses an LM2575 which is a switching regulator, but the clever bit is that a control transistor has been added so that the regulator remains shut down until the rectifier output exceeds 21V.
The net effect of this extra transistor is to raise the apparent input impedance of the DC to DC converter. The problem of the efficient interface between the generator and the DC to DC converter had been solved.
Battery interface
The voltage at the output of the DC to DC converter can be adjusted to any desired voltage (within limits) so now all that remains is to connect it to the lamps and battery. We have to ensure a few things though.
- When the generator is not running the battery will power the lamps.
- When the generator is not running and the lamps are off the battery is not discharged by electronics.
- If the battery is discharged and the generator runs then we want the lamps to light. Only the excess power is used to charge the battery.
- We should have the option of charging the battery from an external mains powered charger.
- The possibility of using the battery to power or charge external equipment would be desirable, e.g. GPS or mobile phone.
A few extra components can be used to satisfy these requirements.
Circuit description
See Fig 1.
The AC feed from the generator feeds a conventional voltage doubling rectifier comprising C1, C2, D2, D3. The voltage across C2 is typically 22 to 24V. This feeds the switching regulator IC1, this is an LM2595T from National Semiconductor. It has a switching frequency of 150kHz and maximum output current of 1A. The input to this regulator must not exceed 37V so this is limited to 30V by D4, D4A and D5 which are 5W zener diodes.
The combination of D4, D4A and Tr1 ensure that the regulator does not switch on until the rectifier output voltage reaches 21V. This provides some load matching between the rectifier and the regulator by switching the regulator on and off so that the voltage across C2 remains higher than 21V.
The resistors R1 and R2 set the output voltage at 7.5V across C5, this is slightly lower at the system output after passing through Tr4.
Three transistors are used to control the battery connection.
Tr4 switches on only when the generator is working, when it is off there is a slight leakage current of a few microamps. This is less than the self discharge of the battery.
Tr2 and Tr3 switch on the charging current to the battery when the regulator output is more than 6V. When they are switched off Tr2 prevents the battery from discharging through R7.
D8 carries the current from the battery when it is powering the lamps, i.e. the discharge current.
The accessory socket is connected directly to the battery so it can be used for an external charger or to power external equipment.
Fig 1.
Wiring
See Fig 2.
The externally mounted switch has three positions, charge, off and on. In the charge position the generator is connected to the system AC input, this allows charging while cycling. The on switch position connects the generator as before and the system output to the lamps.
The battery pack is also connected to the accessory socket which allows a mains powered charger to recharge the battery, or the battery could be used to power a separate piece of equipment.
Fig 2.
Construction
The circuit board is a long thin design which uses a number of surface mount components. See Fig 3. You have three choices for construction.
- You can get someone to manufacture the PCB for you. A .brd file is available which can be used for this purpose. (Ref 2.)
- You can use the .brd file to print the PCB design and make it yourself. You will need to download and install 'Eagle' software to do this. The software is free for evaluation and is available on all popular operating systems.
- You can make the circuit on stripboard. This gives you the option of not using surface mount components. However you may not be able to make it small enough to fit in a seat tube.
The large capacitors hang off the ends of the board so that it will fit inside a plastic tube. This places some restrictions on the size of the inductor L1 but a suitable component is available from Farnell. This and the use of suitable batteries allows the electronics and battery pack to be mounted inside a plastic tube which itself can be hidden inside the seat tube of the bicycle. If your construction skills aren’t up to using surface mount components then I suggest you go for option 3 above. The switch, fuse and accessory socket are mounted on an aluminium plate which fits under the seat. I found that crimp on ¼” spade terminals with strain reliefs work best for the wiring. Terminal blocks seem to be unreliable on a bicycle. If required, larger capacity 1.4Ah batteries will just fit inside the seat tube.
Fig 3.
Performance
Rectifier efficiency
The unit was bench tested by using an audio power amplifier as the source of the power. This was fitted with a series resistor of 12 ohms to simulate the high output impedance of the generator and fed with a sine wave at 50Hz at various voltages.
The overall efficiency of the system is around 75 to 80%.
Road test
Test conditions
The battery is made up of five Nicad cells and is about 2 years old.
The battery had been fully charged and then used for one journey of 5 miles at an average speed of 12mph. The battery capacity was measured at 1.4Ah
The generator is a Schmidt dynohub supplied by the CTC shop.
An ammeter was wired in series with the battery so that the battery current could be measured directly and the speed was measured with a Cat Eye Mity 3 cycle computer. The bicycle was cycled up and down a quiet street at various speeds and readings taken with the lights on and then with the lights off.
The current drawn by the front and rear lights is 0.42A. This is less than a normal set of dynamo lights as the rear lamp has been replaced by a modified LED lamp. This saves about 80mA of operating current.
With the lights on
The generator starts to provide current at about 6mph but it is not until you reach 14mph that the current drawn from the battery falls to zero. As a result in normal use the battery discharges slowly, in practice taking about 28 night time journeys of 5 miles at an average speed of 12mph for the battery to become exhausted. As the battery becomes discharged the rate of discharge slows down.
With the lights off
Again the generator starts to provide current at about 6mph but the charge current reaches a maximum of 0.53A at 20mph.
The charge and discharge current are equal at about 9mph. If the system were used to get to and from work and the generator ran all of the time then as long as the average speed was more than about 9mph the battery would never go flat.
See Fig 4 for a graph and Fig 5 for the results in a table.
Fig 4.
Speed
Current
Lights on
Lights off
mph
A
A
0
0.42
0
6
0.31
-0.04
7
0.28
-0.11
8
0.22
-0.17
10
0.15
-0.28
12
0.06
-0.37
14
0
-0.45
16
-0.06
-0.49
18
-0.08
-0.51
20
-0.1
-0.53
Fig 5.
Further work
The circuit as presented will work fine with the majority of generators on the market. However if you have a Schmidt dynohub you may find that the generator output is too high under some circumstances. If the lamps are off and you try to charge batteries which are already fully charged then it is possible for the excess power to cause a circuit failure. Normally excess power is dissipated in diodes 4, 4A and 5. Each of these is rated at 5W so in total they can dissipate 15W of excess power. Uniquely the Schmidt dynohub can exceed this which causes the components to become un-soldered from the PCB and the remaining components then fail. This can be avoided by using a modified circuit which does not have the voltage doubler. The Schmidt dynohub does not have one side earthed to the bicycle frame so a standard full wave rectifier is satisfactory in this case. Overvoltage protection diodes are also fitted to the switch and connector unit which is fitted under the bike seat. An alternative set of drawings is available which describes this modification.
Component sources
Most of the components should be easy to source, most of them came from Farnell including:
- LM2595T switching regulator 789-768
- The toggle switch 147-779
- AA Batteries 152-073
- 100uH inductor 432-088
and assorted surface mount components.
Appendix
(1.)
Input impedance of a DC to DC converter.
The negative input impedance can be easily explained with an example.
Let us imagine that we have a perfect DC to DC converter which is 100% efficient. It is designed to supply 6V at the output and we have connected a load which will draw 1A This means that the output power is 6*1=6W
We supply the circuit with 12V. Since the converter is 100% efficient then we require 6W from the source. At 12V this means we need 0.5A
Now suppose there is a problem with the input and the voltage falls to 9V. We still have 6V at the output because the circuit adjusts itself to maintain this therefore we still supply 6W at the output, the efficiency is still 100% so we still need 6W at the input. At 9V the current required is 0.666A So the input voltage went down but the input current went up, hence the input impedance is negative.
(2.)
PCB manufacture.
The circuit design was done using Eagle software. You can download a trial version of this from Cadsoft.
A company called PCB-POOL can manufacture PCBs in single quantities. It's cheaper to gang together with friends to get more than one made but it is possible to get just one board made if you have to. You need to send them the .brd file.
See PCB-POOL for more details.