I am building a solar powered (hopefully) automated garden watering system. I have put together an Arduino Nano, Capacative soil Moisture sensor, WiFi module (to store data on a server in my house), and LCD.
I am hoping to power it via a 2.5w 5v solar panel which can charge an 18650 via a TP4056 at the same time.
I bought some solenoid valves, but realized they were 12v ~400ma devices, so I bought 2A step-up power converters to run the valve. The solenoid valves are here. I did the math on the step up and realized that I was going to be having to pull approximately 1.25A at the 3.3v that the 4056 outputs in order to power the solenoid valve.
So the question: Is that reasonable, or am I trying to draw too much load? Is there another alternative? I looked at 12v solar panels and charge controllers. My biggest hesitation is the extra cost. It adds quite a bit to the price of the project. (for my budget at least!)
You mean an ESP-01? Not a great combination, with the Nano being 5V and the ESP being 3.3V. I hope you don't plan to run the ESP from the Nano's 3.3V pin, it can't provide the necessary current. My advice would be to use a Wemos D1 mini instead.
Thanks Paul, I hadn't even thought about that. I assumed that the 5v pin would step up the power...though I suppose I should have verified that. I'll have to research which board I am going to use. Maybe an ESP2866 will be sufficient.
The TP4056 chip just does the charging and prevents overcharging. Some boards containing TP4056 you can buy on eBay/Ali also have an extra protection chip to prevent over-discharging. But be warned, not all boards have the protection chip. But I've not seen any with 3.3V regulator also.
In the above image, the boards in the foreground do not have over-discharge protection chips, but the boards in the background do. If your 18650 cell has built-in over discharge protection, it does not matter which board you use. But most 18650 batteries don't have built-in protection. Those that do have a tiny round PCB at the base of the battery and a metal strip running up the side of the battery that allows the protection circuit to monitor the battery's voltage.
I don't know what type of regulator chips that esp32 boards generally have. They may all be different and some may not be low drop-out, as the tutorial you linked to warns.
But wemos D1 mini has a good low drop-out regulator on board, similar to mcp1700. Not all esp8266 boards are the same. NodeMCU boards normally have an ams1117 regulator, which, although described as "low drop-out", it really isn't compared to mcp1700 or the wemos D1 mini's regulator.
But for your 12V boost converter, you would connect this direct to the protected output pins of the TP4056 board, not to the 3.3V regulator' output, so drawing 1.25A should not be a problem I think.
Even if you do not follow the scheme, I presented above you should use the diode. Even with a 5V cell. Which makes a problem, the voltage drop of the diode.
Look at the solar powered project in 3 sections.
Section one is the solar cell and the controller, section 2 is the controller and the battery, and section 3 from the battery to the load.
Current should only be allowed to flow in section 1 from the solar cells to the controller.
Current, section 2, between the controller and the battery is in both directions.
Current, section 3 is from the battery to the load.
ETA.
A LiPo will work for a short time but the conditions are quite rough for a LiPo. The TP4056 will cause the battery to fail in about 4 months. The TP4056 just does charge and discharge. It does not have a range of charged volts do not charge till volts drop to a level.
LiPo does not work very well when hot but they really do not work well when frozen. If the LiPo freezes, replace it.
I use a LiFePo4 so I do not have to replace LiPo's once every 3 or 4 months. In the end the LiFePo4 saves money over LiPo's.
Thank you for all the instruction! I really appreciate it. All of my previous projects have been indoors and just plugged into the wall. This solar and battery power is all new, especially when it comes to a multi-voltage environment.
May I suggest you consider the sleep modes of the MCU you are using and how will you implement power savings.
One thing I'd consider is all power removed from the 12V devices during deeps sleep of the CPU. This would mean your static sate of operations should not rely on components powered all the time. Like using normally closed solenoid valves.
How often do you want to have the soil moisture checked?
The battery could last up to 6 hours with a 1.25A load of continuous operation.
The MCU volts will be used to drive the solenoids, a relay could provide power to open the valves that can be controlled by the CPU. An opto-isolated relay module.
It's a 1A chip, yes, in that it can charge a battery at a maximum rate of 1A. That chip does not limit the discharge of the battery. Some modules containing TP4056 also have an extra chip such as DW01A which limits the discharge current, and protects against over-discharge, but not all do. Even those that do have the extra chip will not limit the discharge current to a low value like 1A, more like 3A, so should be ok for 1.25A.
I do know that after 3 or 4 months of continuous operations using a TP4056, I had to replace the batteries. The batteries refused to take/retain a charge.
Will you really be doing both at the same time? If so, the tutorial you linked has the load current flowing through the charger when the sun is out, which would provide a maximum of 1A. And since the TP4056 terminates charging when total current drops below 10% of the full charge rate, it may never terminate if the load current alone is greater than that. It's not good to continue to supply 4.2V to a lithium battery after it is already fully charged. Bad things can happen.
The solution to this is to either not do both at the same time, or have regular periods (ie - sleep) when the load current drops way down so the charger can terminate properly, or include a "load sharing" (aka "power path") circuit which will let the panels directly power the load when they are in full sun while independently charging the battery, so no load current ever flows through the charger.
Andreas Spiess has a video on using solar panels with the TP4056, but his circuit doesn't work. I posted a followup Github repo that uses an opamp as part of the load sharing circuit:
Hi,
Don't forget you will at very rare times get anywhere near 2.5W from your panel.
2W possibly during good days with the panel properly aimed at the sun, but most of the time when you do your energy calculations calculate for 1.5W and for 4 or 5 hours a day at that level.
PVs are wonderful things, but it takes a lot to get the best out of them energy wise.
Tom...
PS. If you have your panel, get it outside and do some load tests.
My challenge has been that my solar cells receive direct sun between the middle of June and the middle of July. I've had to learn how to squeeze every bit of battery charge power I can get. From Dec. to March, it takes 80 watts of solar to keep the batteries charged. In mid-March I remove the 50W cell, I am able to generate enough power on 30Watts till mid Dec. Typically, in Dec. the whole project gets covered by snow and runs on batteries 24/7 for weeks at a time.
Yea, @TomGeorge is correct the rated power ratings are under ideal conditions with brand new cells. A modern good quality solar cell should give out 85% of its rated capacity over its life. Thats the figure I use when calculating power availability of a solar cell.
