Oznog
Active Member
I'm hoping to build a fairly large lithium pack for a project- 14ah, 8x series cells=29.6v.
Lithium requires very strict limits on charging voltage of each cell. 4.2V is good, 4.15V is slightly less capacity yet longer life. There are other cases where it'd be wise to charge to another voltage altogether, if the pack's just been run down but will not be used for some time then it should only be charged to 40% not 100%. That can usually be accomplished by using a different charging voltage.
We also need the cell's regs to talk with the power source. For example, we may be using a 10A charger and then one cell hits 4.2V and the cell's reg bypasses the charging current around the cell to prevent overcharge. However, a reg couldn't bypass 10A @ 4.2V, that's 42W of heat, the charger must throttle way back once the first cell begins bypassing.
This is actually more difficult than I originally thought, quite a bit of accuracy is required and being able to set cell voltages independently is a trick. I'm kinda stumped as to the most practical way to do this.
I'd thought about instrumentation amps to have one central ref voltage that could rereference that voltage to each cell so an op amp could drive a bypass transistor off it, but there's catches to that, most instrumentation amps don't deal with inputs with that high of a common mode voltage. Not many are rated for 40V supplies in the first place.
Hmm, kinda tricky here. I'd like to manage this expensive batt properly and I'm just not seeing how yet. Any good ideas I'm missing?
I'd been wondering about the old "Flying cap" where we use a double-ended transmission gate to charge a cap off a single reference then this transmisison gate disconnects from there and reconnects it to a different ground reference for another cell. Sounds pretty dicey though, relying on a cap's charge and a transmission gate's impedance.
I'm kinda leaning towards a PIC on each cell with an analog voltage reference and an isolated bus to talk to a central controller. A problem I see is that if I use PWM and an RC filter to drive a bypass transistor it gets kinda dicey if there's noise in the power supply. The ADC can only sample so fast and the PWM has a limited speed and resolution so if the main power supply has some switching noise it may have trouble adjusting the transistor properly to work with it.
Lithium requires very strict limits on charging voltage of each cell. 4.2V is good, 4.15V is slightly less capacity yet longer life. There are other cases where it'd be wise to charge to another voltage altogether, if the pack's just been run down but will not be used for some time then it should only be charged to 40% not 100%. That can usually be accomplished by using a different charging voltage.
We also need the cell's regs to talk with the power source. For example, we may be using a 10A charger and then one cell hits 4.2V and the cell's reg bypasses the charging current around the cell to prevent overcharge. However, a reg couldn't bypass 10A @ 4.2V, that's 42W of heat, the charger must throttle way back once the first cell begins bypassing.
This is actually more difficult than I originally thought, quite a bit of accuracy is required and being able to set cell voltages independently is a trick. I'm kinda stumped as to the most practical way to do this.
I'd thought about instrumentation amps to have one central ref voltage that could rereference that voltage to each cell so an op amp could drive a bypass transistor off it, but there's catches to that, most instrumentation amps don't deal with inputs with that high of a common mode voltage. Not many are rated for 40V supplies in the first place.
Hmm, kinda tricky here. I'd like to manage this expensive batt properly and I'm just not seeing how yet. Any good ideas I'm missing?
I'd been wondering about the old "Flying cap" where we use a double-ended transmission gate to charge a cap off a single reference then this transmisison gate disconnects from there and reconnects it to a different ground reference for another cell. Sounds pretty dicey though, relying on a cap's charge and a transmission gate's impedance.
I'm kinda leaning towards a PIC on each cell with an analog voltage reference and an isolated bus to talk to a central controller. A problem I see is that if I use PWM and an RC filter to drive a bypass transistor it gets kinda dicey if there's noise in the power supply. The ADC can only sample so fast and the PWM has a limited speed and resolution so if the main power supply has some switching noise it may have trouble adjusting the transistor properly to work with it.
Last edited:
