Solar Disc Light DIY –

This post introduces the design of a little solar disc light. It is based on a handful of electronics components. By “solar disc light” I am referring to a compact light source in the shape of a flat circular plate which also houses a minuscule solar panel and a rechargeable battery. I assume that beginners want to build their model like this one. This design does not have a microcontroller, but it relies on a few off-the-shelf ‘analog’ parts which you can pick from your nearest electronics storefront.

Power Source:

The main requirement I had for this project was a round solar panel, needed to frame within an annular enclosure. This needs to recharge a small lithium-polymer battery pack as well. This is the photograph of the solar panel (6V/80-100mA) I bought from a web merchant.

As for the backup battery pack, the cheap single-cell lithium-polymer (1S LiPo) battery pack (3.7V/600mAh RC Toy Battery) seemed to be perfect for the project on account of its relatively low profile and inbuilt protection features.

The key constituents are going to be managed neatly by a linear regulator in constant voltage configuration, so I tried to rig up an initial setup as depicted below. What I had in mind was to work on a reasonable budget and then do necessary tweaking on the prototype.

As you can see, the above circuitry is simply a constant current source (the solar panel) driving a shunt regulator circuit. The ‘unorthodox’ lithium-polymer battery charger setup, centered on the adjustable shunt regulator chip TL431 and the standard pnp transistor S8550, is configured to render a constant voltage output just about 4.2V (TP1). The charger circuit strictly follows the typical application example found in the TL431 datasheet (see below) but the usual resistor in series with input (Rsup) is not used here. It’s not essential in this particular application as the series impedance of the constant current source (solar panel) takes its place and should be quite high in dynamic resistance if it works as a good constant current source.

The maximum current to be sunk in the regulator is 100mA x 4.2=420mW dissipation essentially all lost in the S8550 transistor, well within its rated maximum!

If the solar charger works 6 hours in a day at 100mA, that is 600mAh. One sunny day is enough to fully charge the 600mAh battery. If you wanted to use a larger solar panel, note that the S8550 is good to around 700mA.

I rigged up the minimalist electronics on a mini breadboard and used a 6V/100mA constant current source to test the concept. I got around a stable 4.2V output at the test point TP1.

Things weren’t quite perfect though as I measured a 300mV drop across the 1N5819 diode at 100mA so the LiPo battery can see only 3.9V! Since the voltage drop across the Schottky diode depends on charging current, adjusting the TP1 voltage to 4.4V should result in good charge rate up to battery voltage of 4.1V, and charge tapering off to mA by 4.2V. Relying on a crude shunt regulator for charging a LiPo battery, rather than using a devoted LiPo charger circuitry is a bit sketchy, however, the LiPo battery pack looks like it’s pretty robust enough to work with the destined scheme.

Merely substituting one 10K multiturn trimpot for the 6K8 (R1) resistor in the basic schematic permits fine adjustment of the charger voltage. After toying with a couple of thoughts I revised my basic design as presented below. The revised circuit is the same as the first one, just with an adjustable voltage set point. I also substituted the 510Ω resistor (R3) with a 470Ω one – a bit lower resistance value, but that’s good in some respects.

What is the role of the resistor R3 here? The TL431 (U1) has a nominal internal reference of 2.495V, and an op-amp driving a transistor. As now we have the pass transistor, we need to feed a feeble current to TL431 to generate the requisite voltage. To conduct current, the S8550 transistor (T1) calls for a base current which is then multiplied by the transistor’s hFE. In my design it is estimated at 2-3mA. This project calls for tight-tolerance resistors.

One word of advice – be very careful using a cheap breadboard jumper wires for projects with more than 50mA of current. I started with a few of them until I looked at the baffling high resistance (~1Ω to 2Ω) introduced by them. A typical cheap jumper cable usually has a thin wire and poorly crimped header pins and adds substantial resistance (unless gold-plated)!

Light Source:

When it comes to the light source, things can be debatable because the LiPo battery’s operating voltage range of 2.75 to 4.20V from discharged (discharge cut off 2.75V) to fully charged can be a bit of a hassle, especially if working with powerful white LEDs!

With the fully charged LiPo battery, the output voltage will start around at 4.2V, and the nominal output will be 3.7V. This is enough to drive a single white LED or several white LEDs in parallel (not a wise method) certainly within the limits of the real capacity of the LiPo battery. I decided to go with a low-current light source to make the most of the battery runtime. A cheap and pretty reliable light source would be a small string of low-current, high-brightness LEDs wired in parallel. I might replace the intended light source later with a bunch of series-connected LEDs driven by a high-efficiency dc-dc boost converter! Does this make sense?

When talking about a LiPo battery, the primary characteristics to understand are the battery’s voltage, capacity, and discharge rate. The voltage of a single-cell (1S) LiPo battery is 3.7V nominal (4.2V fully-charged). The capacity denotes the capacity of the battery in milliamp-hours (mAh). A fully charged 600mAh pack is rated to cater a current of 600 milliamp (600mA) for one hour before it is fully discharged. The discharge rate lets you to determine how many amperes the battery can output continuously without becoming damaged. For example, a LiPo battery with a discharge rate of 10C means you could safely draw it at x10 capacity of the battery pack (The discharge rate is commonly expressed as a multiple of C). Hopefully, this side note has not scared you away from playing with LiPo batteries. For further details, checkout these links:,

Now you can see the light source used in my experimental model. For a quick test I used a parallel combination of 8 generic straw-hat warm white LEDs (4.8mm/3.3V/20mA). The total current consumption estimated is around 160mA, thus retaining the LiPo battery’s discharge rate within cheerful limits.

An LED emits light at an intensity that depends on the current passing through it (not the voltage). Another way to describe this is that an LED is a current-driven device rather than a voltage-driven device. A dropping resistor is normally placed in series with the LED to reduce the supply voltage to the LEDs forward voltage (Vf) as used in the below scheme. Typical Vf of the white LED used here is 3.3V, so the amount of resistance needed is 3.7V-3.3V/20mA = 20Ω. That is okay but often Ohm’s law does not work well for LEDs. The resistance of an LED is not constant, it varies as voltage or current is deviated. This is known as Dynamic Resistance that is a bit hard to calculate (more on this later)!

Random Runs:

Before getting out the soldering iron and making the circuit permanent, my breadboard prototype was successfully tested with another LED light source – 7 white LED aluminum strip (4V/35mA x7) – too. See below images. For the permanent version of the solar disc light (v1) I used a small piece of strip board, and that dirty prototype has been spending night for field-trial as a little gear in a farmer’s shed in darkness before!

This might be too much enthusiasm for a simple solar disc light build, but I’ve learned something from it, and felt like sharing. Thanks for reading!

Leave a Comment