Solar energy harvesting often relies on highly efficient switching ICs with inductors, which are great at extracting tiniest amount of energy power from a solar panel at low-light conditions. However, these chips can be expensive, typically costing between 2 to 4 euro. In one of my earlier projects I found low-cost linear battery chargers, such as the TP4056, priced at just €0.25, can reach 75% efficiency thanks to their low dropout voltage. And when the solar panel voltage is well-matched to the battery voltage, the solar panel operates near its maximum power point, making this approach an intriguing alternative worth exploring.

The perception of light intensity is highly non-linear. Indoors there is roughly 100 times less light energy to harvest than outdoors. Consequently, for many applications that sometimes see outdoor light, only few percent of energy from indoor light, it really doesn't contribute. Therefor many projects don't need an expensive switching energy harvesting IC. It is better to spent money on a larger solar panel.

Lithium Ion Capacitors (LICs) bridge the gap between supercapacitors and lithium-ion batteries. Compared to batteries they offer a wide operational temperature range (-20°C to +60°C), 30k+ charge cycles, no hazardous chemicals, and no shipping restrictions. Where a Li-ion battery would fail in the cold and wears after a few years (after 500 charging cycles), LICs continue to work. They are ideal for outdoor renewable energy projects with lifespans exceeding 5 years.

In 2022 I thought 250F 3.8V LICs are big, in 2023 I was impressed by 750F 3.8V LICs (1840 size), in 2024 I found 1100F 4.0V LICs (1840 size) and in 2025 I acquired a 1300F 4.2V LIC (1840 size) . Today a 250F 3.8V is only 1.65 euro and a 1100F 3.8V LIC is 4.82 euro. The emergence of kilo Farads, the increasing voltage and the dropping price makes them more and more interesting. 

For me, the logical next step is designing designing a linear solar charging circuit optimized for LICs.

In this project I have designed two small PCBs with two LDOs and compare it to an another PCB with an expensive switching solar energy harvesting circuit. The first LDO is a 45 cents 4.0V MCP1702 from Microchip and the second is a 12 cents 4.0V LDO HT7540. Both have a quiescent current less than 5uA. I added a diode with low voltage drop in series which keeps the capacitor at max 3.8V. The diode BAT20JFILM prevents discharging of the capacitor into the LDO at night. This diode was selected for it's low reverse current. Below 4V the LDO input closely follows the capacitor voltage. This is one of the secrets of the high efficiency, as long as the capacitor is charging there is little voltage drop. A 3.8V LIC must be kept within 2.5-3.8V range. Therefor during charging the solar panel voltage will be in the range 2.75-4.05V. So a 4V solar panel may work close to it's maximum power voltage. And due to the very low voltage drop of the LDO and the diode, the system efficiency is very high.

To benchmark efficiency, I compared the LDO-based circuits against an advanced energy harvesting chip E-peas AEM10941. This IC was configured to load the solar panel to 70% of open circuit voltage.

Left is the Holtek HT7540-7 in SOT89-3 package, middle is MCP1702T-4002 in SOT23-3 package and right is the AEM10941 in 28 pins LGA package.

I used two solar panels with same power but different voltage 4V/150mA 70x70mm and 5V/120mA 80x50mm.

First I measured the IV and PV curves of the solar panels in full sunlight outdoors

It showed the 4V-panel had it maximum power at 3.75V and the 5V-panel had maximum power at 4.75V

I connected a 250F 3.8V LIC capacitor and measured LIC charge current, and voltage drop between solar panel, the LDO, diode to the LIC.

I tested both indoors at 400 lux light level and in my window on a cloudy day 2500 to 5000 lux.

I tested...