Top Balancing "How to"

[Hedges]

At a minimum the bulb will eliminate the pre-charge step. If the benchtop power supply is low amperage, might as well leave the bulbs in place while top balancing. In theory if a bad cell shorts, the voltage difference will light the lamp.

"Precharge", I think you mean charge to similar voltage before connecting in parallel. Which probably means something if some cells are near 100 SoC (above knee of curve) but others are below. If all below knee, I would test with the light bulb and see how close voltage comes. Calculate V/0.00025 ohms to see if surge current is a concern.

Some people have completed parallel top balancing, then when using battery found SoC differed between cells. One was low. Going back to charge that one alone to 3.65V fixed the issue. I blamed it on high resistance contact, so during charging that cell lagged.

If you deliberately add resistance in series with cells while charging, I think they will lag, take far longer to drift up to same SoC.

Finally, I think I have it... Are you proposing to connect cells in parallel using bulbs instead of bus bars? If so, I don't see any useful purpose in that.

Even at low current, they have higher resistance than bus bars and will increase the needed Wh to charge the pack.

Not Wh, I don't think, unless charge current is high enough to burn significant watts in the bulb.
But h (hours) definitely.

 

[sunshine_eggo]

"Precharge", I think you mean charge to similar voltage before connecting in parallel. Which probably means something if some cells are near 100 SoC (above knee of curve) but others are below. If all below knee, I would test with the light bulb and see how close voltage comes. Calculate V/0.00025 ohms to see if surge current is a concern.

Some people have completed parallel top balancing, then when using battery found SoC differed between cells. One was low. Going back to charge that one alone to 3.65V fixed the issue. I blamed it on high resistance contact, so during charging that cell lagged.

If you deliberately add resistance in series with cells while charging, I think they will lag, take far longer to drift up to same SoC.

Not Wh, I don't think, unless charge current is high enough to burn significant watts in the bulb.
But h (hours) definitely.

If you apply a given amps at a given voltage for more hours, isn't that more Wh?

 

[venquessa]

New half of a battery. So I put them into a pack and swapped it for the old one.

This is the top balance of the new 8S pack. On the left last night with the Multiplus battery charger and careful manipulation of the absorb voltage and charge current and ... on the right is the sun with my MPPT limiter automating that process. It looks like the automation did a much better job than I did with VEConfig!

Heltec 5A balancer. JK BMS - balancing disabled. As of 10am the loads came online. The MPPT limiter is still triggering on some cell delta at the top end, but it looks balanced enough to me.

 

[Tulex]

Sorry if this has been covered, so much to read through, so probably has been.

Top balanced 16 304ah batteries at 3.65v. When it got down to .1amp, checked batteries, they showed 3.65v. Waited a little while, checked, showed 3.67v, still at .1 amps. Dropped the charger to 3.64v, and it stopped charging, showed zero amps. So, disconnected the charger, checked, they all showed 3.65 amps. Let them sit overnight, and this morning they all show 3.627v. Is this normal?

Also, batteries are Eve, IR: 0.2m ohm, 320-326 Ah, they all started at 3.292v.

 

[sunshine_eggo]

Top balanced 16 304ah batteries at 3.65v. When it got down to .1amp, checked batteries, they showed 3.65v. Waited a little while, checked, showed 3.67v, still at .1 amps. Dropped the charger to 3.64v, and it stopped charging, showed zero amps. So, disconnected the charger, checked, they all showed 3.65 amps. Let them sit overnight, and this morning they all show 3.627v. Is this normal?

LFP is like lead acid. It must be charged to an elevated voltage to attain full charge. It is very common for voltage to drop over time. Given the extremely low current and 3.65V charge voltage, your cells likely have a pervasive "surface charge" on them where residual unusable charge keeps the cell voltages elevated. A bigger drop than you are seeing is more typical of cells in normal use, but this isn't uncommon after a balance charge to such low current.

Please refer to the cell datasheet. It likely says that the cell is fully charged at 3.65V AND 0.05C. In this case, that's 15A, i.e., once the cell hits 3.65V WITH 15A going into it, it's FULL. Any additional input does not yield meaningful usable capacity. By charging 16 304Ah cells in parallel, that's 4864Ah ... 0.05C is 243A, so assuming all cells shared current equally, they were full when they hit 3.65V and 243A. Of course, they likely don't share it equally, and nobody is using a 243A power supply. So, when your 10A power supply showed 10A with the cells measured at 3.65V, they were already full. Anything additional was essentially unusable charge.

 

[Tulex]

LFP is like lead acid. It must be charged to an elevated voltage to attain full charge. It is very common for voltage to drop over time. Given the extremely low current and 3.65V charge voltage, your cells likely have a pervasive "surface charge" on them where residual unusable charge keeps the cell voltages elevated. A bigger drop than you are seeing is more typical of cells in normal use, but this isn't uncommon after a balance charge to such low current.

Please refer to the cell datasheet. It likely says that the cell is fully charged at 3.65V AND 0.05C. In this case, that's 15A, i.e., once the cell hits 3.65V WITH 15A going into it, it's FULL. Any additional input does not yield meaningful usable capacity. By charging 16 304Ah cells in parallel, that's 4864Ah ... 0.05C is 243A, so assuming all cells shared current equally, they were full when they hit 3.65V and 243A. Of course, they likely don't share it equally, and nobody is using a 243A power supply. So, when your 10A power supply showed 10A with the cells measured at 3.65V, they were already full. Anything additional was essentially unusable charge.

Thanks for the reply
My charger didn't behave like you describe, never showed full amps (10 max), but instead stayed on voltage (showed the set voltage the whole time). So instead of showing being on amps and charging full amps and watching the voltage rise, it stayed on voltage and watched the amps drop. Don't think I ever saw higher than 7 amps.

 

[sunshine_eggo]

Thanks for the reply
My charger didn't behave like you describe, never showed full amps (10 max), but instead stayed on voltage (showed the set voltage the whole time). So instead of showing being on amps and charging full amps and watching the voltage rise, it stayed on voltage and watched the amps drop. Don't think I ever saw higher than 7 amps.

This is likely due to the quality or lack thereof of your charge leads. One of the first items in the guide is a discussion on the need to construct higher quality leads. The losses in your cabling resulted in the charger "seeing" a voltage higher than the cells thus reducing the current. Even higher quality leads will see a reduction in current near the end of the charge.

 

[Tulex]

This is likely due to the quality or lack thereof of your charge leads. One of the first items in the guide is a discussion on the need to construct higher quality leads. The losses in your cabling resulted in the charger "seeing" a voltage higher than the cells thus reducing the current. Even higher quality leads will see a reduction in current near the end of the charge.

Thanks again. This unit was advertised as having heavy duty leads, and they seemed bigger than normal, but I will make new ones.

 

[Zwy]

Thanks again. This unit was advertised as having heavy duty leads, and they seemed bigger than normal, but I will make new ones.

10AWG and ring terminals only. Throw alligator clips in the trash.

 

[Hedges]

Other possible explanations are high contact resistance (people here sometimes have one cell lagging the others), or cells were so close to 100% the power supply was CC from the beginning.

The test would be DMM between power supply and cell terminal itself.

 

[DIYrich]

Was thinking of a simple method of balancing the new cells. Wire all cells in parallel, but with a 12V incandescent bulb in series with each cell. Would eliminate having to manually charge each battery to a similar voltage. Any flaw with my hypothesis?

Once you wire in parallel, the cells can share unlimited power between them. That is why they need to be close before putting in parallel. What do you hope to accomplish with the bulb?

Not possible to simultaneously wire in parallel and series the same cells. You end up with a short.

 

[Just John]

Let them sit overnight, and this morning they all show 3.627v. Is this normal?

Yes.

 

[sunshine_eggo]

Thanks again. This unit was advertised as having heavy duty leads, and they seemed bigger than normal, but I will make new ones.

Follow the guide.

 

[Ampster]

This unit was advertised as having heavy duty leads,

Heavy Duty is a marketing term. That is how we get fooled.

 

[bass-o-matic]

misfire.

 

[Hedges]

Capacity test drained individual cells each to low voltage cutoff? Varying number of Ah?
Yeah, I'd say top balance them again.
But charging in parallel will take forever, longer than when you first got them.
You could charge them in series with BMS protecting, until the first cell high voltage disconnect. Then top them off individually or in parallel (different SoC, use some method to get them to same voltage before paralleling, like individual charge.)

 

[hardtop]

I have 4 EVE LF304 Grade A Cells - 3.2V and a JK BMS 4-8S. Each battery measures 3.307v, 3.308v, 3.308v, 3.307 volts. What are the consequences if I don't top balance?

 

[MisterSandals]

What are the consequences if I don't top balance?

If they are matched well enough to charge to your top charge voltage without hitting a cell over volt, then you are good to go.

If you find that you cannot charge to that voltage, you can lower your charge voltage or try to top balance them to see if that helps.

 

[hardtop]

If they are matched well enough to charge to your top charge voltage without hitting a cell over volt, then you are good to go.

If you find that you cannot charge to that voltage, you can lower your charge voltage or try to top balance them to see if that helps.

It sounds like there is a possibility top balancing isn't needed. I'm planning to get the Victron Quattro-II https://www.currentconnected.com/product/victron-quattro-ii-12v-3kva-2x-120v-inverter/ charger/inverter and I have the JK BMS 4-8S. How will I know if I have a cell over volt condition?

 

[FilterGuy]

How will I know if I have a cell over volt condition?

Your BMS app will tell you there is a Cell overvoltage.

If you have one of the JK BMSs with an active balancer, it will balance your cells for you, but you need to have the system set correctly.

(This was corrected in an edit to the original post)
* Set the inverter bluk charge to 13.8V for a 12V system.
* Set the inverter accumulation time to an hour or two.
* Set the BMS cell overvoltage to 3.65V
* Set the BMS to start balancing at 3.45V

 

[42OhmsPA]

Your BMS app will tell you there is a Cell overvoltage.

If you have one of the JK BMSs with an active balancer, it will balance your cells for you, but you need to have the system set correctly.

Please check your listed voltages...
Excellent advice on a BMS with an active balancer or secondary active balancer if your BMS only has passive balancing.
If you don't plan to use an active balancer I would say top balancing is a must, unless you want to let the cells float for days.

I'm running 3 JBDs. 2 packs were top balanced, 1 wasn't. I see a high deviation at the top of the knee on the pack I didn't top balance.

Regardless what BMS you use you must use one!

 

[FilterGuy]

Please check your listed voltages...

OOPS. It should be:
* Set the inverter bluk charge to 18V for a 12V system.
* Set the inverter accumulation time to an hour or two.
* Set the BMS cell overvoltage to 3.65V
* Set the BMS to start balancing at 3.45V

It is corrected above as well.

 

[HRTKD]

18 volt bulk charge seems a bit high.

 

[FilterGuy]

18 volt bulk charge seems a bit high.

crap.... my mind is out to lunch. One more try:

* Set the inverter bluk charge to 13.8 V for a 12V system.
* Set the inverter accumulation time to an hour or two.
* Set the BMS cell overvoltage to 3.65V
* Set the BMS to start balancing at 3.45V

 

[HRTKD]

crap.... my mind is out to lunch. One more try:

* Set the inverter bluk charge to 13.8 V for a 12V system.
* Set the inverter accumulation time to an hour or two.
* Set the BMS cell overvoltage to 3.65V
* Set the BMS to start balancing at 3.45V

I was going to ask who you were and what you did with FilterGuy.