diy solar

diy solar

Are my cells knackered?

andyP

New Member
Joined
Aug 31, 2020
Messages
54
I have a 4s4p bank of 120ah cells, so 480ah @12v in total.

I've noticed that when I apply a high current load (of Circa 150amps) for a couple of minutes, (2000w kettle) the battery voltage drops from 14.4 to about 12.9v. when the load switches off the voltage only climbs up to about 13.6.

2min at 150a is only 5ah. Out of 480. The soc meter from the dalgreen BMS doesn't show any significant drop in SOC. (But i think the soc meter is working properly. By design or by installation?) My system is always around the 40% mark)

Surely removing this much charge shouldn't drop the voltage so dramatically? Or is this normal?

Thanks for your thoughts.
 
I am assuming that you are using Lifepo4 cells.
If so, it could be normal.
Lifepo4 cells naturally drop to to around 3.4v per cell after charging, as the charge gets absorbed deeper in the cell. Some manufacturers actually suggest letting the cells/pack sit for an hour after fully charged, and before doing a capacity test.

3.4v per cell, or 13.6v per a 4S pack is still considered 100% SOC for Lifepo4 chemistry.
LiFe04-soc-voltage-chart.jpg
 
I have a 4s4p bank of 120ah cells, so 480ah @12v in total.

I've noticed that when I apply a high current load (of Circa 150amps) for a couple of minutes, (2000w kettle) the battery voltage drops from 14.4 to about 12.9v. when the load switches off the voltage only climbs up to about 13.6.

2min at 150a is only 5ah. Out of 480. The soc meter from the dalgreen BMS doesn't show any significant drop in SOC. (But i think the soc meter is working properly. By design or by installation?) My system is always around the 40% mark)

Surely removing this much charge shouldn't drop the voltage so dramatically? Or is this normal?

Thanks for your thoughts.
If your kettle actually pulls 2000 ac watts...
2000 ac watts / .85 conversion factor / 12 volts low cutoff = 196.078431373 dc amps.

The gauge of the wiring and the quality of the joinery really matters all along the current path at ~200 amps.
So does the length of the wire.
Pulling that hard on the batteries is going to make them droop even in the best case scenario.
Long undersized wires and poor joinery exacerbate the effect.
What gauge wires are using between battery and inverter?
 
Last edited:
After applying that load, check the temperature of every connection and cable. If you find a location that is hotter than all the other locations, you've found your problem.

Where are you measuring the voltage drop? You have four batteries. Did you measure at every battery?
 
When I read this, thought the OP had FLA batteries, but I see there’s a BMS. The description is just like pulling a load on an older FLA battery bank, or perhaps under wired,
 
Thanks for the replies.
Lifepo cells.
The system is correctly wired for 200amps. No hot connections. The voltage is being measured by the bms at the battery terminals. And this is a question about the internals of the battery so in a way it wouldn't matter if it was badly wired.
The Cell banks are pretty well balanced. Typically a few mV difference (never more than 10mV}, but under heavy load the balance can be out by 50mv.

Is there a method to check each cell in a working pack. We live off it so can't be disconnected for too long.
 
Bad wiring can create voltage issues. No doubt about that.

I suggest that you get a digital volt meter and check the voltage of every cell before and after you run your high amp load.

You say that your system is normally around 40%. Why is that? Some devices need to see a full charge to reset the state of charge value, which can become out of sync.
 
I am assuming that you are using Lifepo4 cells.
If so, it could be normal.
Lifepo4 cells naturally drop to to around 3.4v per cell after charging, as the charge gets absorbed deeper in the cell. Some manufacturers actually suggest letting the cells/pack sit for an hour after fully charged, and before doing a capacity test.

3.4v per cell, or 13.6v per a 4S pack is still considered 100% SOC for Lifepo4 chemistry.
View attachment 73519
Thanks @meetyg. Very helpful. is this voltage based approach likely to be more reliable than my BMSs coulomb based soc measurement over time?
 
Thanks @meetyg. Very helpful. is this voltage based approach likely to be more reliable than my BMSs coulomb based soc measurement over time?
Probably not, because there is very little difference in voltage between say 20% SOC and 80% SOC. The voltage curve for Lifepo4 is fairly flat for most of the SOC (except for the very start/end).
The BMS should be able to measure SOC more reliably.

I mentioned this voltage chart, because you said that when loads are off, the voltage rises again to 13.6v which means that very little capacity has been drawn, and you are basically still high in the SOC.
As for why the voltage drops during the load, I think the others have answered well. You might have problems with your connections or wire guage.
To measure voltage accurately, you should check the voltage with a DVM at the source (I.e. the battery itself), not at the load.
That's basically what a BMS is doing: the balance leads are also used to measure cell and pack voltages at the source.
 
You mention it is a 4s4p bank, so I assume all 4 series stings are in parallel at a single positive and negative Busbar?

If yes I'd check the amps and volts (at battery and Busbar) of each series string between the battery and Busbar. Are there imbalances between series stings under load? Are all parallel cables exactly the same length and size? I'm running a 8s3p and there is less than a half amp difference under large loads between the series stings. Same when charging.

Also you say the cells are out of balance under load. Were they top balanced?
 
You mention it is a 4s4p bank, so I assume all 4 series stings are in parallel at a single positive and negative Busbar?

If yes I'd check the amps and volts (at battery and Busbar) of each series string between the battery and Busbar. Are there imbalances between series stings under load? Are all parallel cables exactly the same length and size? I'm running a 8s3p and there is less than a half amp difference under large loads between the series stings. Same when charging.

Also you say the cells are out of balance under load. Were they top balanced?
Not quite, i have 4 120ah cells connected in parallel making a 3.4v 480 amp supercell. 4 of these are then connected in series to make the battery. The BMS measures voltage at each supercell. But not at each individual cell.

I top balanced at installation, about 11 months ago. BMS has balancing function.

Cells usually within 7mV.

The voltage drop i mentioned is observed by the BMS at the battery. I've checked the bus bars between cells and supercells and no lose spots or warmth. That's why I suspected a bad cell.

today I did a boost charge. Usually 14.40v, but I pushed it to 14.5 and the soc meter jumped from 40% to 100. So it was under reading.

While at the top voltage the balance was out by 50mv. Soon redressed as voltage settled.

Might leave it at 14.45v to give the BMS a chance to equalize the cells.
 
Ok thx. That is a 4p4s bank. You might consider a 4s4p or some other config that would give you individual stings. That way you could test more specifically, be able to take 1 more stings off line for repair or cell swap without turning off the whole system, and have redundancy.

I know, more expense but it does give a lot of flexibility.
 
I can see the advantages. Can I do this with one BMS, just measuring one row of cells, and safely assume the other rows are similar?
 
Not quite, i have 4 120ah cells connected in parallel making a 3.4v 480 amp supercell. 4 of these are then connected in series to make the battery. The BMS measures voltage at each supercell. But not at each individual cell.

Yeah I would think this setup would make it very difficult for a bms to balance individual cells
 
today I did a boost charge. Usually 14.40v, but I pushed it to 14.5 and the soc meter jumped from 40% to 100. So it was under reading.

Some devices need to see a certain voltage to determine a "full" state of charge. I run my Victron charge devices at 14.2 volts at this time. I'm still getting a 100% state of charge reading from my Victron shunt as well as each BMS.
 
I can see the advantages. Can I do this with one BMS, just measuring one row of cells, and safely assume the other rows are similar?
This my 8s3p bank with individual 125 amp BMS's and 175amp T Fuses for each string


Partial view of my 8s3p bank (this is 2 of the strings up close)
1126211131ba.jpg

Single string and BMS
1126211131a.jpg
 
Last edited:
Back
Top