Charging Multiple Lithium Batteries

[dougy83]

>> I dont think the BQ2057 chip is adequate for his application, i would say maybe the BQ2057W would be the choice here.
Quite right, the 2-cell 8.4V is the one we're after.

>> Are you talking about the diagram in post #31 ?
Something like that, yes.

>> I wouldnt mind looking into a non linear charge method for two cells in series either.
Maybe an SMPS in place of the pass transistor with the BQ2057W controlling the feedback... or something else completely...

 

[MrAl]

Hi again,


Yes something with a switcher or two in it.

In your diagram i dont see the current path for the upper cell. I could be missing something so maybe you could explain the circuit a little here.

 

[ronv]

Budget balancer

Here is a real simple charger balancer.
Set the 317 for a nats over 8.4 volts.
Adjust the TL431's to turn on at 4.2 volts.
The transistor limits current to about 2/3 amp.
The TL431's take about 80 ma.
It never turns off, but there is very very small current at the end of charge.

 

[dougy83]

In your diagram i dont see the current path for the upper cell. I could be missing something so maybe you could explain the circuit a little here.

The balancer in the cct above is just an opamp acting as a [unity-gain] buffer with it's output voltage at exactly half the voltage of the 2 series cells... well, it would be if the opamp had a high current output. The opamp has a high output impedance, so it's like connecting the ideal voltage to the centre-point of the 2 cells through a resistor, which will slowly balance it.

There's also the 2 NPN transistors there on the output; they're just acting as a switch to isolate the opamp from the cells if the circuit is disconnected from the charging source (otherwise the bottom cell would discharge through the opamp).

The cell-balancing current path for either cell is through the output of the opamp -- it can both sink and source current. The 2 NPNs are configured in the way shown so that the conduction through the transistors can flow either way easily.

 

[MrAl]

Hi again,


dougy:
Oh ok well you had shown an LM358 in your diagram so that doesnt put out enough current. Yes a higher current buffer op amp might work though.

ronv:
I dont see those TL431 chips as being able to pass 1 amp current (as the usual charge rate) or even 0.6 amps. So if say the bottom cell charges up first it still has to take almost the full current which would charge it even more. Take another look. I think the 431's would have to be able to pass much more current in order to shunt each cell if required. That's why a transistor is usually used with the 431.

 

[Overclocked]

Here is a real simple charger balancer.
Set the 317 for a nats over 8.4 volts.
Adjust the TL431's to turn on at 4.2 volts.
The transistor limits current to about 2/3 amp.
The TL431's take about 80 ma.
It never turns off, but there is very very small current at the end of charge.

That reminds me of the LM317 Li-Ion Charger that I saw. Ive actually used this charger to Charge LiFePO4 batteries, as they need to be charged the same way but are less prone to catching fire . The circuit Works really well, and the current drops to something really low (like 5mA or less) once they get close to 4.2V/cell.

Hmmmm Isnt the LM317 a floating regulator? If so..This gives me an idea. Lets say I go with my original idea, But, since the current drops to something low once the cells are done, I could put a balancer in parallel with each cell. Even though the charger doesnt shut off once 4.2/Cell is reached, having a balancer in parallel would essentially have the same effect.

ADD: derp. I meant use two of those LM317 Chargers in Series with each other, with this setup I wouldnt need a balancer.. Although, One Charger set at 8.4V Would be easier to design around.

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.

Yea your right, BQ2057W, 8.4v . I meant to specify as they also have a 4.2V version.

 

[dougy83]

Oh ok well you had shown an LM358 in your diagram so that doesnt put out enough current. Yes a higher current buffer op amp might work though.

The LM358 has a greater current capability than the bq2920x balancer IC mentioned above.

 

[ronv]

ronv:
I dont see those TL431 chips as being able to pass 1 amp current (as the usual charge rate) or even 0.6 amps. So if say the bottom cell charges up first it still has to take almost the full current which would charge it even more. Take another look. I think the 431's would have to be able to pass much more current in order to shunt each cell if required. That's why a transistor is usually used with the 431.

No they can't do an amp. They are limited to about 80ma. by the series resistor which they can do. Since the charger is in constant voltage mode at the end of charge (where the voltage gets to 4.2 volts) the charge current is very low so not much is needed if the cells aren't way out of balance.

Overclocked, The thing you can't get away from is that with 2 regulators they still share the same ground and the batteries don't.

 

[MrAl]

Hi,

Yeah but you can not bank on both currents reaching the same current (or some low value) at the same time. That's why a balancer really has to be made so that it can shunt the entire max charge current. If that is 600ma then each shunt balancer has to be able to handle at least 600ma. 80ma isnt nearly enough. That's where the transistors usually come into play.

 

[ronv]

Most of the ones I have seen only use 10 to 20% of the charge current.

https://www.electro-tech-online.com/custompdfs/2012/07/chargery20320b20manual20v10.pdf

 

[Overclocked]

But it makes sense to balance at a Low Rate, Its not like Im going to throw two vastly out of proportionally charged batteries into this thing. Im actually going to Precharge them both to a full charge using a single Charger. Once both cells are at 4.2V, THEN I'll plug them into the final circuit. Im under the assumption that there will never be a vast difference in voltage, However, Assumptions are a Bad thing. So The Jury is still out for deliberation, Balance at 1C or Some Fractional Number of C?

My engineering instinct tells me to go with a 1C Balance, Just In case Some where along the lines one of the cell's do malfunction, But the BQ2057 has a temp sensor which I will be using, so there shouldnt be any issues other than ageing. Since the end Application will be a LED flashlight, Its also possible the Boost Converter can malfunction, and short out (On the input side of things that is). A Fuse should be able to handle that sort of malfunction, shouldnt it?

Add (off topic warning): I seem to need a bit of help calculating the resistors needed for the temp sensing portion of the BQ2057 Charger (page 12) . Li-Ions have a max temp of around 45C when charging (well should be below that). I dont need a lower temp as its going to be inside. Heres the datasheet for said thermistor. Its rated 10k @25C, Part Number NTCLE100E3103JT2. Itsa NTC thermistor.

Now the Formula For the resistors for temp monitoring need Two resistances from the Thermistor, Rth and Rtc (hot and cold). So luckily theres a table of values. Im going to be using 45C and 0C (Even though I dont need the lower temp). For those looking, the values are on page 10. For 45C its 4372, and for 0C its 32,554 Ohms. Now my Issue here is, for the denominator on the formula given by the BQ2057 Data sheet, Are negative, thus giving negative resistances.

But heres my answers for those who want to plug and chug: Rt1 = -8.41k and Rt2 = -3.24k. Do I just drop the negative?

 

[Overclocked]

Now I wanted to update this thread. Ive successfully completed my balancer circuit, however, Im a little concerned to as what the trigger voltage should be. I dont have many resistors on hand to make the proper divider, but it works. Ive been experimenting with different values, and Ive gotten it to trigger on at 4v (which is too low), 4.10V and 4.25V

Now, personally, I think 4.10V is a little too low (ideally I would like it to be 4.15v), Unless Im wrong about this. I think 4.10v is about a 90% charge But it also extends the life of all li-poly batteries by charging at that voltage (instead of 4.20). As for 4.25V, Well, Im not sure if lithiums can handle a 50mV Overcharge or not, since they are being charged at a low current during the last part of the cycle. At any rate, it would probably quickly go down once the balancer kicks in.

So which voltage do I shoot for here, 4.10 or 4.25? I dont have the proper resistors for it to trigger at 4.15 or 4.2.

I'll also be releasing the schematics and boards in the project section, I know many people could probably use something like this (as soon as the final revision is done that is).

 

[audioguru]

I don't know the accuracy of your resistors. The "4.1V" might actually be 5% low or high (3.895V which is too low or 4.305V which is too high. Aiming for 4.25V is too high and the accuracy might make it much too high.

It is simple to series or parallel two resistors to make your own value.

 

[Overclocked]

I don't know the accuracy of your resistors. The "4.1V" might actually be 5% low or high (3.895V which is too low or 4.305V which is too high. Aiming for 4.25V is too high and the accuracy might make it much too high.

It is simple to series or parallel two resistors to make your own value.

I always use 1% SMD resistors. Im also using a 2.048 voltage reference, That actually reads 2.039v from my meter. Reference is 0.5% Accurate.

add: Ive left almost no room for a series resistor...But I'll have to see what I can do. I may have to "triangle" them upwards.

 

[ronv]

Usually those circuits are a voltage divider so it might be easier to "tune it in" with a parallel resistor, then you could just "piggy back" it on top and no one would know except you and I.