A Few Words About How the Circuit Works

Inspired by a recent post about a budget Chinese emergency lamp, I decided to design my own circuit from scratch using the fewest components I could find in my scrap box—including a very old Li-Ion battery from a dead phone.

The circuit's purpose is simple: charge a lithium battery during daylight while keeping it within safe voltage limits and power an LED (LD4) in the dark without over-discharging the battery.

It operates with just two transistors—one managing the charging phase and the other the discharging phase—along with a handful of diodes and resistors.

Below is the link to the circuit simulation I created on Falstad.com (thank you, Paul!). Feel free to experiment with the values and observe how the circuit operates: 

https://tinyurl.com/2b22kt2q

-----

Charging 

When the sun shines and the solar panel generates sufficient voltage, the circuit charges the battery (simulated by a capacitor). Refer to the diagram below:

The current passing through the 2.2k resistor (R2) drives the transistor Q1 into saturation. This allows current to flow from the collector to the emitter, charging the battery. Once the battery charges sufficiently and the emitter voltage reaches V_LD3+V_DZ1−VBE, the transistor stops conducting. With an appropriately chosen Zener diode and LED (I had to make some adjustments by selecting the appropriate Zener diode and the correct LED color), this should occur at approximately 4.2V.

It’s important to note that the battery will still be charged slowly through the 47k resistor, but the assumption is that the resulting current is small enough not to damage the battery.

The PNP transistor on the right (Q2) remains in cut-off because, as long as the solar panel is illuminated, its base voltage is higher than its emitter voltage. This ensures that the main LED (LD4) remains off.

-----

Discharging

When the sun goes down and the battery is charged, the light will turn on. See the diagram below for reference.

This occurs because current from the battery flows through the emitter–base junction of the PNP transistor Q2, then through the 2.7 kΩ resistor R1 and the two LEDs (LD1 and LD2). This biases Q2 into saturation, allowing current to flow through its collector and power the white LED (or LED chain, in my case), producing the “fairy light” effect.

As the battery voltage drops below approximately 3 V (depending on the components used), the current through R1, LD1, and LD2 decreases. As a result, Q2 transitions from saturation to the active region and eventually to cut-off, thereby reducing the discharge current.

At that point, the battery continues to discharge slowly through R3, R1, LD1, and LD2, with a current of only a few microamperes. Consequently, as long as the solar panel receives light the following day (or within a few days), this residual discharge should not pose any issue for the battery.

Along with the simulation, it's useful to test the circuit in real life using a capacitor instead of the battery. This lets you verify the voltage levels and fine-tune the components to suit your specific needs, such as considering battery specifications (min and max voltage), output current requirements (set by R1 and R4), and the characteristics of your solar panel.

My final circuit is the following (LD1 and LD2 red, LD3 green):

but you may want to change the values of some components (LD1-LD2 or R4) to match your load and battery requirements.

Have fun!