My plan is a TTN Node, or even a MySensors weather station, or other Arduino based projects to operate completely autonomously as a self-sufficient.
This is probably the most simple and cheapest a solar cell and a LiPo as a battery for dark times.
Arduino power supply through solar cell for many years
Choice of solar cell
The choice is not so easy – so I thought. However, there are not so many possibilities.
The load module I use in all my projects is a TP4056 with protection. This chip can handle 4-7V input.
Consequently, the solar cell should produce a maximum of 7V. Because most solar cells have 5 or 6 volts, I decided to use 6V, as it certainly will never deliver more than 7V.
Why not 5V?
Solar cells deliver their maximum voltage even at maximum light. Say if it is not direct sunlight, it sometimes has only fractions of it. By now we have 6V, have 2V buffer until the charger stops working.
And how much watts?
The watt results from the size. And the size results from what I want 😉
I just tried and ordered different. I am very satisfied with a 6V 0.6W solar cell and I will discuss it further. The dimensions are 90mm by 55mm. As always, you will find links to the products at the end!
As we learned in the last post, deep sleep, or the rare sending of values, is not the problem for us. Self-discharge is the big problem. LiPo’s discharge at 5% per month. Consequently, we have to charge more than 5% per month!
6V and 0.6W are 0.1A, so 100mA. I know, it’s a dream. The important thing is that I have not measured the output of the solar cell, but what finally comes out of the charger. So here is already a lot of loss. Voltage conversion, control, etc.
Our 3.200mAh LiPo we always use (NRC18650) discharges in the first month by 3.200mA * 0.05 = 160mA.
Let’s say we have one week of sun a month. So 7 days. If we calculate with an average of 8 hours, we come to 56 hours sun.
If we have to charge 160mA in 56 hours, it means that the solar cell has to deliver 2.8mA. That should not be a problem .. Or is it? Instead of speculating I measured.
In order to measure reasonably, I built myself a small construction.
For this I have taken the solar cell above, a diode to prevent backflow and the charging module TP4065. In addition, I have installed a step-up converter, which converts to 5V. So I can simply discharge the battery through the USB socket. In order for the set-up converter not permanently consumes power, there is a switch in between.
Between plus of the charger to the battery, i measured.
By the way, it’s full of sunshine in April, my weather station measures 10,000lx.
Interesting is the angle. If the solar cell is always kept in the sun, it is much more efficient. In my case, the solar cell is always installed flat on the top, so it usually has sun. Unfortunately never perfect, but more often.
Here are the two measurements. Values in Apere:
79mA direct, 37mA flat. In about half. 🙁
If only one cell of the solar cell is covered, the voltage drops drastically, causing the charger does not get sufficient voltage. Therefore, no charging (look at the little shadow):
An improvement I have not told to you: The charger has an LED that lights up immediately when power is applied to the input. I have unsoldered these. As we have learned, these SMD LEDs consume about 1mA. That does not have to be:
The solar cell charges the battery with about 40mA. According to our estimates of 56 hours per month we come to 2,240mA. That’s +70% per month. However, our TTN nodes consume so little that only self-discharge becomes a problem. This is 5%. When fully charged, that means 160mA. At 40mA load it means that 4 hours of sun per month are enough to counteract the self-discharge. Excellent!
To fully charge the battery we need 80 hours of sun.
Since our node would last for 2 years without charging, 40 hours of sun per year would be enough to never get empty.
If you order through any of the following AliExpress links, I get about 8% commission 😉