Top Balancing "How to"

[Tomthumb62]

Previously this is the voltage I charged to and then continued to sink in 6A (into 608Ah pack) for ~8 hours with only a 0.01V change in the entire pack (~1mV per individual cell)

That’s going to take a LOOOOONG time on a pack of that size.

 

[alexcl]

most cells ship around 30%, so you need to supply 70% of a 608Ah pack

0.7 * 608Ah / 6A = 71 hours.

Yep for the full charge.

This is what I saw yesterday. These screenshots are ~6.5 hours apart but I only started screenshotting after a couple hours of the pack voltage flitting between 26.94-26.95V. I actually forgot I had dialed up the current to ~7A and finally 10A (from 6A) over the course of roughly 8-9 hours.

The voltages didn't seem to change - as mentioned, the pack voltage oscillated between 26.94 and 26.95V. Cell voltages were relatively consistent over time too.

I did have a poor terminal connection on cell 6. Maybe that affected the charging?

 

[sunshine_eggo]

Yep for the full charge.

This is what I saw yesterday. These screenshots are ~6.5 hours apart but I only started screenshotting after a couple hours of the pack voltage flitting between 26.94-26.95V. I actually forgot I had dialed up the current to ~7A and finally 10A (from 6A) over the course of roughly 8-9 hours.

Not sure why you're not running at 10A all the time. That would cut the charge time down to about 42 hours.

The voltages didn't seem to change - as mentioned, the pack voltage oscillated between 26.94 and 26.95V. Cell voltages were relatively consistent over time too.

LFP is famous for having a very consistent voltage throughout most of its range. This slow voltage rise during charging phenomenon has been documented on this site literally hundreds of times.

I did have a poor terminal connection on cell 6. Maybe that affected the charging?

No. All cells in series MUST pass the same current. It might contribute to your voltage drop between supply and battery terminals.

If the BMS registers a current, it's accepting charge. At the very low rate you're charging, watching paint dry will be like an orgy of joy compared to watching the voltage rise @ 10A on a 608Ah pack.

 

[alexcl]

Not sure why you're not running at 10A all the time. That would cut the charge time down to about 42 hours.

It's my first time using a benchtop PSU and it was offgassing some pretty noxious stuff at the beginning. I'm guessing rosin heating up? At any rate it spooked me a bit so I've dialed down the current several times.

LFP is famous for having a very consistent voltage throughout most of its range. This slow voltage rise during charging phenomenon has been documented on this site literally hundreds of times.

Yeah I've seen it - maybe I'm misunderstanding something though. I've been looking at charts like this one for lifepo4 and thinking I'm in the main flat part of the curve where 0.2V delta means a 10% SoC difference for a 24V battery (this also aligns with how the cells were charging to start with).

Then taking the example from my last screenshots. ~8 hours at 7 amps (I think this is conservative given that I dialed up to 10A pretty soon after that first screenshot) means that I put in ~56Ah or ~9% of the 608Ah battery. The total voltage only changed by 0.01V (and as mentioned it was oscillating back to 26.94 regularly). It should have been much closer to the 0.2V increase.

Also I was seeing that relatively linear increase in voltage up to that point (no screenshots since everything seemed to be working as I expected).

No. All cells in series MUST pass the same current. It might contribute to your voltage drop between supply and battery terminals.

gotcha

If the BMS registers a current, it's accepting charge. At the very low rate you're charging, watching paint dry will be like an orgy of joy compared to watching the voltage rise @ 10A on a 608Ah pack.

Lol - good thing I've also got some painting/polyurethaning to do on my van build.

 

[FilterGuy]

Be patient grasshopper. Be patient.

 

[FilterGuy]

It's my first time using a benchtop PSU and it was offgassing some pretty noxious stuff at the beginning. I'm guessing rosin heating up? At any rate it spooked me a bit so I've dialed down the current several times.

Yeah I've seen it - maybe I'm misunderstanding something though. I've been looking at charts like this one for lifepo4 and thinking I'm in the main flat part of the curve where 0.2V delta means a 10% SoC difference for a 24V battery (this also aligns with how the cells were charging to start with).

Then taking the example from my last screenshots. ~8 hours at 7 amps (I think this is conservative given that I dialed up to 10A pretty soon after that first screenshot) means that I put in ~56Ah or ~9% of the 608Ah battery. The total voltage only changed by 0.01V (and as mentioned it was oscillating back to 26.94 regularly). It should have been much closer to the 0.2V increase.

Also I was seeing that relatively linear increase in voltage up to that point (no screenshots since everything seemed to be working as I expected).
View attachment 182239


gotcha


Lol - good thing I've also got some painting/polyurethaning to do on my van build.

There are people on the forum that disagree with me but I consider that chart garbage. The only thing you can tell with voltage is that the battery is nearly empty or nearly full. Trying to use voltage for anything in-between is a mistake

 

[sunshine_eggo]

There are people on the forum that disagree with me but I consider that chart garbage. The only thing you can tell with voltage is that the battery is nearly empty or nearly full. Trying to use voltage for anything in-between is a mistake

This!

@alexcl

Take a look at the chart from a consistency standpoint...

30-99% is covered by 13.0-13.4 (0.4V), but there's a full volt from 99-100% (13.4-14.4). Various 0.1V drops vary from 10-30% losses in SoC.

At best the chart is good for a ±15% approximation of a battery's SoC after it's been sitting completely unused for 30 minutes.

 

[Daddy Tanuki]

It's my first time using a benchtop PSU and it was offgassing some pretty noxious stuff at the beginning. I'm guessing rosin heating up? At any rate it spooked me a bit so I've dialed down the current several times.

Yeah I've seen it - maybe I'm misunderstanding something though. I've been looking at charts like this one for lifepo4 and thinking I'm in the main flat part of the curve where 0.2V delta means a 10% SoC difference for a 24V battery (this also aligns with how the cells were charging to start with).

Then taking the example from my last screenshots. ~8 hours at 7 amps (I think this is conservative given that I dialed up to 10A pretty soon after that first screenshot) means that I put in ~56Ah or ~9% of the 608Ah battery. The total voltage only changed by 0.01V (and as mentioned it was oscillating back to 26.94 regularly). It should have been much closer to the 0.2V increase.

Also I was seeing that relatively linear increase in voltage up to that point (no screenshots since everything seemed to be working as I expected).
View attachment 182239


gotcha


Lol - good thing I've also got some painting/polyurethaning to do on my van build.

trying to gauge lithium using voltage is like trying to gage full on a 13 year olds stomach by asking them... its worthless as 5 minutes later they want more.

you are dealing with 8 cells.... last month i top balanced 48 cells in parallel using a 40 amp power supply. patience is the key, any attempts at shortcuts will bite you in the ass.

I started with them in series parallel with a BMS and ran the pack up till it shut off due to one cell (3 cells) being over 3.600 volts that was quick and the rest of the cells were at 3.580. i then wired them all in parallel and put the power supply back on them and left for the week. came back and they had just finished top balancing by about 4 hours.

sorry got cut off.... so that was me top balancing after they had been running a year and had started to drift due to a couple of bad cells. So I consolidated the pack and top balanced once more. they were already close to full when i started and it still took close to a week with a 40 amp power supply. if you 8 cells take a week with a 10 amp power supply i would count that as lucky.

 

[alexcl]

In case it helps anyone in the future, patience was the key. Thanks for the intel.

 

[Nonlin]

I only got a 30V 10 amp charger but have a 16S for 48 Volt bank. Do I need to split my banks up and top balance them 24Volts at a time 8s and 8s or just go a head and top balance them all together?

 

[time2roll]

I only got a 30V 10 amp charger but have a 16S for 48 Volt bank. Do I need to split my banks up and top balance them 24Volts at a time 8s and 8s or just go a head and top balance them all together?

Typically the top balance would be all parallel charged to 3.650 volts before the stack is assembled in series for service.
I advocate to just assemble in series with an active balancer. Place in service protected with a BMS. If needed limit the top voltage until the cells are balanced.

 

[Zwy]

I only got a 30V 10 amp charger but have a 16S for 48 Volt bank. Do I need to split my banks up and top balance them 24Volts at a time 8s and 8s or just go a head and top balance them all together?

I use a 24V JBD BMS and charge up 8 cells until one hits the cell high voltage cutoff I had set around 3.60V and BMS disconnects. I then put the cells in parallel and finish the top balance of those 8 cells. My power supply's don't go to 48V. I then do the other 8 cells. I do however capacity test the 8 cells using the 24V BMS and recharge before assembling all 16 cells into a 48V battery.

Don't charge in series without a BMS. If you only have a 48V BMS then use it with all the cells in series but you will need a 48V charger. After one cell hits cell high voltage disconnect, then put cells in parallel to finish top balance.

With a BMS like the JK with 2A cell balancing, you could get by cutting down charging amps and let the BMS do the balancing.

The other option is all 16 cells in parallel, power supply set to 3.65V but it will take about 2 weeks to top balance.

 

[Nonlin]

I use a 24V JBD BMS and charge up 8 cells until one hits the cell high voltage cutoff I had set around 3.60V and BMS disconnects. I then put the cells in parallel and finish the top balance of those 8 cells. My power supply's don't go to 48V. I then do the other 8 cells. I do however capacity test the 8 cells using the 24V BMS and recharge before assembling all 16 cells into a 48V battery.

Don't charge in series without a BMS. If you only have a 48V BMS then use it with all the cells in series but you will need a 48V charger. After one cell hits cell high voltage disconnect, then put cells in parallel to finish top balance.

With a BMS like the JK with 2A cell balancing, you could get by cutting down charging amps and let the BMS do the balancing.

The other option is all 16 cells in parallel, power supply set to 3.65V but it will take about 2 weeks to top balance.

Seems cheaper to buy two 30V max 10 amp max DC chargers. I guess I could use both to each charge 8s? Would cut the charge time in half right? 1 Week instead of 2 weeks?

The BMS I have has only a 0.6 amp active balancer so that will be really slow to strictly depend on for balancing.

 

[time2roll]

Seems cheaper to buy two 30V max 10 amp max DC chargers. I guess I could use both to each charge 8s? Would cut the charge time in half right? 1 Week instead of 2 weeks?

The BMS I have has only a 0.6 amp active balancer so that will be really slow to strictly depend on for balancing.

Relying on the active balancer would get the system running two weeks earlier. Even if top voltage is limited I speculate at least 90% of capacity will be immediately available. Most cells seem to be far better quality than even a few years ago and very close to same charge level. Worst case a couple of the low cells could be boost charged while in service.

 

[Zwy]

Seems cheaper to buy two 30V max 10 amp max DC chargers. I guess I could use both to each charge 8s? Would cut the charge time in half right? 1 Week instead of 2 weeks?

In 8S, the cells would alternate on connections. + to - on each cell to the adjoining cell. This would require using a BMS as the voltage would be nominal 24V. Do you have a 24V BMS? My guess is you are building a 48V pack and have a 48V BMS but not sure if you have a 24V BMS.

Are you using a power supply or charger? With the power supply you can turn the voltage down to 3.65V and charge using C.V. (constant voltage) with cells in parallel.

Would be best to let us know what you have for charging and what you have for any BMS's. Photos or links to what you have.

The BMS I have has only a 0.6 amp active balancer so that will be really slow to strictly depend on for balancing.

What do you for a BMS? The idea behind using the BMS is to charge cells up quickly until one cell causes the disconnect, then finish balancing cells in parallel. It doesn't take long to finish balance on 8 or 16 cells in parallel after charging using the BMS A few hours and it will be done.

 

[Nonlin]

In 8S, the cells would alternate on connections. + to - on each cell to the adjoining cell. This would require using a BMS as the voltage would be nominal 24V. Do you have a 24V BMS? My guess is you are building a 48V pack and have a 48V BMS but not sure if you have a 24V BMS.

Are you using a power supply or charger? With the power supply you can turn the voltage down to 3.65V and charge using C.V. (constant voltage) with cells in parallel.

Would be best to let us know what you have for charging and what you have for any BMS's. Photos or links to what you have.


What do you for a BMS? The idea behind using the BMS is to charge cells up quickly until one cell causes the disconnect, then finish balancing cells in parallel. It doesn't take long to finish balance on 8 or 16 cells in parallel after charging using the BMS A few hours and it will be done.

JK BMS 8-17S - 80A Continuous - 150A Peak - 0.6A Active Balancing - JK-BD6A17S8P

www.18650batterystore.com

Is my BMS.

I wasn't going to charge them either BMS just a DC charger but all I have is this Jesverty DC Power Supply Variable, 0-30V 0-10A Adjustable Switching DC Regulated Bench Power Supply with High Precision 4-Digit LED Display, 5V/2A USB Port, Coarse and Fine Adjustment SPS-3010 https://a.co/d/5vuiH5g


But I was thinking I get a second one of those and then could charge 2x 24 volts at 10 amps? Maybe I can wire the chargers in series to achieve that.

 

[Zwy]

JK BMS 8-17S - 80A Continuous - 150A Peak - 0.6A Active Balancing - JK-BD6A17S8P

www.18650batterystore.com

Is my BMS.

I wasn't going to charge them either BMS just a DC charger but all I have is this Jesverty DC Power Supply Variable, 0-30V 0-10A Adjustable Switching DC Regulated Bench Power Supply with High Precision 4-Digit LED Display, 5V/2A USB Port, Coarse and Fine Adjustment SPS-3010 https://a.co/d/5vuiH5g


But I was thinking I get a second one of those and then could charge 2x 24 volts at 10 amps? Maybe I can wire the chargers in series to achieve that.

No need for a second power supply. Use the BMS for initial charging in series using 8 cells with your power supply set to 29.0V, set the BMS to high cell volt disconnect at 3.65V. When the BMS disconnects, break down the pack and put cells in parallel. Set power supply to 3.65V and when amperage drops to almost 0A, you are done. Then do the other 8 in series using the BMS the same way.

This will shorten the time considerably.

The other way to let the BMS balance cells is to charge to around 28V or less. When the 8 cell pack pack hits 28V, the BMS should have started balance. You can limit the amps on the power supply so the BMS can passive balance the cells. Once cells are balanced at that level, the voltage is increased by 0.5V, amps kept low and let the BMS balance cells again to that voltage level. You can balance all the way up to 29.2V but it isn't really necessary. If you plan on charging to 3.5V per cell as an example, (which is 28V) when you are using the battery, then just balance cells to that point. I run my 48V pack to 56V bulk and 55.5V float. This allows balancing to occur and cells are about 99%+ SOC.

 

[FilterGuy]

The other way to let the BMS balance cells is to charge to around 28V or less. When the 8 cell pack pack hits 28V, the BMS should have started balance. You can limit the amps on the power supply so the BMS can passive balance the cells. Once cells are balanced at that level, the voltage is increased by 0.5V, amps kept low and let the BMS balance cells again to that voltage level. You can balance all the way up to 29.2V but it isn't really necessary. If you plan on charging to 3.5V per cell as an example, (which is 28V) when you are using the battery, then just balance cells to that point. I run my 48V pack to 56V bulk and 55.5V float. This allows balancing to occur and cells are about 99%+ SOC.

This method works best with active balancers that have a reasonably high balance current. Ideally, the charge current during this balancing process would be less than the balance current. Otherwise one of the cells will hit the cell voltage limit too soon.

 

[Nonlin]

No need for a second power supply. Use the BMS for initial charging in series using 8 cells with your power supply set to 29.0V, set the BMS to high cell volt disconnect at 3.65V. When the BMS disconnects, break down the pack and put cells in parallel. Set power supply to 3.65V and when amperage drops to almost 0A, you are done. Then do the other 8 in series using the BMS the same way.

This will shorten the time considerably.

The other way to let the BMS balance cells is to charge to around 28V or less. When the 8 cell pack pack hits 28V, the BMS should have started balance. You can limit the amps on the power supply so the BMS can passive balance the cells. Once cells are balanced at that level, the voltage is increased by 0.5V, amps kept low and let the BMS balance cells again to that voltage level. You can balance all the way up to 29.2V but it isn't really necessary. If you plan on charging to 3.5V per cell as an example, (which is 28V) when you are using the battery, then just balance cells to that point. I run my 48V pack to 56V bulk and 55.5V float. This allows balancing to occur and cells are about 99%+ SOC.

So the BMS's 0.6 amp balancing isn't to low to depend on? You say passive balance but I think the BMS does active? Or is passive just something that occurs anyways?

Also you say ~28Volts but 16s would be for 48V system?

 

[Zwy]

This method works best with active balancers that have a reasonably high balance current. Ideally, the charge current during this balancing process would be less than the balance current. Otherwise one of the cells will hit the cell voltage limit too soon.

One can turn the amperage down with the power supply. My Batrium uses 0.75A for balancing bypass and it has to end the load as the heat sink builds up heat. I've seen 120A heading into a 64 cell bank and it handles it fine. I have one major runner that will be heading for use in the golf cart but the Batrium still handles it well in bypass.