Suspected faulty battery in AGM bank

[off.the.grid]

I suspect I have a battery thats failed in my bank of AGM batteries, they are configured in 2S2P as a 24V bank.
I have had a victron battery balancer installed and ive noticed fairly recently that the alarm warning light has been coming on in the morning for some time.
I guess the only way to test this is to disconnect them all and measure the voltages?
Any point in performing a load test on the bank by running the kettle or something?
My suspicion is that one battery in the bank has failed. I also have the impression that the bank is not holding up a good charge anymore as the power went out for the first time the other week when the inverter shut off at 50%.

 

[sunshine_eggo]

2S2P means you should be able to measure them all individually unless you've wired the balancer to parallel in between the 12V units (I don't agree with Victron on this scheme).

You should be able to break it down to pure 2S2P with nothing connecting the strings except at the ends and measure each 12V individually, i.e., basically just disconnect the balancer and the middle interconnects as annotated in this 2S3P config:

 

[off.the.grid]

2S2P means you should be able to measure them all individually unless you've wired the balancer to parallel in between the 12V units (I don't agree with Victron on this scheme).

You should be able to break it down to pure 2S2P with nothing connecting the strings except at the ends and measure each 12V individually, i.e., basically just disconnect the balancer and the middle interconnects as annotated in this 2S3P config:

View attachment 224921

Thanks, I will do that.
Yes I wired the balancer as per their instructions and was recently reading a number of threads questioning their method of wiring.
Seen others suggest that 2 balancers are actually required to do a proper job in this situation.
Not too sure if thats really necessary or not on a bank this size, but one thing i also noticed recently was that the balancer wire interconnect linking the centre of each string together has gotten hot at some point and melted the insulation in places which indicates it had a ton of current going through it which is unusual.
Im not too sure when this would have happened, but it would have been fairly recent.
I replaced it with a heavier wire, but it should not be drawing that much current to need anything too significant in wire gauge?
Not too sure what gauge it is, but would have to be at least 18 gauge wire, possibly 16 gauge.

 

[sunshine_eggo]

Seen others suggest that 2 balancers are actually required to do a proper job in this situation.

Yes.

By only paralleling strings at the end, you maintain the ability to individually monitor each battery - simply check the voltage on each battery and compare. Very helpful when troubleshooting.

Not too sure if thats really necessary or not on a bank this size,

The need for balancers depends on battery health and charge patterns more than size.

but one thing i also noticed recently was that the balancer wire interconnect linking the centre of each string together has gotten hot at some point and melted the insulation in places which indicates it had a ton of current going through it which is unusual.

very

Im not too sure when this would have happened, but it would have been fairly recent.
I replaced it with a heavier wire, but it should not be drawing that much current to need anything too significant in wire gauge?

Correct. It further supports your suspicion of a misbehaving battery.

Not too sure what gauge it is, but would have to be at least 18 gauge wire, possibly 16 gauge.

If interconnects are going to be made intra-string, they should be the same gauge as the wires paralleling the batteries at the main (+) and (-).

 

[Bob142]

I suspect I have a battery thats failed in my bank of AGM batteries, they are configured in 2S2P as a 24V bank.
I have had a victron battery balancer installed and ive noticed fairly recently that the alarm warning light has been coming on in the morning for some time.
I guess the only way to test this is to disconnect them all and measure the voltages?
Any point in performing a load test on the bank by running the kettle or something?
My suspicion is that one battery in the bank has failed. I also have the impression that the bank is not holding up a good charge anymore as the power went out for the first time the other week when the inverter shut off at 50%.

I had a setup just like that with 4 Trojan group 27 AGMs and the Victron balancer. Same symptoms. One battery was toast.
Rather than replace it, I reconfigured the three remaining into a 12V system to back up a sump pump and built a native 24V battery (8S 280AH LFP) as the new main battery bank.

Since then I've decided to never put batteries in series. I don't trust that I can keep them properly balanced and maintained.

 

[off.the.grid]

I had a setup just like that with 4 Trojan group 27 AGMs and the Victron balancer. Same symptoms. One battery was toast.
Rather than replace it, I reconfigured the three remaining into a 12V system to back up a sump pump and built a native 24V battery (8S 280AH LFP) as the new main battery bank.

Since then I've decided to never put batteries in series. I don't trust that I can keep them properly balanced and maintained.

Interesting, I do have a spare AGM battery, but if one turns out to be toast, then its not really ideal to put in a new battery among the 3 remaining batteries is it? I would see it causing more balancing issues.
I suspect the melted wire is a result of one side of the bank seriously going out of balance, presumably during charging.
But here is what I dont get, all very fine to balance between each battery, but each cell inside those batteries cant be balanced at all because there is no physical way to connect to the cells inside, surely there must be room for those to go out of balance quite easily too, but how would you know when you are only measuring the voltage across the 6 cells in the battery itself?
Im really put off any lead acid batteries as a result of this and after hearing similar stories with those ive talked to, with that being said lots of people still use AGM batteries, but in the time thats passed since I installed the bank, lithium batteries are way cheaper and actually cost less for the same usable capacity. Im surprised that AGM batteries are still being sold when you look at the price of prismatic cells these days.

 

[sunshine_eggo]

Im surprised that AGM batteries are still being sold when you look at the price of prismatic cells these days.

LFP is absolutely, positively not for everybody. The number of posts we get here about brand new batteries being unable to fully charge or produces voltage spikes that shutdown inverters is high.

AGM is still a better/simpler solution for:

Starter batteries.
Standby applications.
Cold applications.

 

[off.the.grid]

LFP is absolutely, positively not for everybody. The number of posts we get here about brand new batteries being unable to fully charge or produces voltage spikes that shutdown inverters is high.

AGM is still a better/simpler solution for:

Starter batteries.
Standby applications.
Cold applications.

OK, well at the time I went with it because it was the cheapest option and all the system largely is doing is keeping a fridge running.
Seemed like it would suffice.
Interesting what you say about voltage spikes, not sure what would cause that, but should I go to lithium I would not be charging the bank up anywhere near full capacity, which I assume would reduce those chances.
Ive got a pretty good inverter and even though its rated at 24V, its tolerances for charging means its running with up to 30V on a sunny day while boost charging the bank, so if thats not a voltage spike, I dont know what is!
If people are having issues with new batteries failing to charge, I would assume the BMS would likely be having a role to play in that?
Im seriously looking at building a prismatic cell bank, but want to do my research first, I can get 8x300Ah CATL cells for less than i paid for these AGMs, all are A grade, will do it properly and clamp them together, etc.

 

[sunshine_eggo]

OK, well at the time I went with it because it was the cheapest option and all the system largely is doing is keeping a fridge running.
Seemed like it would suffice.

Agree that LFP prices have come down, but most consumers that just want things to work, and they may be frustrated by LFP.

Interesting what you say about voltage spikes, not sure what would cause that, but should I go to lithium I would not be charging the bank up anywhere near full capacity, which I assume would reduce those chances.

Very common for brand new batteries to arrive imbalanced - all brands, all price points. The issue is that they have to be shipped at < 30% SoC for hazmat regs and it's rare that you get your hands on a battery less than 90 days since it was last charged. This means it's almost impossible to guarantee a battery will arrive with the cells top balanced.

What ends up happening is ONE cell hits the cut-off point before the others, and the BMS CUTS the battery out of the charge circuit instantly. The charger is still sending current when this happens, so there is a brief period where the voltage spikes because the only place the current can go is internal to the charger before it clamps down on it.

Ive got a pretty good inverter and even though its rated at 24V, its tolerances for charging means its running with up to 30V on a sunny day while boost charging the bank, so if thats not a voltage spike, I dont know what is!

While most 24V inverters can handle up to 32V, it's very easy for the aforementioned spikes to go notably higher and trigger an inverter shutdown.

If people are having issues with new batteries failing to charge, I would assume the BMS would likely be having a role to play in that?
Im seriously looking at building a prismatic cell bank, but want to do my research first, I can get 8x300Ah CATL cells for less than i paid for these AGMs, all are A grade, will do it properly and clamp them together, etc.

DIY or higher priced batteries with visibility into cell data gives you insight into what's going on and a means to gauge health and troubleshoot issues. Too many of these cheap batteries that are competitive with AGM do not give any visibility to cell voltages. BMS is a black box.

You being willing to DIY your own battery puts you well outside the vast majority of battery users.

 

[off.the.grid]

Agree that LFP prices have come down, but most consumers that just want things to work, and they may be frustrated by LFP.

Very common for brand new batteries to arrive imbalanced - all brands, all price points. The issue is that they have to be shipped at < 30% SoC for hazmat regs and it's rare that you get your hands on a battery less than 90 days since it was last charged. This means it's almost impossible to guarantee a battery will arrive with the cells top balanced.

What ends up happening is ONE cell hits the cut-off point before the others, and the BMS CUTS the battery out of the charge circuit instantly. The charger is still sending current when this happens, so there is a brief period where the voltage spikes because the only place the current can go is internal to the charger before it clamps down on it.

While most 24V inverters can handle up to 32V, it's very easy for the aforementioned spikes to go notably higher and trigger an inverter shutdown.

DIY or higher priced batteries with visibility into cell data gives you insight into what's going on and a means to gauge health and troubleshoot issues. Too many of these cheap batteries that are competitive with AGM do not give any visibility to cell voltages. BMS is a black box.

You being willing to DIY your own battery puts you well outside the vast majority of battery users.

Quite true, most DIY'ers dont have all the ins and outs and not everyone has much electrical background either.
Most without the experience would be better off purchasing a lithium battery with the BMS and cells already assembled and tested, even if there are less savings to be had, they are still reasonably priced.

That makes sense what you are saying about the spikes, looks like an easy thing to avoid.
Ive read a ton about top balancing, straightforward enough procedure but a great deal of people wont even have a suitable power supply to do that themselves, it appears that many DIY'ers are oblivious to it.
Most decent retailers appear to top balance them but its still best to rebalance yourself to make sure everything is good to go and not trust their word on it.

My inverter is a Steca HP compact, see manual here:

https://www.steca.com/frontend/standard/popup_download.php?datei=173/17396_0x0x0x0x0_Steca_HPC_instruction_EN_DE.pdf

Manual says nominal input voltage is up to 34V, and maximum input voltage is quoted at 44V, so I think I would be safe here from spikes.
My only concern is the battery charging voltages(while genset is running) and the manual says not suitable for lithium, but im assuming the right BMS would mitigate any issues there, so I want to make sure I get the right BMS module.

When I have the genset running, ive never seen the voltages go above 29V, which makes me think the equalization voltage on this inverter/charger unit is set to the lowest it can go at 30V

 

[sunshine_eggo]

Quite true, most DIY'ers dont have all the ins and outs and not everyone has much electrical background either.
Most without the experience would be better off purchasing a lithium battery with the BMS and cells already assembled and tested, even if there are less savings to be had, they are still reasonably priced.

Yep.

That makes sense what you are saying about the spikes, looks like an easy thing to avoid.

top balance is the solution. Charging to 3.40-3.45V/cell can mitigate the situation, reduce/eliminate spikes and allow for extended BMS balancing times.

Ive read a ton about top balancing, straightforward enough procedure but a great deal of people wont even have a suitable power supply to do that themselves, it appears that many DIY'ers are oblivious to it.

Not if they use this site... there's a top balancing guide in resources. Very thorough...

Most decent retailers appear to top balance them but its still best to rebalance yourself to make sure everything is good to go and not trust their word on it.

Nope. None do. It's incredibly time consuming and then they have to discharge them to < 30% to legally ship them.

My inverter is a Steca HP compact, see manual here:

https://www.steca.com/frontend/standard/popup_download.php?datei=173/17396_0x0x0x0x0_Steca_HPC_instruction_EN_DE.pdf

Manual says nominal input voltage is up to 34V, and maximum input voltage is quoted at 44V, so I think I would be safe here from spikes.

Impressive. I suspect the 44V is an absolute limit, and the inverter will fault above 34V.

I've seen 12V systems spike into the 20s.

My only concern is the battery charging voltages(while genset is running) and the manual says not suitable for lithium, but im assuming the right BMS would mitigate any issues there, so I want to make sure I get the right BMS module.

Meh. Most chargers can be made to work with LFP.

Per 5.3:

LED 12 to 27.0V float
LED 11 to 28.8V absorption
LED 10, 11, 12, 13 together can set absorption time to 1h.

You should be able to disable equalization.

You should disconnect the CT-35 temperature sensor if present to prevent temperature compensated charging. Your battery will need its own low temp charging protection.

 

[off.the.grid]

Yep.

top balance is the solution. Charging to 3.40-3.45V/cell can mitigate the situation, reduce/eliminate spikes and allow for extended BMS balancing times.

Not if they use this site... there's a top balancing guide in resources. Very thorough...

Nope. None do. It's incredibly time consuming and then they have to discharge them to < 30% to legally ship them.

Impressive. I suspect the 44V is an absolute limit, and the inverter will fault above 34V.

I've seen 12V systems spike into the 20s.

Meh. Most chargers can be made to work with LFP.

Per 5.3:

LED 12 to 27.0V float
LED 11 to 28.8V absorption
LED 10, 11, 12, 13 together can set absorption time to 1h.

You should be able to disable equalization.

You should disconnect the CT-35 temperature sensor if present to prevent temperature compensated charging. Your battery will need its own low temp charging protection.

OK, thats good to know.
3.4V or so appears to be the sweet spot for LiFePO4 cells, if im not mistaken that would take it up to about 90% capacity?

Going by reports here, it seems like retailers such as Docan are top balancing their cells prior to shipping, but I still wouldnt trust that.
Unfortunatley shipping cells internationally to New Zealand is cost prohibitive, would almost work out the same as getting from my local retailer which is about $300 per 300Ah cell in $NZD

Ive been looking at the inverter manual and there is a bit of messing around with it to change the charge voltages, I cant see anywhere to disable equalization, but looks like it can be lowered right down to 20mins which is its default setting.
I think I would still need to rely on the BMS regardless but looks like I can get away with it.

Seems that the JK BMS is the most popular out there, but I could be wrong.

Now speaking of my existing AGM bank, I tested it this morning and one of the batteries does indeed have a dropped cell.
It is not obviously noticeable at first until you put the bank under load.
I could put the kettle on and one battery started dropping to 10V while the other battery in series with it was staying at 13V, the second chain in series were both hanging around close to 12V at the same time.
Its certainly caused the bank to go a fair bit out of balance as a result and will be the reason why the interconnect wire had started to draw current and melt.
The balancer cant keep up with all of this and hence why the alarm light is staying on.
The batteries are still under warranty so should be able to make a claim. I still think im better off to upgrade to lithium regardless though.

 

[sunshine_eggo]

OK, thats good to know.
3.4V or so appears to be the sweet spot for LiFePO4 cells, if im not mistaken that would take it up to about 90% capacity?

No. it depends on the absorption period. I took a 4S battery to 13.6V (3.4V/cell), and it got to 99.7% SoC with a 5 hour absorption. Charging without that absorption could yield as little as 50% SoC:



Trying to charge to a given SoC based on a voltage is essentially impossible without a thorough evaluation of the system, and a quality shunt that could be used as the basis for calculating a target voltage and terminating with a tail current.

Going by reports here, it seems like retailers such as Docan are top balancing their cells prior to shipping, but I still wouldnt trust that.

Unfortunatley shipping cells internationally to New Zealand is cost prohibitive, would almost work out the same as getting from my local retailer which is about $300 per 300Ah cell in $NZD

OUCH!

Ive been looking at the inverter manual and there is a bit of messing around with it to change the charge voltages, I cant see anywhere to disable equalization, but looks like it can be lowered right down to 20mins which is its default setting.

DIP switch E can be turned off to disable equalization.

I think I would still need to rely on the BMS regardless but looks like I can get away with it.

Best to avoid using the BMS to engage protection for routing operations. Always best to operate INSIDE the limits of the BMS such that it only triggers protection when a true limit is succeeded.

Seems that the JK BMS is the most popular out there, but I could be wrong.

Very popular. Active balancing makes maintaining top balance easier.

Now speaking of my existing AGM bank, I tested it this morning and one of the batteries does indeed have a dropped cell.
It is not obviously noticeable at first until you put the bank under load.
I could put the kettle on and one battery started dropping to 10V while the other battery in series with it was staying at 13V, the second chain in series were both hanging around close to 12V at the same time.

Concur.

Its certainly caused the bank to go a fair bit out of balance as a result and will be the reason why the interconnect wire had started to draw current and melt.

Definitely.

The balancer cant keep up with all of this and hence why the alarm light is staying on.
The batteries are still under warranty so should be able to make a claim. I still think im better off to upgrade to lithium regardless though.

NICE!

 

[littleharbor2]

When you have a bad battery in a 2sXp bank the problem is each 2s group are at the bank's voltage. One bad battery will cause the other one to carry the balance of the voltage. I saw this in a friends battery bank which was 2s5p. One 6 volt GC battery was at about 4 volts which was causing the other battery to charge up to over 10 volts. It was literally steaming and fizzing so much that it and all batteries around it were dripping wet with electrolyte. Not a pretty picture. Removed the pair of batteries and they were able to carry on with a 2s4p bank.