Needing some advice on a system a family member has installed.

System is 2 x Sacolar 5kw inverters (Growatt SPF5000ES) with 4 x TheSunPays 4.8kw LiFePO4 batteries.

Batteries were purchased in January 2023 with 750 cycles on them. For the last couple of months, the batteries have had sudden drops in SOC.
In July 2024, one of the batteries suddenly dropped for 60% SOC to 15% SOC which can be seen here at the total SOC dropping from 59% to 47% just before 18:00

We contacted their support and upped the voltages which brought everything back into sync

Since then, we continuously had problems of the batteries SOC dropping, again in July 2025 multiple times.

Again, we contacted TSP and this time they connected remotely and updated the batteries firmware.

Then in August, the system shut down entirely. Dropping from 21% SOC to 0%. Most notable the battery voltage was sitting at 45.3v at 20%. It would normally be at 48.8v at 20% SOC.

Again, reported to TSP, they connected, and said everything on the batteries is fine.

Again, it happened on 26 August 2025, except this time failing from 44% SOC to 0%. Voltage at 44% was 45.3%

In this failure, the 1 inverter also failed. We sent it in, they charged us R3400 for a new main board even though I protested saying the failure was due to the continuous battery failures.
Their response was that the inverter is out of warranty.

Nevertheless, we paid and a week later we had the inverter installed.

And yet again, 18 September, it happens again

Batteries drop from 41% SOC to 0%, voltage 45.3v.

Now considering these constant failures, and that we've reported it each time, and in each instance, their support cant see any issue with the batteries whatsoever, I'm strongly of the opinion that these batteries are duds.

They seem to have a workable 55% SOH even though the reported is 100%.

How should I proceed here? Should I send all 4 batteries in and insist on replacements or a refund? They have a 5 year or 3500 cycle warranty.

sounds like some of the cells may be out of balance.
those batteries mostly use pace bmses so try getting the pacebms software and check the settings, they may need adjusting to help you bring all the cells back in to a balanced state.

How long do you go without the batteries reaching 100% soc fully charged?

If its quite a few days, this can happen, imbalance. SoC value is an estimate the inverter makes, but if not charged fully often enough then it is not accurate.

2 hours ago, Bl4d3 said:

Now considering these constant failures, and that we've reported it each time, and in each instance, their support cant see any issue with the batteries whatsoever, I'm strongly of the opinion that these batteries are duds.

Most probably there's just one or two weak cells in one of the four batteries. Even if you would be successful with returning them, the chances are that the new batteries will develop quite the same behaviour after a year or two of usage. That's because the guts of these batteries are always the same: 16pcs of prismatic cells + OEM metal enclosure + PACE/DALY/JK BMS.

2 hours ago, Bl4d3 said:

How should I proceed here? Should I send all 4 batteries in and insist on replacements or a refund? They have a 5 year or 3500 cycle warranty.

IMHO, identify which one of the four batteries is the "lemon" and then ask TSP for swapping THIS ONE for a new (or refurbished) unit. If TSP refuse to take action, then just forget about the (non-existent) warranty and go this way:

1) Find an electronics repair shop and let them swap the weak cell(s) for the new one(s).
2) One 105Ah cell costs less than 30 USD now (here in EU, not sure about SA costs).
3) Weak cells are those, that will drop their voltage sharply at the same time when BMS reports drop in SOC.

As an alternative before swapping the cell, you can let repair shop to try this method, which I already saw working once:

• Identify weak cell first

• Charge whole battery to 100%

• Then charge weak cell via bench power supply to 3,5V manually

• Cycle battery SOC 100-20-100% two times to check whether the cell is stabilised or not

3 hours ago, abd7 said:

How long do you go without the batteries reaching 100% soc fully charged?

If its quite a few days, this can happen, imbalance. SoC value is an estimate the inverter makes, but if not charged fully often enough then it is not accurate.

They are charged fully most days.

2 hours ago, Youda said:

Most probably there's just one or two weak cells in one of the four batteries. Even if you would be successful with returning them, the chances are that the new batteries will develop quite the same behaviour after a year or two of usage. That's because the guts of these batteries are always the same: 16pcs of prismatic cells + OEM metal enclosure + PACE/DALY/JK BMS.

IMHO, identify which one of the four batteries is the "lemon" and then ask TSP for swapping THIS ONE for a new (or refurbished) unit. If TSP refuse to take action, then just forget about the (non-existent) warranty and go this way:

1) Find an electronics repair shop and let them swap the weak cell(s) for the new one(s).
2) One 105Ah cell costs less than 30 USD now (here in EU, not sure about SA costs).
3) Weak cells are those, that will drop their voltage sharply at the same time when BMS reports drop in SOC.

As an alternative before swapping the cell, you can let repair shop to try this method, which I already saw working once:

• Identify weak cell first

• Charge whole battery to 100%

• Then charge weak cell via bench power supply to 3,5V manually

• Cycle battery SOC 100-20-100% two times to check whether the cell is stabilised or not

Would a single cell in 1 pack cause this problem across all packs? As all of them would suddenly drop in capacity. There was only 1 instance where 1 pack dropped a year ago, but since then, all 4 packs drop to 0% and shut down due to low voltage.

Are you monitoring SOC of the each individual pack in the realtime?

Because if you are reading SOC via inverter and all of the packs have their comm ports daisy-chained, then the total SOC is calculated as an average. Therefore, if one pack falls to zero, total SOC falls proportionally downwards.

It's very unlikely that all four batteries would have a bad cell and that SOC of all four woud fall down within the exact same moment. If you want to be sure, just connect one battery at a time and then charge-discharge in order to check its behavior. (Keep in mind not to go over C-rating while charging discharging).

8 hours ago, Bl4d3 said:

4 x TheSunPays 4.8kw LiFePO4 batteries.

What is would do. Disconnect all 4 batteries and run them individually as @Youda mentioned. If possible do a capacity test on each battery discharge to bms low voltage cutoff( normally 10% dod. Then individually charge to 100 SOC but configure the absorption and float at the same voltage for maximum balance. Should 1 pack go into overvoltage discconect you have a cell they call a runner empty first charged first. If you have a method of checking cell balancing it would make the process easier. By discharging to bms cutoff you reset Coulumbic Efficiency / Battery Status Meter Efficiency for LFP = 99%.
This get's the bank charged to full with high amps (Constant Current) and then float (Constant Voltage) tops off so the cells are on average between 3.400-3.420. I am running 24/7/365 so float is used up by the Inverter + provides whatever the packs will take to top off.
* Do Not forget to adjust for Voltage Offsets between Actual Voltage @ Battery Terminal & at Solar Controller.
AGAIN: Calibration is VERY Important !
This is a step many miss and can't understand why they have issues at the Top of the SOC Level and at the bottom (most notably inverter cutoff at wrong Batt Terminal Voltage.)

2 hours ago, Youda said:

Are you monitoring SOC of the each individual pack in the realtime?

Because if you are reading SOC via inverter and all of the packs have their comm ports daisy-chained, then the total SOC is calculated as an average. Therefore, if one pack falls to zero, total SOC falls proportionally downwards.

It's very unlikely that all four batteries would have a bad cell and that SOC of all four woud fall down within the exact same moment. If you want to be sure, just connect one battery at a time and then charge-discharge in order to check its behavior. (Keep in mind not to go over C-rating while charging discharging).

Yes. Monitoring each battery, and all 4 fail simultaneously due to low voltage

The packs all appear fine, voltage is actually not bad in this instance.

While charging on Thursday, there was some bike imbalances though

5 minutes ago, Bl4d3 said:

Yes. Monitoring each battery, and all 4 fail simultaneously due to low voltage

If that is really true, and all the batteries, while being discharged, do report undervoltage roughly at the same time, then the most probable reason is the wrong capacity being set in the firmware.

I saw that in the past already: the actual cells were 100Ah, but in the BMS 120Ah was set by a mistake. Sharp SOC drop almost every morning, because of undervoltage. Sharp SOC bump when charging, because batteries were full way before BMS expected that.

Not sure whether this is your issue, but if you are monitoring each battery individually and they behave very same then it might be worth checking.

3 minutes ago, Youda said:

If that is really true, and all the batteries, while being discharged, do report undervoltage roughly at the same time, then the most probable reason is the wrong capacity being set in the firmware.

I saw that in the past already: the actual cells were 100Ah, but in the BMS 120Ah was set by a mistake. Sharp SOC drop almost every morning, because of undervoltage. Sharp SOC bump when charging, because batteries were full way before BMS expected that.

Not sure whether this is your issue, but if you are monitoring each battery individually and they behave very same then it might be worth checking.

They actually behave exactly as you described, however I can confirm that the BMS is set to 100Ah as the TSP tech set all of them to 100Ah on the last update they did. They adjusted them down as each one was at roughly 105Ah

1 hour ago, Bl4d3 said:

The packs all appear fine, voltage is actually not bad in this instance.

While charging on Thursday, there was some bike imbalances though

those images you shared show the batteries at midpoint in the flat area of the charge cycles, you will not easily see bad cells that way, you should rather check at the top end of a charge cycle to try pinpoint bad cells. and check each cells voltage to see/find runaway cells.

2 hours ago, Bl4d3 said:

The packs all appear fine, voltage is actually not bad in this instance.

While charging on Thursday, there was some bike imbalances though

Do you have these pictures at the time that the batteries report 0% SOC? I am assuming that the BMS of each battery stays powered on, but the discharge mosfets are closed due to some protection being triggered.

Can you overlay the SOC graphs with the graphs showing the load placed on the batteries.

Are the SA readings taking directly from the master battery RS232 or RS485 ports ? If they are, you can use the SA cable to read the data directly into PBMS tools or PBmodbus tools depending on which port you are using.

Based on your earlier posts the size of the SOC drops seems to be getting larger (10%, 20% and then 40%) which to me indicates cell imbalances that are progressively getting worse. I am not sure if the BMS will turn off the entire bank if one battery reaches and UVP state.

Edited September 21, 2025Sep 21 by I84RiS

11 hours ago, I84RiS said:

Do you have these pictures at the time that the batteries report 0% SOC? I am assuming that the BMS of each battery stays powered on, but the discharge mosfets are closed due to some protection being triggered.

Can you overlay the SOC graphs with the graphs showing the load placed on the batteries.

Are the SA readings taking directly from the master battery RS232 or RS485 ports ? If they are, you can use the SA cable to read the data directly into PBMS tools or PBmodbus tools depending on which port you are using.

Based on your earlier posts the size of the SOC drops seems to be getting larger (10%, 20% and then 40%) which to me indicates cell imbalances that are progressively getting worse. I am not sure if the BMS will turn off the entire bank if one battery reaches and UVP state.

Sadly not. But when it happens again, I will grab a screen shot.

Good news is, today the batteries seem to have acted normal. Discharged down to 26% over night and voltage only dropped to 48.2v.

Also mind posting the cell voltages on all of them at 100% soc

Really would like to see them all above 3.45v to see the imbalance on the top side.