ESP8266 Solar Panel & 18650 Battery Charging Issue with TP4056

I have a mini station made that monitors soil moisture and weather conditions on an ESP8266 powered by an 18650 battery and a small 6W solar panel + charger on a TP4056.
Generally the system worked for 35 days without problems until it shut down due to the battery discharging.
The problem is that when the sun comes up and the voltage on the panel comes up, the charger does not charge the battery.
The meter shows 7.2V on the panel and a current of 0.04A. There is a red LED on the charger showing that the battery should be charging.
After inserting another battery (not fully charged) in today's nice sunshine, the current from the panel is 0.8A (the panel's rating is a maximum short circuit current of 0.92A) - the battery is charging with about 0.6A the rest goes to power the ESP and sensors.
It appears that when the battery is discharged, the TP4056 does not induce charging and does not load the panel. It doesn't matter if the power to the ESP is connected or we just have the battery itself plugged into the charger.
Anyone know if the TP4056 has it like this or if mine has gone flat? And if this type has it then what should I use instead?
What I want is for the system to start up as soon as there is enough sunshine to start charging and power the ESP.

Krawietz

Charging a largely discharged Li-Ion battery starts with a low current. This is probably less than the current drawn by the device. For this reason, the whole thing does not start. In this case, it is necessary to disconnect the power supply from the battery and wait for it to be at least partially charged.

TvWidget wrote:

Charging a largely discharged Li-Ion battery starts with a small current. Probably this is less than the current drawn by the device. For this reason, the whole thing does not start. In this case, it is necessary to disconnect the power supply from the battery and wait for it to be at least partially charged.

The voltage on the battery is 2.8V and it doesn't look like the TP4056 is trying to charge.... It doesn't matter if the ESP is plugged in or not.

Krawietz.

The TP4056 is not suitable for charging from a 'bare' photovoltaic panel. The solar charger should keep an eye on not taking so much current that the voltage drops below the maximum efficiency of the photo cell, and the TP4056 will take the current as much as it has set with the resistor and the panel won't work properly, the TP5056 will detect the lack of voltage and shut down charging, and so on and so forth.
And if it's a nice sun that can give that current more than the TP4056 requires then everything will work OK, for a while. Maybe it can be handled somehow with an additional Attiny working as a guardian of that PV voltage, but I guess it's simpler to use a dedicated module: "CN3065 Mini Solar Lithium Battery Charger Board Continuous Charge Current to 500mA Li-Po battery solar panel 2 pin JST connector".

Oh thanks for the hint. So will have to replace with something dedicated. This CN3065 doesn't really work for me, because the panel is 7.2V and the CN has a maximum terminal voltage of 6.5V. You'd have to put a diode in series with the panel to bring the voltage down, only that would probably introduce some loss.
All in all, the internet is full of solutions using TP4056 for solar charging...hence my experiment using them.


Krawietz

krawietz wrote:

The voltage on the battery is 2.8V and it doesn't look like TP4056 is trying to charge.... It does not matter if the ESP is plugged in or not.

.
When the voltage on the battery is below a certain limit (~2.9V) the charging current is significantly reduced. If you connect a discharged battery it will be charged but at a very low current. This is the recommended algorithm for the charger and presumably this TP4056 uses it too.

When the voltage on the battery is below a certain limit (~2.9V) the charging current is significantly reduced. If you connect a discharged battery it will be charged but at a very low current. This is the recommended algorithm for the charger and presumably this TP4056 uses it too.[/quote].

But then shouldn't the battery charge at least a little all day in the sun? And here generally nothing happens....
I have to check what happens if I plug in a 0.5A USB charger... instead of the panel...

Krawietz

krawietz wrote:

But in that case, shouldn't the battery charge a bit all day in the sun?

.
Not if there is an ESP8266 connected to the battery. It probably draws more current than the initial charge current. You mentioned a current of 0.04A.

TvWidget wrote:

krawietz wrote:

But in that case, shouldn't the battery charge a bit all day in the sun?

.
Not if there is an ESP8266 connected to the battery. It probably draws more current than the initial charge current. You mentioned a current of 0.04A.

But then, would a dedicated solar charger solve the problem and give a higher current to the power output or will the same action happen?

Tailor

krawietz wrote:

.
But then, will a dedicated solar charger solve the problem and feed more current into the power output or will it be the same action?

It is impossible to answer such a general question unequivocally.
What you refer to as a "solar charger" only ensures better use of the energy provided by the solar cell. What is important are the values of the charging current and the load current as a function of the voltage on the battery. You must ensure that in each case the former current is greater than the latter. You must also not exceed the values recommended for a particular cell. Sometimes the solution is to disconnect the load when the voltage is too low.
If you have a regulated power supply, measure how much current the weather station draws as a function of the supply voltage, e.g. from 2V to 4V.

As a solar charger, I am thinking of a circuit based on the CN3065 mentioned here, or rather CN3791 because it has a larger input voltage range.
The panel in good sunlight gives this 0.8A and the ESP draws 150 - 200mA.
So the question is whether there is a circuit that will draw enough current from the panel (if it is sunny, of course) to make the ESP work and the battery charge at the same time.
Well, unless these systems regulate the current in function of the battery voltage and will not give more when the battery is discharged.
Maybe I can add a step-up converter so that there is always this 5V at the ESP input, but from what I have experimented with it causes quite a large energy loss.
Alternatively, is there any ready-made protection system against discharging the battery with voltage regulation, I would set it for example at 3V and then the ESP would be disconnected and the battery could easily start charging.

Krawietz

It's not as if the solution to every electronic problem is to buy the right chip or module.
Take a look at the graph at the bottom of the first page of http://www.tp4056.com/d/tp4056.pdf and poke around to see what might be causing the lack of charging. With this load disconnection it is not such a simple matter. The voltage on a battery especially a discharged battery depends on the discharge/charge current. Care must be taken to ensure that the battery as a whole does not become excited. A further complicating factor in the analysis is the changing sunlight.

From the graph it seems that charging should start at around 100mA, and this is not the case even with the ESP disconnected.
Tomorrow I'll try a 5V 500mA power supply, I wonder if it will behave in the same way as with the solar panel.

Krawietz

ESP should not work like that, it should measure something, send it to a server, e.g. ThinkSPeak, and put it to sleep. And you use the Thinkspeak widget to monitor the parameters on your Android phone. Even a tractor battery can be discharged when it snows, a 120Ah battery like this can be discharged in 4 weeks, and actually faster, because it has less capacity in the cold. If you wake it up every 15 minutes and it measures in 10s, that's almost 100x less power consumption.
If the device has to be switched on all the time then it is better to have a sth->3.3V inverter. If the device is put to sleep it can be a linear HT7333, it draws 4uA + ESP with 14uA, it can sleep for 14 years on a 2500mA battery, or run for 16h while working with WIFI, from the ratio it comes out how much it will last in mixed mode, there are calculators for your phone that calculate after giving the current consumption in sleep, while working and after giving the duration of these modes.

Due to the nature of the application I'm using this with (when the ESP is not connected it doesn't even display the last data read) I'm keen for the device to work as much as possible when there is current. The problem is that the system doesn't want to "get up" when there is sufficient sunlight to charge the AKU and power the ESP.

Krawietz

Sometimes the limitations are only in our heads. Maybe it is cheaper to fire up a second ESP on the same network, but already hanging on a phone charger in a 230V contact, if the application does not know how other than to be connected permanently to the ESP. This ESP on battery power will go to sleep, when waking up it will send data to the server and to the second ESP, or just to the server, the second ESP will fetch data from Thinkspeak for itself in the same way as the Thinkspeak Widget on Android does.