Build your own bicycle electric generator
Written by William Lloyd. Updated on 26th December 2009.
1. Introduction
Do you have a bicycle turbo trainer? Have you ever wondered if you could harness some of that wasted energy?
This page is for anyone who is interested in building an electricity generating system based on a commercial bicycle turbo trainer (figure 1) or the DIY equivalent (figure 2).
Figure 1. Bicycle on a modified turbo trainer charging a 12 V leisure battery
Figure 2. Bicycle on a DIY support stand connected directly to a portable radio
The principle elements of the electricity generating systems shown in figure 1 and figure 2 are described first. Modifications and ideas are discussed in the final sections.
2. Bicycle requirements
The systems shown in figure 1 and figure 2 are based on "ordinary" bicycles mounted on a commercial turbo trainer and DIY support stand respectively.
Commercially available turbo trainers usually require that the rear wheel is secured using a quick release system. Older bicycles may not have quick release wheels. My DIY support stand is for such a bicycle.
I have not tried a bicycle with knobbly off road tyres, but the instructions for the turbo trainer state that "…knobbly tyres cause noise and vibration. Using a slick tread will improve the performance of your trainer".
Gearing needs to be sufficient such that the generator turns quickly enough (see Section 5). For example consider a bicycle with a 52 tooth front chain ring, a 14 tooth rear sprocket and a 27" rear wheel. A typical diameter for the roller mounted on the generator shaft is 42.1 mm (see Section 6). At a reasonable pedal speed of 60 rpm (revolutions per minute) the generator speed would be 60*52/14*675/42.1 = 3573 rpm. Note the figure of 675 represents the effective diameter of the rear wheel in millimetres as the tyre presses on to the generator roller.
3. Modifying the turbo trainer
The turbo trainer I have used is a CycleOps Magneto. These are readily available. I bought mine from www.wiggle.co.uk. Figure 3 shows the turbo trainer in its original state.
Figure 3. CycleOps Magneto turbo trainer
The idea is simply to replace the original resistance unit bracket assembly with a new bracket that can be used to support the generator. My new bracket fits in to the same mounting points on the support stand as the original bracket. Thus, if so required, the turbo trainer can easily be returned to its original state. The geometry of the bracket is the same as the original, except that allowance has been made for a larger roller diameter (55.5 mm compared to 35 mm for the original). It is made from two 190 mm lengths of 60 mm * 60 mm * 6 mm thick metal angle cut accordingly and bolted together to form a channel shape. I have made a new tensioning rod (for pulling the roller on to the bicycle wheel) from an M6 threaded bar bent to suit. The necessary pivoting action for this rod is provided by a 16 mm diameter bar resting in a cut out in the bracket. Figure 4 shows a close up of the bracket assembly and figure 5 shows the tensioning rod pivot.
Figure 4. Generator bracket assembly
Figure 5. Tensioning rod pivot
4. Building a support stand
The support stand shown in figure 2 is made from 60 mm * 60 mm * 6 mm thick metal angle bolted together. The two "legs" on either side of the bicycle are joined with threaded bar. The generator bracket pivots on the rearmost bar. Tension of the generator roller against the bicycle wheel is maintained simply by tightening the nuts on the threaded bar. A tensioning spring is shown in the figure but this is not really necessary. Figure 6 is a view of the rear of the stand.
Figure 6. Close up of DIY support stand
5. The generator
The generators in figure 1 and figure 2 are both permanent magnet DC brushed motors. This type of motor when run in reverse, that is when the motor shaft is turned, will act as a DC generator. The permanent magnets provide the field excitation thus the relationship between output voltage and current is linear with negative slope equal to the armature resistance. The open circuit voltage (no current flow) is directly proportional to the speed. In other words, the faster the generator shaft is turned the higher the voltage. The relevant equation is:
V=E-RaI
where V is the DC voltage at the armature (output) terminals, E is the emf (E=kaΦn where kaΦ is a constant and n is the rotational speed) and Ra is the armature resistance.
The motor has a rated voltage and power at a particular speed. When used as a generator this essentially means that it can transmit this amount of power continuously at this voltage and speed.
The generator in figure 1 is a 24 V PM60 model from Parvalux (www.parvalux.com) rated at 210 W at 3000 rpm. When charging a 12 V battery the generator voltage needs to be greater than the battery voltage otherwise no current flows and consequently no charging takes place. In fact without a blocking diode current will flow in the reverse direction and the generator will act as a motor (see Section 7).
The generator in figure 2 is a 12 V PM90 model from Parvalux rated at 300 W at 4000 rpm. In figure 2 the generator powers a portable radio with a required DC input voltage of 4.5 V. The power usage of this radio is very modest (6 W) meaning that the pedal load is low. In this situation you are doing little more than spinning the pedals. All you have to do is watch the voltmeter (see Section 7) to keep it at a more or less constant 4.5 V. Radio signal interference can be a problem (see the note at the end of Section 9). This generator is less suitable for 12 V battery charging because it is difficult to achieve the required pedal speed to maintain a charging voltage of 14.4 V at reasonable pedal loads.
6. The roller for the generator
A roller needs to be manufactured to fit on to the generator shaft. The larger the diameter the roller the less frictional losses there are between tyre and roller. Smaller roller diameters require that you press the roller harder in to the tyre to transmit the same power. The original roller diameter on my turbo trainer is 35 mm. I have tried several different diameter rollers: 29.8 mm, 42.1 mm and 55.5 mm. I found that the 29.8 mm diameter roller was not at all satisfactory because it required pressing in to the tyre so hard to avoid slipping that most of the pedal energy was wasted in friction. The 55.5 mm diameter roller is shown in figure 1 and figure 2.
I made the rollers from mild steel bars from a scrap metal yard. I turned and faced them on a lathe. The roller is secured to the shaft using a grub screw acting on a flat ground on the shaft.
7. Monitoring electrical output and preventing motoring when battery charging
In figure 2 the generator is connected directly to a DC load. The DC load is a portable radio. The generator positive is connected to the positive in the radio battery compartment and the generator negative is connected to the negative. The only monitoring of electrical output required is measurement of the voltage across the positive and negative terminals of the generator or radio. I have used an analogue multimeter (no batteries required).
When battery charging, figure 1, there are three main issues to consider. Firstly, a blocking diode is necessary to prevent the generator acting as a motor when the generator output voltage falls below the battery voltage. Secondly, the battery is best charged at a certain rate (see Section 8), so an ammeter is useful. Thirdly, battery voltage is an indicator of the state of charge of the battery (see Section 8), so a voltmeter is useful. Figure 7 shows the circuit diagram for my battery "charge controller". The fuse prevents damage to the cable and components.
Figure 7. Circuit diagram for battery "charge controller"
Solar photovoltaic or wind turbine battery charge regulators normally incorporate some method of "dumping" energy when the battery is full (indicated by the terminal voltage across the battery reaching a certain threshold). The purpose of this is to prevent dangerous over charging of the battery. I have not included this in my circuit simply because unlike wind or solar systems, you are the energy source and you can simply stop pedalling when the battery is full.
My battery "charge controller" can be seen tied on to the handlebars of the bicycle in figure 1. The parts are as follows:
I have sized the components and cable based on the charge rate requirements of a battery with a capacity of around 100 Ah (as seen in figure 1). I have used a 1.5 mm2 cable for which the tabulated current carrying capacity in the IEE Wiring Regulations (BS7671) is 16 A, but I have used a 20 A fuse. A cable with a current carrying capacity greater than the fuse rating should really be used, however you are not going to be producing maximum current flow continuously so I didn’t consider the extra cost of a bigger cable justifiable. I have used 3 core cable instead of 2 core simply because it is more readily available in short lengths.
The voltmeter listed above does not seem to be particularly accurate compared to a digital multimeter. For more accurate voltage readings I have used a digital multimeter.
8. Battery charging
Deep cycle lead acid leisure batteries such as those used in caravans are the most suitable and readily available type of battery to use. One cautionary point about their use is that they should not be recharged in enclosed spaces due to potential gassing.
How the battery is charged is a significant factor in determining the battery life. Consider the charging rate in amps as a proportion of the battery capacity (C). For a fully discharged battery the current charging rate should be limited to C/20 until the battery has reached 20% state of charge (5A for a 100Ah battery). From 20% to 90% state of charge the current charging rate should be limited to C/10 (10A for a 100Ah battery). From 90% to 100% state of charge the current charging rate should be limited to C/20 or, if possible, charging voltage maintained at the battery’s fully charged voltage level. Typically the battery should not be discharged to less than 20% state of charge.
I have a maintenance free, sealed for life 100Ah battery from Elecsol. In the guidance information on charging it is stated that fully charged the open circuit voltage (voltage with no current flow to or from the battery) is 12.80V, 50% discharged the open circuit voltage is 12.40V, and fully discharged the open circuit voltage is 10.70V. The guidance goes on to say: Never over charge the battery. The battery is fully charged when the voltage reaches 14.40V. Once disconnecting the charge and after 24 hrs a fully charged battery should have a voltage of 12.80V. Never over discharge the battery. The battery is fully discharged when the load voltage is 10.70V. Also, keep the battery in the highest possible state of charge.
Another battery that I have (Exide 85Ah) has guidance information stating that the battery should be recharged when the voltage drops to 12.3V.
When charging using the bicycle generator you will unlikely be much above the 5A to 10A (C/20 to C/10 for a 100Ah battery) level for very long. To charge a 100Ah battery from 20% to 90% state of charge at 10A (C/10) will take 7 hours. This is a lot of hard cycling! I have found that the main problem with the set up in figure 1 is actually assessing the state of charge of the battery. As alluded to in the guidance information for my 100Ah battery (described above) the actual state of charge of the battery should be determined by measuring the voltage 24 hrs after charging. However, it does say that the battery is fully charged when the charging voltage reaches 14.4V. I am therefore aiming for a voltage around 14.4V as I pedal. This would indicate a full battery. The battery voltage after 24 hrs should then be 12.8V. To complicate matters further my 85Ah battery seems to charge to a higher voltage value thus implying a higher charging voltage when pedalling.
An additional practical point is that even when pedalling at a more or less constant rate you will notice that the current output varies a lot. Although it feels like you are putting in an even effort, you aren’t. To confirm this I have tried powering the generator with a power drill. The result was a constant current reading. A possible modification (Section 10) to smooth the current output would be to fit a flywheel to the generator shaft.
9. Using battery power
The most efficient use of the battery energy is to power 12 V DC appliances directly from the battery.
If 230 V AC power is required an inverter is necessary to convert the 12 V DC from the battery to 230 V AC. Obviously, the power rating of the inverter needs to be greater than the power rating of the appliance. How much greater depends on the appliance because some appliances will draw significantly more current at start up than their power rating would indicate.
For battery charging the bicycle generator could be connected just to the battery as shown in figure 1. However, to allow me to use the battery power I have wired up a simple system following the schematic diagram shown in figure 8.
Figure 8. Wiring of battery system
The idea behind this system is that I can run 12 V DC and / or 230 V AC appliances while cycling and while not cycling (with or without the bicycle generator connected). In the former case (while cycling), if input effort is matched to the electrical load the battery is simply a convenient power smoothing device. Figure 9 is a picture of the battery system with a 12V DC radio and a 12VDC lamp. Figure 10 is a picture of the television that you can watch powered by your own efforts. It is 230V AC.
Figure 9. Battery system
Figure 10. Portable television
A practical point to note with respect to television and radio is possible signal interference from the generator. A piece of wire attached to a portable aerial may be required! External aerials solve the problem.
10. Modifications
With the aim of improving the efficiency of power transmission I have tried the following modifications:
1. Cutting notches in to the 29.8 mm diameter roller to increase the surface area available for contact against the tyre. This did help, but not significantly. I could have tried again with a better shaped notch, but I didn’t think it was worth the effort.
2. Fixing two joined bicycle chains around the rear tyre, using tyre pressure to hold it; then mounting a derallieur idler gear on to the generator shaft and pressing it against the chain. Figure 11 is a close up of the gear and chain. This worked and the pressure needed to hold the gear against the chain was not too excessive (I used the spring shown in figure 2 and figure 6 to maintain the pressure). However, it was noisy.
Figure 11. Close up of gear and chain with chain fitted directly on to the tyre
3. Removing the rear tyre and inner tube and attaching the two joined bicycle chains around the wheel and idler gear on the generator shaft. Figure 12 shows the set up. I maintained the chain tension simply by tightening the nuts holding the generator bracket on to the threaded bar. This was better than the previous solution (2), especially with the generator bracket moved further away from the wheel, but it was still noisy.
Figure 12. Gear and chain with chain around rim
4. As for (3) above, but replacing the chain with a long tumble dryer belt and fitting a pulley on to the generator shaft (figure 13). The belt is the longest I could find at 2083 mm. It has a j4 belt profile and is obtainable as a belt replacement for a Bosch T470, T473 and T475, Haka DR72, and Siemens Siwamat 400, 470 and 480. I machined the pulley to match the belt profile and added a tensioning bar between the motor bracket and the support stand. This is probably the best solution, however, I did find that there was a curious belt resonance at certain speeds.
Figure 13. Belt and pulley
In summary, these modifications have been worth looking at, but the reward is difficult to justify versus the simplicity of the plain roller against tyre solution.
I have not tried fitting a flywheel to the generator output shaft as described in Section 8. My theory is that a flywheel would provide inertial mass to smooth out the current output. The disadvantages would be the extra loading on the generator bearings and the possibility of imbalance.
If you have a wind / photovoltaic system there is no reason why you cannot connect the bicycle generator to the charge controller of the system as another energy source, provided of course that the voltage range of the bicycle generator is appropriate and that the maximum current is not exceeded. I have tried this with a Marlec RWS60 charge controller (a small wind turbine and photovoltaic panel controller designed for battery charging).
In theory you could connect the bicycle generator to the mains electricity grid and sell the energy back to the supplier. However, you will need to obtain permission from the local Distribution Network Operator and use an inverter meeting G83 quality requirements. The cheapest suitable inverter I could find costs around £1450 (Windy Boy WB-1100-LV with a maximum DC voltage of 60Vand a maximum current of 62A). As this is much to expensive as well as much to big for one bicycle generator, I have shelved this idea without making further inquiries.
11. Comments and Suggestions
All comments are my personal findings and opinions at the time of writing. Any errors are unintentional but are certainly possible, especially as I am still experimenting.
In this age of cheap electricity from the mains electricity grid this type of energy generating system does not have mass market potential. Why do it? The answer must be because you want to or because you have a particular "off-grid" power requirement.
Any comments, suggestions, stories or business ideas would be gratefully received. Thank you.
My email address is bill@bicycleelectricgenerator.co.uk