To Boost or Not to Boost..That is the Question!

[Beeper]

While I am waiting on a few parts for a toaster hack, I have a burning question:

I have built a bicycle light out of a large array of high lumen white LED's using a MCU. The MCU is blinking the LED array with a 5V Vdd supply. Several magnets mounted on the spokes pass by a coil on the frame and generate pulses that are rectified by a Schottky diode bridge to feed a circuit to charge a Li-Ion battery.

Since the battery will eventually be charging at 4.15+ V near the end of charge, the rectified voltage from the wheels must reach approx 4.75V to deliver energy to the battery.

Because of the 4.75 threshold of charging, Is it that I am only using a small portion of the available energy pulse (that voltage greater than the charging threshold voltage) delivered from the coil?

If instead, I feed the rectified output of the generator coil to a boost converter IC (it will start to work way down at .9V Vin) to boost-convert the generator output to 4.18V  to charge the battery, would I get better results?

Someone else has told me that this is just spinning my wheels, the 75% efficiency and loading of the generator coil would offset the benefit of a boost-converter.
 

[tom66]

In any kind of full-wave rectifier off a generator, you will only be using a portion of the WAVEFORM, however you will not be wasting significant energy as if you load the generator more, it will require greater input energy to turn. During the period in which your circuit is not conducting, the generator has no effective load. The only loss is friction and perhaps some parasitic magnetic effects / eddy currents etc.

If the energy required to turn it is negligible, you might want to consider a way to harvest the energy - perhaps some kind of psuedo active-PFC circuit, which would utilise more of the waveform, although with a low input voltage, the advantages are going to be fairly minimal.

[Beeper]

Once thing I have observed with my meter, the voltage output of the generator pickup coil is proportional to speed. So, if I am going slow on the bike, I will not have sufficient drive voltage from the generator to become charge voltage, so methinks a boost converter will give the charging a boost!

Right now, after approx 1 month of use of a fully charged Li-Ion battery(I charged before putting it in the bike light), the battery needs to be taken into the house for a charge.

The whole thing seems to due to the Winter Effect, I am using the bike more hours in the dark than in the day, it seems the light of day is growing short.

[tom66]

In which case a boost converter may work. You might want to investigate controlling the switching current of such a converter to be proportional to the input line voltage (active power factor correction.) This would reduce resistive losses in the coil whilst extracting the maximum energy possible.

[Seekonk]

I would go with a simple diode voltage doubler in that application.

[Beeper]

I can't find a PFC correction boost converter for .9V Vin to 4.15V boost conversion.

The other thing: little room left in the light for another large electrolytic capacitor to make a voltage doubler.

Right know I am using Schottky rectifiers to make a bridge with a 1200uF/6.3V cap for filtering prior to the charger circuit.

The MCP1623/24 seems to me to be the best candidate I've found so far if boost beats what I have to blink'n charge with now.

PS I am using the coil inside a 1.8V LED ordinary self-powered bike blinker  I stole the coil output and ran wires up the frame to charge the battery mounted inside tne case of  my bike light mounted on the front bike basket.

[mikerj]

A coil with more turns is possibly an easier way forward.  The output will be proportional to speed (or more accurately rate of change of magnetic flux) and loading so in the event of a fully chanrged battery you will need to have some scheme to prevent the voltage rising to potentially damaging levels.

[Beeper]

The generator pickup coil is coated/potted in epoxy, so the idea of rewinding looks harder to stomach than blackened burnt toast for breakfast.

I just don't have the desire to play with fine wire or chip away epoxy.

I already have a shunt regulator to clamp the Max charge voltage applied to the battery  to  <4.2V, excess voltage lights LEDs to regulate, A2D controlled by the MCU.

The question is, "Does a boost approach allow me to coast and charge?"

Am I trying to pedal the idea that low-voltage coil output has a charging potential?

Isn't that the condition of high coil voltage output  also presents the greatest charge current,  and that while I am coasting downhill,  lower output voltages are also sources of  low current anyway?

Swapping out the Li-ion cell once a month has the benefit of giving me a little more much-needed additional exercise.