Solar charger

user warning: Unknown column 'i18n.language' in 'where clause' query: SELECT DISTINCT b.* FROM blocks b LEFT JOIN blocks_roles r ON b.module = r.module AND b.delta = r.delta LEFT JOIN i18n_blocks i18n ON (b.module = i18n.module AND b.delta = i18n.delta) WHERE (i18n.language ='fr' OR i18n.language ='' OR i18n.language IS NULL) AND ( b.theme = 'garland' AND b.status = 1 AND (r.rid IN (1) OR r.rid IS NULL) )ORDER BY b.region, b.weight, b.module in /var/alternc/html/o/orangeseeds/usr/drupal-6/modules/block/block.module on line 456.

Following our research on rechargeable batteries and the meeting with Bartaku we ran experiments using a simple solar charger.

The circuit

The circuit is pretty easy to build: basically you plug the solar panel straight into NiMH batteries and in your circuit. You need to put a diode between the solar panel and the batteries to avoid the current to flow back from the batteries to the solar panel. To avoid the diode to take a big part of the voltage (silicon diode take 0.7V) preferably use a low consumption diode such as a Schottky (0.2V) or Germanium diode (0.3V). You also need to make sure the solar panel won't damage the batteries by adding a resistor (if necessary). An alternative would be to use a regulator such as LM317T.

Experiment

We ran experiments using two solar panels:
  1. Panel A was bought at a local store in Brussels, it is of type "crystalline" (best effective under direct sunlight) and generates 2V max.
  2. Panel B was part of a "garden lamp" bought in Montreal (the same with which we made the bird modules, it is of type "amorphous" (which works well even if there is shadow) and generates around 3-4 V in plain sunlight

We also used two types of NiMH batteries:

  1. Batteries 1 are 1.2V, 600mA batteries that were taken from those same "garden lights". We know they are pretty bad cause we tried to charge them with a battery charger and it didn't prove very efficient.
  2. Batteries 2 are also 1.2V ??? mA, we bought them in Montreal

Results

Best setup is Panel B with Batteries A (we use two batteries ie. 2.4 V). After 1h charging under lamps it provides 1.8 V and the batteries are able to keep a LED alive without light. Measured current in circuit if we cut sunlight is 200 mA.

I was not able to drive a single ATMEL chip circuit with that however.

Conclusion

Good practices:
  1. Use amorphous-type solar panels
  2. Use solar panels that drive enough current/voltage
  3. Use good NiMH batteries

We need to work on the question of current (A) as well as that of voltage (V). Now we get proper voltage but seemingly not enough current to drive our circuits.

References

We suggest you follow the following detailed instructions if you plan to built your own circuits:

Commentaires

Overcharging issue

This article points out to an important issue that we had not taken care of: overcharging. Apparently, NiMH batteries are prone to overcharging: when they are full, the input current should just stop flowing or should at least get lower.

Simple trick

The most simple trick to prevent overcharging seems to be: make sure your solar panels... won't overcharge the batteries! If you have sufficiently big batteries (enough mAh) the solar cells will just never be able to fully recharge them and thus overcharge them.

To estimate the number of hours needed to fully recharge batteries, use the simple calculation: I / C * 120% where I is the batteries capacity (in mAh) and C is the solar panels output (in mA). The 120% is an approximation that we will lose about 20% of the charge. Of course, for a solar panel, we also need to consider the efficiency of the cells under the sun (e.g. in cloudy weather the efficiency drops).

You can alternatively use an online charging calculator such as this one to ease your work.

I've read elsewhere that it's possible to limit overcharging by simply adding a resistor between the diode and the batteries.

AVR-based circuit

Yet another circuit using an AVR instead of a PIC. It's pretty complicated but maybe there's a way to simplify it.

Solutions?

Most of the solutions I found to prevent overcharging are rather complicated. After a lot of research I found a few ones that look rather easy.

A bit of theory is given here about charging algorithms and voltage slope detection. It's worth reading. If NiMH batteries are charged at a constant current the voltage will increase but will eventually reach its peak: this is the moment where we need to stop charging.

The MAX712/713 are electronic ICs that do all the job for you for only 3 USD.

This link suggests to use a thermistor put close to the batteries. When they start to overcharge, they will heat. With a thermistor plugged to one of the analog pins of the AVR you can measure that heat and stop charging (e.g. using a transistor or relay) once it's reached a specific value.

This post shows a rather complex circuit using a PIC to charge batteries. It includes a fast-charge algorithm that I believe could be dropped. The PIC could easily be replaced by an AVR. There is a great discussion about NiMH and NiCd batteries, with the following advices:

  • You don't have to flatten your battery before you recharge it,
  • Don't flatten your battery below 0.8V per cell,
  • Don't overcharge your battery beyond 100% of C, and
  • NiCd/NiMH don't like to get too hot, or too cold (0 to 40 degrees C is usually best)