Today I will show you a project with several
aspects. Even if you do not need to charge batteries with voltages of 35 volts and above
using a solar panel, you will find learnings for your projects. Or you enjoy that not only you
encounter some difficulties in your projects.
Grüezi YouTubers. Here is the guy with the Swiss
accent. With a new episode and fresh ideas around sensors and microcontrollers. Remember: If you
subscribe, you will always sit in the first row.
My wife's sister asked me if I could build
her a solar charger for her bicycle batteries. Without thinking and without having a concept in
mind, I said: Yes, I can. This should not be too complicated. Right? Back then, I did not know
that this project would take more than a year.
First, I made a list of tasks:
1. I had to find the appropriate plug for the battery
2. I had to hack the battery interface. Because it did not charge when I
only applied a voltage to the two thick pins
3. I had to find a solar panel strong enough
to charge the 635Wh battery in less than a day
4. I had to boost the voltage of the solar panel
to the level needed to charge the 36V battery
The plug is from HiGo, and after a lot
of searching, I found their distributor in Belgium. They sent me the “minimum” number
of 5 connectors, including short cables. I do not want to tell you how much this costed,
including shipping. After all, it is for my wife's sister and, therefore, essential.
Point one was solved in less than a month.
After doing some measurements, I found that the
battery started charging when I applied 5 volts to this small control pin and that this pin is
connected to the green wire. And the maximum voltage of the original charger is 42 volts.
Cool! Point two is solved. This was fast.
For the solar panel, I looked at these flexible
and portable panels. They are ok, but their price/performance is not good. So I decided
on a standard panel with an aluminum frame. The panel will be stationary, and its weight is
an advantage, as we will later see. So I went for a 100W peak panel of eBay. It should be able to
charge a not fully flat battery in less than one day, even if there is not always sun. 635Wh/100W
= 6-7 hours of solid sun. Point 3 is solved, too.
Then I came to point 4. How would you solve this
problem? First, I thought: Easy, I use a cheap boost converter and boost the roughly 20 volts to
the 42 volts needed by the battery. As you see, this works fine on the bench. I even added this
lovely volt and ampere meter. After adding a 5V regulator to create the voltage required for the
control pin, I was ready for a test in the sun. I connected this converter to the input because
its maximum input voltage is only 24 volts, not the 42V of the converter's output.
Unfortunately, it did not work at all if connected to the solar panel. Do you know why?
Because the battery has a low internal resistance, it acts like a short. And what does the boost
converter do if its output is nearly shorted? It tries to draw the current from the solar
panel, of course. And the solar panel? Unfortunately, solar panels have such a curve.
If the current is reasonable, they keep their voltage. But if it is a bit higher, the voltage
drops rapidly till the booster switches off. Then, the panel's voltage increases, and the whole
thing starts over. So, what to do? I exchanged the boost converter with one where I could limit the
output current. Excellent! This works if we limit the current to a level below this bend. But what
happens if the solar radiation is a bit smaller? Then, the curve is reduced, and we are in the
wrong spot again till we adjust the potentiometer to a new level. What we need is an MPP tracker.
Because the project already was quite expensive, I tried to go cheaply. And because I read some good
reviews, I ordered this Solar charge controller. It sometimes is sold as MPPT but usually as a
PWM controller. At least it has “MCU control”. I hoped it would produce “optimal” power at the
output for my boost converter. When it arrived, I discovered that it does not work without a battery
to charge. Only connecting the boost converter did not work. So I added the LiFePo4 battery I use
for my transmitters. Now it worked. The controller charged the battery, and the boost converter
charged the bicycle battery. Exactly as planned!
I even was able to get good power. Do you see the
next problem? Yes, you are right. This battery is much smaller than the bicycle battery. When there
is not enough sun, the small battery discharges fast and switches off. Not exactly what I need.
Also, this controller is not very good. It seems it only “chops” the current from the panel to
the battery. This is why they call it a “PMW” or “pulse width modulated” controller. The effect:
The panel's voltage is way below its maximum power point, and its power is only around 50W. I am
moving in the wrong direction! Only 50% of the expected power, and because of the needed battery,
the price of this setup is relatively high. So I need a better solution. But how
does it look like? And where to get it?
Before starting KiCAD to create my own design,
I again searched AliExpress. After some time, I found this device: An MPPT controller
with a boost, not a buck converter. These are pretty rare. After all my setbacks, I
was very skeptical. Nevertheless, I had a promise out, and my wife asked from time to time about
the project's status. Of course, she did not want to know the details, just a date when it would be
ready. So I hit the “order” button. Because of the new “15 days delivery time”, the controller came
fast. And, because spring is here, also some sun is available for testing. But one problem has to
be solved: I need 5V for the small pin. As before, I wanted to use the buck converter connected
to the solar panel. But other than before, the input and output ground of the boost converter
is not connected. On the PCB of the 5V converter, however, ground on the input and ground on the
output is connected. So my small converter would shorten the two grounds and probably draw
a high current. Not exactly what I want. As mentioned before, the maximum input
voltage of this 5V converter is only 24v. So I cannot solve the problem by connecting it
to the output. So, what to do? Again, not easy. I needed some insulation between the grounds.
I remember buying very strange power supplies: 5V to 5V DC. I no longer remember why I bought
them, but fortunately, Aliexpress keeps all my orders in its database, and I could verify that I
should have a few of them somewhere in my lab.
After searching a bit, I found them in one of my
many boxes. I connect its input to the 5v of the converter and the output to the control pin
for the battery. Now I am sure the battery is protected, and the converter is not shortened.
After adding some heat-shrink tubes, everything is ready for a test. I connect
the solar panel to the converter, and really, it shows around 20 volts input voltage. The
display does not flicker in reality, BTW. It is because of an interaction with the camera.
I can adjust the output voltage to the needed 42 volts. We are on the right track! But
what happens now? The output voltage is 0V. Only if I press the buttons, it is 42
volts again. But only for a few seconds.
Not what I expected. It seems I got a
defective device, and all my hopes are gone! This is the hard life of an engineer. And the
even harder life of a husband who must tell his wife that she has to tell her sister that the
man she married is not as helpful as expected.
But before I give up and file a dispute on
AliExpress, I connect the bicycle battery. Look what happened! The voltage is around 40
volts, as expected from a half-empty battery. And even better, currents are flowing.
So the battery is charged. Not too much, but it is early in the morning. So maybe
my reputation is saved. And at least in low sunlight, the converter behaves much
better than all my former designs.
So let’s wait till noon.
Now we have nearly full sun. Look at that. The voltage is almost 20 volts, and more than 5A
flows into the charger. Nearly 100 watts. And it is only April. Cool! Like that, the battery will
be charged till the evening. Exactly as promised.
And really, after a few hours, the battery
switched off charging. Is it really full? Let’s check with the bike. Yes, it reached
nearly 100%. The missing few percents are not the responsibility of my charger; the battery does
it also when charged with the original charger.
After all the setbacks, I am happy with this
converter. It does what I expected. It varies the input voltage all the time to find the
sweet spot for maximum power, as you see here. With minimal sun, it regularly starts at around 15
volts and searches for the best point. The output voltage is precisely what the battery needs,
but there is hardly any current flowing because the panel cannot deliver. This is proof that it
is also stable at the lowest radiation levels.
But I must add two things to the to-do list:
The controller must be protected against rain. And more important, against thieves. Frequent
viewers of this channel know precisely why. But this is work for the partner of the
sister of my wife. He is a mechanic and much better suited for such problems. And he
probably also wants to prove his usability…
So finally, this story has a happy ending. My wife is happy, and I can
continue with my next project.
This was all for today. As always, you find
all the relevant links in the description.
I hope this video was useful or at
least interesting for you. If true, please consider supporting the channel to
secure its future existence. Thank you! Bye
