Charging Multiple Lithium Batteries

[ronv]

Just gonna throw out an idea Here. Im thinking a Comparator circuit that checks the battery's voltage. Cells only need to be balanced if they are Over 4.2V, Correct? Under charging really isnt too much of a issue. So, If i have the comparator trigger a small load (say 100 mA) if the batteries voltage is over 4.3V and then turn off once the voltage reaches 4.2V. Im thinking this would achieve balance, and not risk damaging the cells. The only thing is, would I use a Comparator with Hysteresis?

That is what the circuit in post 17 does. The Tl431 is a voltage reference that can be set to turn on the shunt transistor to "steal" the current from the battery that is already at 4.2 volts. These work okay for low currents like you were talking about (100 - 200 ma. or so).

 

[ChrisP58]

I was under the assumption that balancing takes place after charging. I mean, while your charging at 1A, You would also be consuming 1A at the same time, so the battery wouldnt get charged, would it?

Yes, it is during charging. In this case, it happens right at the end of charge when one cell is full but the other(s) are not. It is to prevent one cell from overcharging while the rest finish.

The better way of passive balancing is during the entire charge time, but the circuitry to do that is more complex. You need to constantly compare the voltage of all cells and shunt some of the charge current past any cell(s) with the higher voltage.

 

[audioguru]

Balancing a multiple-cell lithium battery allows each cell to chrarge to a maximum of 4.20V separately because each cell reaches 4.20V at different times due to the difference in capacity.

If you try to charge two cells in series from a 8.4V source then the cell with the least capacity will reach 4.20V first then continue increasing its voltage until it catches on fire. A lithium fire is not nice.

 

[Overclocked]

That is what the circuit in post 17 does. The Tl431 is a voltage reference that can be set to turn on the shunt transistor to "steal" the current from the battery that is already at 4.2 volts. These work okay for low currents like you were talking about (100 - 200 ma. or so).

One for each battery, correct? I would probably need to bump up the transistor if I were to balance at 1 Amp, or substitute a FET. I was only doing low currents because I dont want to drain the battery with it sitting there, but then again the balancer will only be active when its above 4.2V. The Batteries are actually 4800mAH per Back (2 in parallel per 4.2V back-so basically 2P 2S). Is there any difference between the Discrete version and one using a Op-Amp in terms of function?

So I have many options here:

1) Discrete Like Post #17 and #21
2) Switch in between the Pos + neg of the two batteries being charged (each having their own charger)
3) Comparator Just dumping the load if the cell is above 4.2V
4) Dedicated IC (Small But Might be not solderable..I havent delved into really small parts yet)
5) Just make One huge Battery pack, with all the cells in Parallel. Avoids complicated Charging (I could Change my design for it-But I think its taking the "easy" way out)
6) Separate Buck Converter and Charger for Each Battery.

All of this has me wondering, how is this done on a commercial scale? I mean there has to be something Done on a commercial scale to balance cells thats cheap and effective, Unless they just design everything to run on single cell's. I mean, how is balancing done in a laptop?

ADD: Ok Linkage time. Im adding some Links since I know others might come here for the same reason. Looks like someone else had the same idea of using a comparator:

Comparator Balancer:
https://www.rcgroups.com/forums/showthread.php?t=157641

Discrete Balancer -looks like the schematic was updated a few times.
https://www.rcgroups.com/forums/showthread.php?t=270580

 

[strantor]

Balancing a multiple-cell lithium battery allows each cell to chrarge to a maximum of 4.20V separately because each cell reaches 4.20V at different times due to the difference in capacity.

If you try to charge two cells in series from a 8.4V source then the cell with the least capacity will reach 4.20V first then continue increasing its voltage until it catches on fire. A lithium fire is not nice.

Could you charge each battery separately, while keeping them in series, with separate isolated 4.2V supplies?

 

[audioguru]

Could you charge each battery separately, while keeping them in series, with separate isolated 4.2V supplies?

Yes you can.

 

[wall]

I'm going to build a battery powered LM2596 CC/CV step down converter with two 18650s in series and was also wondering how could I charge the batteries with all the components inside a housing.
I'm no expert in electronics, but I had thought about using individual charging circuits for each cell, but didn't know if it was possible with both cells still in series.
Could I use this **broken link removed** on each cell, with both cells permanently in series, and connected to this **broken link removed** with a switch?
I'm trying to keep things simple and cheap so only one AC adapter would be ideal, and the components would have to be permanently assembled inside a housing, is there anything that wouldn't work with the setup I just described?
Also, could I draw power from the batteries while they're charging?
I've attached a very simple schematic of what I had in mind.

 

[simonbramble]

what about something like this:
https://cds.linear.com/docs/Datasheet/6801fb.pdf

 

[ronv]

One for each battery, correct?

Yes
[quote}I would probably need to bump up the transistor if I were to balance at 1 Amp, or substitute a FET. I was only doing low currents because I dont want to drain the battery with it sitting there, but then again the balancer will only be active when its above 4.2V.

It all gets more complicated as the current goes uo. You need to match the constant current source to the shunt resistor. So both will disapate more power. No worry about battery drain. Your right they are only on over 4.2 volts.

The Batteries are actually 4800mAH per Back (2 in parallel per 4.2V back-so basically 2P 2S). Is there any difference between the Discrete version and one using a Op-Amp in terms of function?

I think they all use the same principal. Just that some assume the batteries are not far out of balance so don't steal all the current from the one that is fully charged.

Others don't balance at all.

 

[MrAl]

Hi,

I joined this thread late so i may have missed some comments.

Charging two cells that are in series requires either two isolated power supplies, or two power sources that can act in series, or two shunt circuits to shunt the current around a cell once it is charged (with both being charged from a higher power supply voltage).

Balancing isnt that important unless the cells are to be connected in parallel during actual use. In series as long as either cell does not get discharged too low it is ok, but that does mean some sort of protection circuit.

 

[dougy83]

Here's a quick sketch of an idea for a balancer in the charge circuit.

As you just want to _balance_ the cell voltages, there's no need for any zener references, just a resistor divider to get the target voltage between the two Li cells. The opamp attempts to correct the mismatch during the charge. The two transistors stop the opamp draining the cells when the charger is disconnected.

The balancing current is going to be small (but so was the above TI solution, so maybe it's not an issue); limited by the opamp output current. This can be boosted using e.g. an emitter-follower transistor configuration on the opamp output, if required.

 

[ronv]

I kind of like the comparator one. Fewer parts.
What are you thinking about the charger part of it?
Here is a little write up.

https://sites.google.com/site/tjinguytech/charging-how-tos/how-lipo-chargers-work

 

[Overclocked]

It all gets more complicated as the current goes uo. You need to match the constant current source to the shunt resistor. So both will disapate more power. No worry about battery drain. Your right they are only on over 4.2 volts.



I think they all use the same principal. Just that some assume the batteries are not far out of balance so don't steal all the current from the one that is fully charged.

Others don't balance at all.

I wont be getting any higher than 1A, and now that I think about it, I think I'll charge at 1A. Id have to leverage a power supply for this somehow. I was originally going to use a power brick, but I dont think I have anything that can put out that much current and thats higher than 8V. I originally planned to charge at 500mA But They would take too long (9.6 Hrs).

Now the thing is, When the cells are approaching Full Charge (4.2v each or 8.4V total), The charger Tapers the current off to something like 0.1*C . Wouldn’t it be Better to Have the balance drain at something a little more than 0.1*C; rather than at 1C, say at 0.5*C? What Im getting at is, Even if I charge at 1C, the cells would have to be Drastically out of balance with each other (say ones at 4.2 and the others at 3.8V) in order to balance at 1C, so couldn’t I get away with the Lower Balance current?

 

[ronv]

Yes I think you can. Probably .1C would be enough unless something is "bad". In that case a 1 watt resistor and small FET would do it.
Ebay has some cheap supplies.
**broken link removed**

Or charger balancer:

**broken link removed**

 

[Overclocked]

I kind of like the comparator one. Fewer parts.
What are you thinking about the charger part of it?
Here is a little write up.

https://sites.google.com/site/tjinguytech/charging-how-tos/how-lipo-chargers-work

I was going to use the BQ2057 Charger. Has pins for temp and uses a external PFET.

https://www.ti.com/product/bq2057

Here's a quick sketch of an idea for a balancer in the charge circuit.

As you just want to _balance_ the cell voltages, there's no need for any zener references, just a resistor divider to get the target voltage between the two Li cells. The opamp attempts to correct the mismatch during the charge. The two transistors stop the opamp draining the cells when the charger is disconnected.

The balancing current is going to be small (but so was the above TI solution, so maybe it's not an issue); limited by the opamp output current. This can be boosted using e.g. an emitter-follower transistor configuration on the opamp output, if required.

You know what, Using the Output (8.4v) is a awesome idea for a reference voltage. Hmm but it seems you have the emitters and collectors on the transistors opposite of each other (Unless this was on purpose). Would it matter if the CS resistor was High-Side or Low side? I was going to go with High-Side. Hmm, this basically looks like a emitter-follower.

How does it actually Balance the batteries though?

 

[MrAl]

Hi again,


Another simple method would be to use a charging circuit set to 1A with double the voltage (8.4v) and use a shunt regulator across the bottom cell.
That would mean a circuit that can charge a Li-ion cell but puts out 8.4v instead of 4.2v, and a shunt regulator. The shunt regulator naturally conducts current *around* the load when the voltage reaches the set point (4.2v).

 

[dougy83]

You know what, Using the Output (8.4v) is a awesome idea for a reference voltage. Hmm but it seems you have the emitters and collectors on the transistors opposite of each other (Unless this was on purpose). Would it matter if the CS resistor was High-Side or Low side? I was going to go with High-Side. Hmm, this basically looks like a emitter-follower.

The 2 transistors are rigged that way on purpose and all they're doing is acting as a switch (ON when the charger is connected). You could also get away with just a single MOSFET.

Would it matter if the CS resistor was High-Side or Low side?

I just put it there to show how the reference is connected; the CS can be anywhere... just the divider goes across the two cells.

How does it actually Balance the batteries though?

The opamp pulls the voltage at the cell's junction towards half the total cells' voltage (i.e. where both cell voltages will be equal). This might only be 40mA or so. When the cells are balanced, the opamp won't be pulling or pushing.

Another simple method would be to use a charging circuit set to 1A with double the voltage (8.4v) and use a shunt regulator across the bottom cell.
That would mean a circuit that can charge a Li-ion cell but puts out 8.4v instead of 4.2v, and a shunt regulator. The shunt regulator naturally conducts current *around* the load when the voltage reaches the set point (4.2v).

But what if the top cell has more juice than the bottom cell? Wouldn't you need a shunt on both?

 

[MrAl]

Hi dougy,

Yes a shunt regulator on the top cell is also needed. That means the power source itself doesnt have to be regulated.
There's also another drawback though and that is that the shunt regulator would be shunting the full 1 amp once the cell(s) is charged, thus wasting power for the entire time it takes to charge the other cell.
I guess we also have to keep in mind that most of these chargers shut the charge off completely once the current gets down to maybe 1/20 of the cell capacity.

I have a feeling there is another simple way to do this too without having to use isolated power supplies. I say this because they dont really have to be isolated if the positive of one cell connects to the negative of the other cell. I think this deserves more thought.

 

[dougy83]

I have a feeling there is another simple way to do this too without having to use isolated power supplies.

I think using the charger IC the OP has chosen (BQ2057) and something like the balancer I posted should be a pretty simple solution.

 

[MrAl]

Hi,

I dont think the BQ2057 chip is adequate for his application, i would say maybe the BQ2057W would be the choice here.

Are you talking about the diagram in post #31 ?

I wouldnt mind looking into a non linear charge method for two cells in series either.