Let's say you have a remote weather station, remote being defined as located on top of a pole. How would you power it?
I invented that problem in order to demonstrate the solution, which was the practical use of solar power. Because the nature of the project is actually educational, and the subject is solar power, I wanted to design the circuits to mimick as closely as practical the behavior of a "real" solar power installation - with panels, a converter, charger, and distribution system.
When you take a panel that puts out 8.4V and use it to trickle charge lithium ion batteries, you are breaking every rule in the book. You can't trickle charge lithium ion batteries. You might destroy them. Yet it will work fine, and there are examples of this type of lithium battery charging on the web. I have seen circuits as simple as the one below.
The circuit above would result in fire were it not for one immutable truth: the sun goes down every night. That gives the lithium battery time to stabilize back down to 3.6V/cell. So if you are looking for a cheap and easy way out, that might be it. In this project we are not looking for either cheap or easy. The simple charger above has only limited application. If your solar panel isn't putting out 8.4V, it won't work. If you need a single lithium cell, it won't work. If the Earth stops spinning on it's axis, it won't work. Well, it is a variable in the design.
Li-Ion batteries are particular about how they are charged, and require a constant current deep charge and a constant voltage saturation charge. Trickle charging a lithium ion battery is a very bad idea, so the charger has to monitor the charging current and battery voltage, and determine when the charging current needs to be turned off completely. See this Battery University article for a detailed explanation of the process. A good charger would be flexible enough to allow for different battery and solar panel configurations.
The converter in this system is a switching step-down regulator using an LM2596. They come as complete modules from China, for less than the cost of the IC. It drops the solar panel voltage to 8.4V, which is the maximum charge voltage for a pair of series connected Li-Ion batteries. Current from that supply will be used to both run the weather station and charge the batteries.
There are some assumptions made up front:
So the situations we need to design for are laid out and we can put some numbers to them.
The conclusion from all of this is that the solar panels will provide enough power to recharge three cloudy days' usage in one day. Now the batteries. The smallest reasonably priced batteries I could find were 14500 size, which is so close to AA as to be functionally identical. They are rated at 2000mA/Hr each. Way overpowered, but it won't matter. They can run the system for 8 days before recharging, but it would take three daylight cycles to get them back to a full charge. It's Ok - they satisfy the requirements.
The schematic of the finished supply is seen above. There are a few extra measurements for tracking what the power supply is doing. You probably wouldn't want that. The regulators are completely adjustable, so as long as Vsolar > Vbatt > Vload this configuration should work. An alternative might be to run a step-up from a lower voltage panel to one or more batteries, then step-down to the required voltages.
The power supply normally does no battery charging. The solar panels feed the 8.4V regulator, which provides power to the power supply circuit, and the weather station. When the weather station MCU senses the 8.4V output, it checks the battery voltage. If the battery voltage is below 8.2V the current is increased in the charging circuit, by increasing the duty cycle of the PWM on the "Set In" line. It is increased slowly, until the solar voltage begins to drop below 13.8V, or the current exceeds 75mA. Then it stops increasing the current. Since the control circuit is a current source, it does not require realtime monitoring to keep the current steady. Once each minute the system will poll the current and voltage. In my case, all of the currents and voltages.
When the battery voltage increases to over 8.3V, the current is monitored, and when the current drops below 75mA, the charging is stopped. Of course, if the solar voltage goes away, charging stops.