charging/discharging LiFePo efficiently to maximize life

[sunshine_eggo]

are you saying that the bms hitting limits too many times, which it is designed to do, will kill it? that if it shuts off charging because a cell got to 3.61v more than 30 times, it's (maybe) done for? looking back, sunshine did say the same thing....that seems strange to me, that it's made to do that but doing so will destroy it... from the current connected literature, and from what Dexter said, it seemed normal for it to go to OVP, at least while it's getting things in order.

We're not talking about 30 times. Imagine a poorly designed/horribly imbalanced system that hits OVP 5+ times a day... that's almost 2,000/year. We mention it because we've seen it happen on this forum. The BMS is a safety device. It should only activate when absolutely necessary, and the system should work INSIDE the BMS limits.

Besides the affect on the BMS itself, consider the impact on connected chargers. Suddenly removing the load from a charger means the charger is going to subject the system to a voltage spike because there's a short delay in the circuitry reacting to the loss of load. Again, we see this here too where 12V systems are spiking to over 20V due to BMS OVP. Inverters setting of over-voltage alarms and shutting down due to voltage being too high to operate.

if that's the case, i guess i will change my bulk/absorb to 3.45v and see what happens....where the cell voltages stop. i'm only guessing that the bms shuts it off at just over 3.6v.

Good.

i just checked, after running them for a few days w/o charging (and it's been sitting idle for 10 hrs as i shut everything off before i left for work), i'm at about 14% SOC w/ cell voltages at 3204mv. almost all are at 3204mv, a cpl are either 3203 or 3205.

Recommend you get into the habit of seeing only X.XX on cell voltages. 3.20-3.21 is a narrow range, and it's to be expected even when imbalanced because they're in the operating range (3.1-3.4V).

 

[curiouscarbon]

With a MOSFET BMS especially, damage on disconnect occurs when amperes/current is flowing and suddenly has to go somewhere.

For over cell voltage or overcurrent, naturally current will be flowing and elevate the risk of hardware damage of BMS or other components.

If a MOSFET BMS disconnects with 0 Amps flowing, no damage should occur even after thousands of times.

But current will be expected to be flowing in these situations, so rely on it as an absolute last resort and program chargers etc to stay within BMS limits. my 2¢

 

[defed]

We're not talking about 30 times. Imagine a poorly designed/horribly imbalanced system that hits OVP 5+ times a day... that's almost 2,000/year. We mention it because we've seen it happen on this forum. The BMS is a safety device. It should only activate when absolutely necessary, and the system should work INSIDE the BMS limits.

Besides the affect on the BMS itself, consider the impact on connected chargers. Suddenly removing the load from a charger means the charger is going to subject the system to a voltage spike because there's a short delay in the circuitry reacting to the loss of load. Again, we see this here too where 12V systems are spiking to over 20V due to BMS OVP. Inverters setting of over-voltage alarms and shutting down due to voltage being too high to operate.

Recommend you get into the habit of seeing only X.XX on cell voltages. 3.20-3.21 is a narrow range, and it's to be expected even when imbalanced because they're in the operating range (3.1-3.4V).

i only said 30 cycles because MisterSandals had mentioned figures...and i guess he said more like 100 could be a problem. if that is the case, WHY does current connected recommend a charging voltage that is apparently going to trip the OVP EVERY cycle!?! at least until it's better balanced (i presume, if it ever improves). if i hadn't come here asking stupid questions, i would have just never even known it was possible to kill the bms.

i haven't watched close enough to see when it goes to OVP, because who can be there at the exact moment things happen...is it during bulk where there could be a bit of current or after it switches to absorb, where the current should be tapering off? the most i can charge at is about 35a and i don't think i get there very often, so i am generally charging at lower rates for what it's worth.

 

[defed]

i was going to make a comparison of the idea of the bms wearing out to car brakes, and how crazy it sounds to me...car company: use your brakes to stop, but use them too much and they will fail...

well that just may be the perfect analogy...because that IS what happens. if you brake really hard all the time, they won't last long...brake gently and they will last a long long time. the same can be said about the bms (or any electronic perhaps).

 

[MisterSandals]

if that is the case, WHY does current connected recommend a charging voltage that is apparently going to trip the OVP EVERY cycle!?!

Lets bring Current Connected in with a little incantation: @Dexter

I'm willing to learn and/or eat crow if necessary...

 

[defed]

Lets bring Current Connected in with a little incantation: @Dexter

I'm willing to learn and/or eat crow if necessary...

i'm just trying to learn too! the last thing i want to do is shorten the battery life by a significant amount doing something that shouldn't be done. which is why i'm always asking (probably) dumb questions.


from their manual:

• Charge Current for normal use: Up to 63A (Per Battery)
• Charge Current for fast charge: Up to 95A (Per Battery)
• Absorption Voltage: 57.6v recommended; up to 58.4v max
• Absorption Time: 15 Minutes
• Float Voltage: 55.2v
• Charging Target: 100% SOC
• Low Voltage Shutdown: 48v

 

[sunshine_eggo]

Lets bring Current Connected in with a little incantation: @Dexter

I'm willing to learn and/or eat crow if necessary...

Is that really Dexter?

@HighTechLab is usually what works for me...

i'm just trying to learn too! the last thing i want to do is shorten the battery life by a significant amount doing something that shouldn't be done. which is why i'm always asking (probably) dumb questions.


from their manual:

• Charge Current for normal use: Up to 63A (Per Battery)
• Charge Current for fast charge: Up to 95A (Per Battery)
• Absorption Voltage: 57.6v recommended; up to 58.4v max

3.60-3.65V/cell

Same manual says cut-off is 3.7V, so by design, a balanced battery will not regularly trigger OVP. So,

if that is the case, WHY does current connected recommend a charging voltage that is apparently going to trip the OVP EVERY cycle!?!

They don't.

• Absorption Time: 15 Minutes
• Float Voltage: 55.2v

3.45V/cell - this is higher than what is typically recommended (3.375V) due to low current overcharge concerns. At this level, balancing can be used at float.

In other words, both charge voltages give the BMS opportunity to passively balance.

What I would love to know is if this particular BMS will balance based on voltage alone, or does it require charging for balancing.

@HighTechLab ?

 

[MisterSandals]

Is that really Dexter?

@HighTechLab is usually what works for me...

Oops, you're right. Thanks.

 

[defed]

3.60-3.65V/cell

Same manual says cut-off is 3.7V, so by design, a balanced battery will not regularly trigger OVP. So,

you are right, something else i missed (or didn't understand all the data at the time). but pretty sure mine has gone into OVP every charge cycle so far (3x) and i hadn't ever seen it at 3.7v. the very 1st cycle i was there pretty quick after it completed but then again, i don't know how fast it drops down once it's off. seems like it stays in the 3.6's for a little while.

yes, probably because it's not balanced, but i don't know how to get it there except by doing what it says....10-15 cycles, but then that's 10-15 bms overloads by then.

 

[HighTechLab]

What I would love to know is if this particular BMS will balance based on voltage alone, or does it require charging for balancing.

The BMS does not require current flowing in. Balancing triggers based on voltage threshold and deviation. Balancing starts when a cell reaches 3.45v with a 20mV deviation. Balancing will continue during charge AND OVP but stops when the battery has current exiting (discharging). We recommend floating at 3.45 so the balancing can continue even in the float stage. This has been very helpful in getting new batteries balanced out quickly.

OVP triggers at 3.7v in batteries shipped to customers in 2022. We changed to 3.65 on batteries shipped in 2023 to reduce triggering "high voltage" warnings on some CANbus connected Victron systems and also changed to give us more margin. This is because the curves are so steep at this voltage, that if you are charging at rated max current without CAN communication, the voltage curve can overrun the reaction time of the BMS. While very few ever max out the charging c-rate of the BMS without communication connected, we needed to remain safe in every corner case. Danger point never to cross according to cell MFG is 3.8v, so with this new setting we safely stop before 3.7 and never come close to 3.8 even in the worst of scenarios.

As far as wearing out the BMS with OVP, it's not going to happen on this hardware. We setup a test jig on this BMS hardware in 2022 and put over 250,000 OVP cycles (one cycle every 5 seconds) on a BMS interrupting a 50A current. Fotunately and unfortunately, there were zero failures - I wish we could have had a failure to know what that point looks like... This test is equivalent to triggering OVP over 65 times per day, every day for 10 years. It's very unlikely that you will ever trigger OVP this many times.

That said, as pointed out, it's not as good on the hardware connected to it due to potential voltage spikes, but it's not like it's going to cause the BMS to short and allow the battery to be overcharged and blow up. If communication is used, the current tapers down before triggering OVP so it's not a very big "water hammer" voltage spike effect at all, even if OVP is still triggered.

Some parts of this thread I skimmed over, so if I missed a question or something please let me know. Hope I covered everything!

 

[sunshine_eggo]

you are right, something else i missed (or didn't understand all the data at the time). but pretty sure mine has gone into OVP every charge cycle so far (3x) and i hadn't ever seen it at 3.7v. the very 1st cycle i was there pretty quick after it completed but then again, i don't know how fast it drops down once it's off. seems like it stays in the 3.6's for a little while.

Imbalance in new batteries is extremely common. Top balancing cells prior to build would be prohibitively expensive, AND batteries sitting during transport or in inventory can cause a loss of top balance. Hazmat shipping requirements limit LFP batteries to be at 30% SoC when shipped.

These factors nearly insure a battery will arrive at the customer in an imbalanced state.

yes, probably because it's not balanced, but i don't know how to get it there except by doing what it says....10-15 cycles, but then that's 10-15 bms overloads by then.

This number of events is not a concern. Our comments have been directed at deploying a system where the BMS routinely has to engage protection. There are MANY posts where folks state they simply plan to allow the BMS to cut charge every time.

 

[defed]

Imbalance in new batteries is extremely common. Top balancing cells prior to build would be prohibitively expensive, AND batteries sitting during transport or in inventory can cause a loss of top balance. Hazmat shipping requirements limit LFP batteries to be at 30% SoC when shipped.

These factors nearly insure a battery will arrive at the customer in an imbalanced state.



This number of events is not a concern. Our comments have been directed at deploying a system where the BMS routinely has to engage protection. There are MANY posts where folks state they simply plan to allow the BMS to cut charge every time.

balance starts at 3.45v as per Dexter. let's say i charge to 3.45v and the cells all balance out there over a few cycles w/o tripping OVP since i'm quite a bit below the bms cut off limit of 3.7v. could one, in theory, then gradually increase the voltage to 3.5v, 3.55v, 3.6v (because hopefully the cells all reach these lvls together as they were balanced at 3.45v so there should be less variation between them) and eventually get it balanced there as well and then use the recommended charge value of 3.6v and not trip OVP? let me just repeat, this is in theory, as you all have stated that the SOC difference between 3.45v and 3.6v is very minimal and not really worth chasing. i'm just trying to wrap my head around how it all works.

it also appears that even if i did just leave it at the 3.6v charge level, it wouldn't hurt the bms as per Dexter's testing. but it is good to know the value at which balancing begins and where the bms cuts it off.

i really appreciate all of you taking the time to try and get this stuff through my thick skull! thank you.

 

[sunshine_eggo]

balance starts at 3.45v as per Dexter. let's say i charge to 3.45v and the cells all balance out there over a few cycles w/o tripping OVP since i'm quite a bit below the bms cut off limit of 3.7v. could one, in theory, then increase the voltage to 3.5v, 3.55v, 3.6v (because hopefully the cells all reach these lvls together as they were balanced at 3.45v so there should be less variation between them) and eventually get it balanced there as well and then use the recommended charge value of 3.6v and not trip OVP?

Yes. Once cells are balanced at 3.45V, there will be very little, if any, imbalance up to 3.65V.

let me just repeat, this is in theory, as you all have stated that the SOC difference between 3.45v and 3.6v is very minimal and not really worth chasing. i'm just trying to wrap my head around how it all works.

Yep.

it also appears that even if i did just leave it at the 3.6v charge level, it wouldn't hurt the bms as per Dexter's testing. but it is good to know the value at which balancing begins and where the bms cuts it off.

Agreed. I'm encouraged to know they went to this level of effort to establish a BMS life. Not all BMS are created equal, and while we've seen several instances of "lesser" BMS failing - and failing closed, i.e., they will allow charging even if OVP criteria are met, this is a more robust animal.

i really appreciate all of you taking the time to try and get this stuff through my thick skull! thank you.

 

[featherlite]

I just read the specs for the SOK batteries per the Current Connected web page. I am impressed.

 

[sunshine_eggo]

I just read the specs for the SOK batteries per the Current Connected web page. I am impressed.

Yep. It feels like they've taken all the shit out of the Pylontech-type system and made it all make sense.

 

[defed]

Yes. Once cells are balanced at 3.45V, there will be very little, if any, imbalance up to 3.65V.

i wasn't home when the sun went down so i can't be sure what they got to, but the c/c said it produced 9.11kwh. that should have been enough to get them to 3.45v.

so i check my cell voltages when i get home from work, all are 3.338v and 3.339v (53.4v)....so pretty well balanced after resting. let's say it did get to 55.2v (3.45v) i'd imagine some cells probably were a fair bit higher than 3.45v...does this mean the lowest were 3.33v and that's the level it brings all the high ones down to match?

i checked the bms history and it already has 20 OVPs, but only 4 charge cycles (after this one). i suppose this means that when i had it at 57.6v, and cells got to 3.7v, it shut down....did some balancing, settled at, let's say 55v, the charger kicked back on the next day and took them back to 57.6v and some obviously tripped at 3.7v again, even though it wasn't a 'charge cycle'...that would account for all the OVPs despite having few cycles. i thought it was only at 3 OVPs....but not that i know, i can see if it trips when charging to 3.45v. i can't imagine any are that far out, but i will should find out tomorrow.

 

[Ampster]

i check my cell voltages when i get home from work, all are 3.338v and 3.339v (53.4v)....so pretty well balanced after resting.

Resting voltage is not a good indicator of how well cells are balanced at the top. The only way to know if you have a pack that is top balanced is to observe the cell voltages when the pack average is above 3.45 volts per cell.

 

[defed]

Resting voltage is not a good indicator of how well cells are balanced at the top. The only way to know if you have a pack that is top balanced is to observe the cell voltages when the pack average is above 3.45 volts per cell.

i wasn't sure if it meant anything or not, as i said, i was at work when they finished so i didn't get to see the top. but i'm still trying to keep track of the numbers as best i can. i should get to see it at the top saturday or sunday, depending on work.

 

[defed]

i think i missed the absolute top, but my cells range from 3.49v to 3.36v shortly after going to float. i presume it's balancing in 'idle'. i did bump the bulk/absorb up to 3.5v just to make sure i got them past the 3.45v required to balance.

 

[defed]

today, i saw it close to the top....55.8v (have it set to 56v). cell voltage ranges are 3383mv-3588mv. Dexter stated that it only balances while idle or charging, not while discharging. does that mean i shouldn't put a load on them until the cell voltages level out?

by the time it balances, it will probably be around 54v, which means it should start to recharge the next cycle. i'm not sure, but i think if it gets low enough it will get a charge the same day. do i just let it do its thing a few days or should i draw it down some once it levels out and then let it start again, then repeat?