Hello,
I just bought a DS3231 breakout (ZS-042), later I found a topic about it having a "circuit" that constantly charge the battery.
I've read somewhere that constantly charging a Li-ion battery, even if it's fully charged already, is safe and does not hurt the battery. Is this true?
Thanks
No. See batteryuniversity.com for a detailed discussion, page BU-409.
By the way, the charging circuit on those RTCs is an extremely bad design that can cause the DS3231 to malfunction. A number of posts on this forum and elsewhere describe some simple mods to remove all the junk, so that you can use a standard coin cell as intended. It will last for years.
guix:
I've read somewhere that constantly charging a Li-ion battery, even if it's fully charged already, is safe and does not hurt the battery. Is this true?
You've misunderstand what you read.
Lithium-ion can be held at a steady-state voltage, because the charge current will slow to practically nothing. The problem is that holding the cell at that steady-state voltage is not easy. Additionally, I would worry less about hurting the battery and more about what happens to lithium ion when it is misused. (Hint: they have a nasty habit of exploding.)
To properly charge a Lithium-Ion cell you need to have a good constant-current supply that is programed to ramp the voltage properly for a lithium-ion. Conditions for over-tempeature and over-voltage must be taken into account, other wise, the battery can (and will) explode.
Just applying a 3.3V line to a lithium-ion battery is NOT at all safe.
guix:
I just bought a DS3231 breakout (ZS-042), later I found a topic about it having a "circuit" that constantly charge the battery.
Did you also read reply #5 in that topic?
I found no schematic for the ZS-042 breakout board, but most likely the charging is done the same way as with the RTC Tiny DS1307 modules: The charging is not done with 5V from VCC, but with a voltage divider that divides the 5V by using resistors into a 4.2V/0.8V voltage.
This voltage divider will lead to extra power consumption when the Arduino is powered, and the voltage divider will provide a small current for loading the LIR accumulator, but only until a voltage of 4.2 volts is reached.
4.2 volts is the high voltage for a fully loaded LIR cell with a nominal 3.6V rating.
So the LIR is not "constantly charged", but it is only "charged until the cell voltage is 4.2 volts", which is the same as "fully loaded".
jurs:
I found no schematic for the ZS-042 breakout board, but most likely the charging is done the same way as with the RTC Tiny DS1307 modules: The charging is not done with 5V from VCC, but with a voltage divider that divides the 5V by using resistors into a 4.2V/0.8V voltage.So the LIR is not "constantly charged", but it is only "charged until the cell voltage is 4.2 volts", which is the same as "fully loaded".
I do have the schematic, and the board. The charging is not done that way. The battery is fed a constant current from Vcc, via a diode and resistor in series. So the LIR is constantly charged. I have removed the resistor from all of mine.
Thanks.
Here is the schematic:
Unless I'm missing something, I don't see anything that would stop the battery from being constantly charged (nor do I understand how a voltage divider could help). It is probably bad design as jremington said, but I don't think it could damage the DS3231.
Nonetheless I will remove that diode, but I was deceived by a previous RTC module I had, with a tiny battery that was soldered to it, which lasted only one year, then the RTC was loosing time because battery voltage was too low I guess ( < 2.3V according to DS3231's datasheet )...
So I still would like an integrated charge circuit for that LIR2032, that I will put in the ZS 042 module, so I don't have to remove it when it needs charging... What do you think if I add one of those cheap TP4056 breakout board (just search TP4056 on ebay), of course after replacing Rprog with a reasonable value ( 33 Kohm, right? )...would it be good/safe enough ?
Not sure why you would want to do that, since the timekeeping current of the DS3231 would exhaust a CR2032 battery in 28 years.
The battery is fed a constant current from Vcc, via a diode and resistor in series. So the LIR is constantly charged.
The resistor seems to be 200k or similar, so it's a kind of trickle charging (which is not recommended for Li-ion batteries like the LIR2032).
The voltage drop across a 1N4148 at 0.0x mA is not really high enough to keep the max. charging voltage at 4.2V, if Vcc is 5V or even slightly higher.
But it's not really lousy or catastrophic, neither...
However, a CR2032 without charging circuit is a better solution, I agree.
The schematic posted does not correspond to the module I have, which has a resistive divider between the battery and the Vbat input. It violates the impedance requirements for the oscillator circuit and affects battery backed timekeeping accuracy. I did not say it would damage the module.
I don't believe it will last decades, because, unlike DS1307 modules, the high current needed for the temperature conversion that is done every minutes, and alarms, also self discharge and probably other factors... I think 5 years is closer to reality 
Anyway I agree too I would prefer not to use a rechargeable battery, but I have other considerations for my project, which make me think otherwise. So do you think using a TP4056 (or do you know a more adequate charger IC?) will work nicely?
Thanks
guix:
I don't believe it will last decades, because, unlike DS1307 modules, the high current needed for the temperature conversion that is done every minutes, and alarms, also self discharge and probably other factors... I think 5 years is closer to reality
Notwithstanding your other requirements, I believe you are mistaken about the temperature conversion. From the datasheet, "When the serial interface is inactive, timekeeping current (IBATT), which includes the averaged temperature conversion current, IBATTC, is used".
IBATT is about 1uA. Also, a rechargeable battery has a shelf life, just as a non-rechargeable battery.
michael_x:
The resistor seems to be 200k or similar, so it's a kind of trickle charging (which is not recommended for Li-ion batteries like the LIR2032)..
It's a 200 ohm resistor.
