Two 12V LiFePo4 batteries in Series one shows Reverse Polarity

[Solgato]

So I have an odd thing happening that I’m trying to get to the bottom of.

I have 2 12V LifePo4 batteries wired in Series for 24V with a Victron shunt and BMV-712.

The batteries are charged via dual Victron 75/15 MPPT charge controllers fed by 600W of panels.

All Victron devices are networked together.

I never fully discharge the batteries which power a Solar electric boat.

The chargers are set to float at 27V and Absorb at 29V.

The problem:

A few times now I have taken the boat out given it throttle only to see the voltage drop significantly in direct relation to throttle input. The motors then go into a protection mode since the voltage has dropped below 20V, a Limp Mode of sorts if you will. During that time the voltage from the solar system (60V) will provide some power to allow the motors to limp along.

In this situation it’s obvious the batteries are not giving power under load, but rather the solar panel voltage.

Testing once I get home, I find that while the batteries are still wired up/connected in Series with the solar chargers turned off, one battery measures 14V and the other -12.4 or so.

If I disconnect the negative terminal that is fed from the shunt of the battery that shows the reverse polarity battery, it measure +13V.

As soon as you hook it back up, it measures negative again.

The shunt and BMV in this scenario don’t function as it must not be getting/measuring 24V.

I forgot to but should have measured across the whole bank to see what it measures (the batteries are in separate hulls).

If I turn the solar chargers back on the Shunt and BMV work and everything appears okay Voltage-wise except that there is no current being reported as moving into or out of the bank as if the Shunt is only able to report voltage.

The best theory I can come up with is that perhaps one battery is disconnecting in between charge cycles due to low voltage, or perhaps during a charge cycle due to over charge, and that is somehow causing an imbalance between the batteries? Or they are drifting apart in capacity? But how would that explain the behavior?

I don’t know, I’m really stumped on this one.

The last time it happened I disconnected the batteries and charged them independently using a straight charger until they reached an equal state of charge and resting voltage. Seemed okay after that and was for a number of months, but maybe there is some other issue at play.

I guess the next step is to capacity check each battery.

Both were purchased new at the same time and are about a year and a half old, only used in this boat from day one which sits mostly in the yard until I take it out weekly. During the night there is a very small draw from the BMV and remote linear actuators controllers, and so everyday it goes through a minor charge cycle. I’m hoping now with Victrons ability to Data Log, I’ll be able to see exactly when this phenomena happens although I suspect it would show as a 0V and not negative if it happens at night because the BMV would turn off completely since it wouldn’t be getting 24V.

I know batteries can reverse polarity, but that doesn’t seem to be the issue since it measures correctly when disconnected. Instead it seems to be somehow related to the Series connection or perhaps the Shunt since it’s the battery closest to it.

I didn’t think I would need a Balancer with these with the way they get used, but I may have to install one.

Right now I’m drawing some power out of the 14V batt to get it closer to the other, then I’ll separately charge them again.

Maybe before that I’ll try to get a voltage reading across the whole bank without the charger on if the one battery is still doing its thing.

Anything else I should check while I’m trying to get to the bottom of this issue?

Thanks!

 

[chrisski]

My guess is giving the throttle exceeds the amps for the BMS and it shuts down or the batteries don’t have enough power and have voltage sag and start to drop on voltage when the throttle is depressed.

Im sure you thought of the above.

IMO unless the battery spec. Sheet says so, absorption should not be above 28. 29 starts to have potential for a runaway cell in an unbalanced pack.

The Victron shunt has no ability to record data and will only graph your current reading, and once a different reading is taken, the old graph is not available to compare data. There’s always a rumor of an update to the software to fix this, but I think its a hardware issue.

There is a reading for min and max voltage, but won’t say the date or time of the readding.

 

[Texas-Mark]

Testing once I get home, I find that while the batteries are still wired up/connected in Series with the solar chargers turned off, one battery measures 14V and the other -12.4 or so.

If I disconnect the negative terminal that is fed from the shunt of the battery that shows the reverse polarity battery, it measure +13V.

As soon as you hook it back up, it measures negative again.

Can you please explain how you are measuring things (i.e. with a multi-meter) and where you are putting the leads. A drawing or pic of the setup would help.

 

[Solgato]

My guess is giving the throttle exceeds the amps for the BMS and it shuts down or the batteries don’t have enough power and have voltage sag and start to drop on voltage when the throttle is depressed.

Im sure you thought of the above.

IMO unless the battery spec. Sheet says so, absorption should not be above 28. 29 starts to have potential for a runaway cell in an unbalanced pack.

The Victron shunt has no ability to record data and will only graph your current reading, and once a different reading is taken, the old graph is not available to compare data. There’s always a rumor of an update to the software to fix this, but I think its a hardware issue.

There is a reading for min and max voltage, but won’t say the date or time of the readding.

Thanks for your response and input.

Boat performs flawlessly otherwise. Can pull 120A out of them without any trouble and they charge back up quickly when at rest. I had the batteries built to spec with the BMS’s built/programmed to handle max continuous amp draw of the motors under load. All wiring and switches are correctly sized. No other issues except for this. Every week or so I look at the data in Victron connect and all values are very consistent day to day week to week except for when I take the boat out of course.

With regard to settings, it is my understanding that you need to push the cells and hold them there for a period of time so they can do an intra-cellular balance. This typically is in the 28-29V range. With the batteries being 14.6V and wired in Series, I figured 29V would be okay. I like to float them at a lower V since I don’t use them regularly and am willing to give up a bit of capacity at the start of an outing. I would however prefer if there was a feature/ability to set a ‘balance schedule” where say once a month absorption went to 29V, and then on a daily basis it say went to 28V.

I do however agree that one of the batteries may be disconnecting due to a Absorbtion setting of 29V if the other battery is falling behind or reaching capacity more slowly, and that it may be the cause, but I’m struggling to understand how that would create the strange polarity issue.

The latest firmware update logs data on Victron Smart devices. You can view any number of variables by the minute, hour, day, week, etc. Stoked they added it as I had requested it some time ago. The min and max voltage is useless in situations like this since the BMV shuts down without power, reporting 0 as a value for Min. The new update also lets you see real time reported data of multiple devices without pairing directly with one device in particular.

The Data Logging is called Stored Trends. You can read more about it on Victrons Change Log page:

 

[TomC4306]

I know batteries can reverse polarity

Never heard of such a thing.

 

[Solgato]

Can you please explain how you are measuring things (i.e. with a multi-meter) and where you are putting the leads. A drawing or pic of the setup would help.

Sure. Batteries are wired in Series with the shunt near one battery and the power wire for the shunt pulling from the other battery. Have correct 1A fuse inline for BMV.

With the chargers off but the batteries still wired up as a Series bank with shunt, if I use a multimeter to measure the battery furthest from the shunt independently (plus to +/minus to -) I get 14V, if I measure the other (plus to +/minus to -) near the shunt I get a negative voltage of 12.4V or so.

If I then disconnect the negative cable that attached to that batteries negative terminal which then runs to the shunt, breaking Serial connection, and measure the battery independently, it measures 13V using the multimeter.

Although the batteries are far apart, all wires are of equal gauge and length, and there are no other loads connected.

The batteries and MPPT and motors are all tied together with a DPDT master switch that allows the MPPT solar to flow back to batteries at all times regardless of whether motors are powered up or not. Basically using the poles of the DPDT switch as Bus Bars.

Please excuse my chicken scratch, but it should be good enough to give you a basic idea of the system and how it is wired.

 

[Solgato]

Never heard of such a thing.

It’s rare, but it does happen, at least with FLA. If you drain a battery completely you can charge it back up in reverse. Once empty it doesn’t really care and will take a charge, but will be damaged.

That’s not the issue I have, but the BMS’s work differently in different batteries. Some batteries have to be completely physically disconnected and removed from the bank after a BMS fail safe disconnect or even “jump started” before they will open up again and show/provide voltage.

Lots of possibilities when you have individual internal BMS protection built into batteries that comprise a bank.

It’s possible of course that one battery has a dead short or a bad cell.

Unfortunately I can’t monitor my cells individually. Best I can do is a capacity load test of each after fully charging.

 

[Solgato]

So after some more testing I think the problem may be related to one battery, although I don’t understand how it would cause the strange voltage readings. If in fact the issue is due to a severe imbalance, internal BMS, cell, or short failure, it would be nice to know that this is the kind of behavior one might experience.

Here are the tests I performed tonight in order step by step with results:

1. Measured each batteries resting voltage. Chargers off. Disconnected from Series connection. Battery #1 = 13.35V after leaving an LED overhead light attached for a few hours. Battery #2 (closest to Shunt) = 13.16V

2. Make serial connection for 24V. Battery #1= 13.2V. Battery #2= -8.79V

3. Measure from either side of shunt to positive post of Battery #2. Both sides show -8.85V

4. Measure across entire 24V bank while wired in Series (one batts + to other batts -) and it measures 4.1V

5. Turn MPPT charger on. Charger goes to Bulk and BMV now Powers up because it has 24V and Shunt/BMV reports 27V or so as bank V.

6. Measure shunt side battery #2 individually while wired up in Series and receiving solar charge and it measures 14.8V.

7. Measure Battery #1 and it is 13.4V

8. Turn charger off and V drops to 4V or so and BMV shuts off because it no longer has minimum Voltage to work

9. Turn charger back on and shunt side Battery #2 goes to 15V while Batt #1 holds at 13.4V

10. Turn charger off and totally disconnect each battery (both posts) and Batt #1 has resting V of 13.8V and the shunt side #2 Batt measures 13.15V.

11. Attach multi LED dome light to Batt #1 and it powers it fine.

12. Attach multi LED dome to Batt #2 and it only lights up one or two surface mounted LED’s. Measure the voltage under load and it’s only 6.5V.

13. Attach a charger for to Batt #2 and voltage goes to 14.4V but no Amps flowing.

Conclusion: Something is wrong with Batt #2!

Next step: BMS is supposed to LVD at 10V and obviously didn’t since batt is still powering LED light at 6.5V under load. Will leave on charger over night and see what happens.

Thoughts after reviewing voltages: if you subtract Batt #2 resting voltage in step 2 test above from Batt #1 resting voltage, you get step 4’s result of 4.1V for the overall bank voltage.

My guess is Batt #1 is trying to charge Batt #2 when wired in Series and is producing a negative voltage reading. I know this can sometimes happen although it’s rare in Series connected cells when one is severely depleted and the other full. The full battery can reverse the polarity of the depleted battery as it forces its current through it drawing it even further down, but I find it hard to believe the BMS isn’t intelligent enough to protect this from happening. So my assumption is I could have a faulty BMS and that in turn has damaged the battery, especially given that it will deliver power well under it 10V cutoff.

On a side note, I see now that I may have mistakenly started this thread in the wrong topic category. I misunderstood the Thread Title relating to over the counter LifePo4 sealed battery products. My apologies.

 

[mikefitz]

It seems at first that the batteries are at different states of charge, battery 1 at 13.35 , SOC = 95%, battery 2 at 3.16 = 50% , however when loaded battery 2 volts fail, suggests issues with battery 2
Some of the other tests make no sense, batteries cannot go negative, perhaps a measurement error. Battery 2 has the BMS in protection mode, preventing current flow and giving odd voltage measurements. Try charging as a 12v battery and carrying out a load and capacity test as a 12v battery. As
Before reconnecting the batteries in series, assuming both batteries are functioning correctly, charge as 12v batteries to full.

After reading all you comments it appears this issue can be solved by charging each battery separately as a 12v battery.
Overtime its possible the batteries are developing imbalance, perhaps due to the long cable joining the batteries or batteries with different performance characteristics

 

[TacomaJoe]

If you draw a couple of batteries in series, add a load resister and measure the voltages is makes sense. Now mentally replace a battery with a high resistance (BMS open) and the voltage measured on the terminals of the open battery will be equal to voltage of the good battery but the polarity is reversed.

 

[TomC4306]

Ok. But this isn't the actual battery changing polarity.

 

[Zwy]

So I have an odd thing happening that I’m trying to get to the bottom of.

I have 2 12V LifePo4 batteries wired in Series for 24V with a Victron shunt and BMV-712.

2 BMS's? Why not just purchase a 24V battery or build one?

The batteries are charged via dual Victron 75/15 MPPT charge controllers fed by 600W of panels.

All Victron devices are networked together.

I never fully discharge the batteries which power a Solar electric boat.

The chargers are set to float at 27V and Absorb at 29V.

I wouldn't run Absorb, just bulk to less than 29V.

The problem:

A few times now I have taken the boat out given it throttle only to see the voltage drop significantly in direct relation to throttle input. The motors then go into a protection mode since the voltage has dropped below 20V, a Limp Mode of sorts if you will. During that time the voltage from the solar system (60V) will provide some power to allow the motors to limp along.

In this situation it’s obvious the batteries are not giving power under load, but rather the solar panel voltage.

Testing once I get home, I find that while the batteries are still wired up/connected in Series with the solar chargers turned off, one battery measures 14V and the other -12.4 or so.

-12.4V

If I disconnect the negative terminal that is fed from the shunt of the battery that shows the reverse polarity battery, it measure +13V.

Given the negative voltage above, my guess is the one battery has a bad cell/s and pulls power from the other battery when the 2 are connected in series.

As soon as you hook it back up, it measures negative again.

The shunt and BMV in this scenario don’t function as it must not be getting/measuring 24V.

I forgot to but should have measured across the whole bank to see what it measures (the batteries are in separate hulls).

If I turn the solar chargers back on the Shunt and BMV work and everything appears okay Voltage-wise except that there is no current being reported as moving into or out of the bank as if the Shunt is only able to report voltage.

The best theory I can come up with is that perhaps one battery is disconnecting in between charge cycles due to low voltage, or perhaps during a charge cycle due to over charge, and that is somehow causing an imbalance between the batteries? Or they are drifting apart in capacity? But how would that explain the behavior?

I don’t know, I’m really stumped on this one.

The last time it happened I disconnected the batteries and charged them independently using a straight charger until they reached an equal state of charge and resting voltage. Seemed okay after that and was for a number of months, but maybe there is some other issue at play.

I guess the next step is to capacity check each battery.

Both were purchased new at the same time and are about a year and a half old, only used in this boat from day one which sits mostly in the yard until I take it out weekly. During the night there is a very small draw from the BMV and remote linear actuators controllers, and so everyday it goes through a minor charge cycle. I’m hoping now with Victrons ability to Data Log, I’ll be able to see exactly when this phenomena happens although I suspect it would show as a 0V and not negative if it happens at night because the BMV would turn off completely since it wouldn’t be getting 24V.

I know batteries can reverse polarity, but that doesn’t seem to be the issue since it measures correctly when disconnected. Instead it seems to be somehow related to the Series connection or perhaps the Shunt since it’s the battery closest to it.

I didn’t think I would need a Balancer with these with the way they get used, but I may have to install one.

Right now I’m drawing some power out of the 14V batt to get it closer to the other, then I’ll separately charge them again.

Maybe before that I’ll try to get a voltage reading across the whole bank without the charger on if the one battery is still doing its thing.

Anything else I should check while I’m trying to get to the bottom of this issue?

Thanks!

 

[Solgato]

Ok. But this isn't the actual battery changing polarity.

Right. That’s what I stated twice. It’s not the problem as the battery measures correctly on its own when disconnected.

Batteries also leave the factory sometimes with reverse polarity BTW. You might not believe it until you’ve experience it first hand.

 

[Texas-Mark]

Batteries also leave the factory sometimes with reverse polarity BTW. You might not believe it until you’ve experience it first hand.

Say what?

 

[Solgato]

If you draw a couple of batteries in series, add a load resister and measure the voltages is makes sense. Now mentally replace a battery with a high resistance (BMS open) and the voltage measured on the terminals of the open battery will be equal to voltage of the good battery but the polarity is reversed.

Thanks TacomaJoe for taking a stab at trying to theorize what might be happening that is causing the strange voltage readings/behavior.

Can you explain further your theory?

I would think others might run into a similar issue with similar troubleshooting results if this is the case since there are plenty of people using 12V batteries in Series with shunts and no balancer between them.

I have been looking into the behavior of batteries with BMS’s when they are in LVD or HVD mode and some documentation suggests a battery can still measure up to 6V across the posts when disconnected.

I had assumed it would read 0V and that one would not be able to pull any energy out of it since a disconnection is designed to protect the battery from being damaged by further discharging or charging.

Perhaps this is partly what is causing the strange Voltage readings.

BTW, I do know how to use a Voltmeter and the readings are accurate. I am in the electronics business.

The strange readings and behavior is why I came here asking questions because it is very odd and I have not been able to find anyone else that has experienced such a thing.

 

[Solgato]

Update:

I left the battery on an old Schumacher charger over night. Woke up this morning, physically inspected battery and no heat or bulging.

Battery was not drawing current.

Removed charger and battery measured 14.2V.

Attached 12V LED dome light and battery dropped to 6.65V under load only lighting up a couple surface mounted LED’s.

Next step, going to try to jump start battery using the other fully charged battery and see if the BMS will wake up.

As I mentioned in my last post, I didn’t expect to be able measure any voltage or draw any current from a battery with a disconnected BMS, but have found documentation that suggest one might at least get a voltage reading. And so I think trying to jump it using a Parallel connection is the next logical thing to try.

I will also check (per mikefitz’s comments) the resistance of all my wire run while everything is disconnected to make sure I don’t have any breaks or poor connection points causing a resistance deviation between batteries with regard to discharging and charging and the two drifting apart in capacity.

 

[Texas-Mark]

Is this battery you can access the cells? Just wondering what the cell voltages are after the BMS opens. Sort of sounding like a bad BMS.

 

[Solgato]

Is this battery you can access the cells? Just wondering what the cell voltages are after the BMS opens. Sort of sounding like a bad BMS.

Agreed. Unfortunately not easily accessible although I will likely have to take a crack at a DIY repair based on their age and my location with regard to shipping logistics.

Maybe I could at least get the manufacturer EWT to supply a replacement BMS if the cells check out.

A few more tests/attempts to get it working and then I’ll probably have to open it up. I got it working again once before (the only other time this happened) about 6 months ago.

I’m kicking myself now for not adding a Balancer after it last happened as I had planned and intended to, but got side tracked designing my scaled up (30’) version of this boat.

 

[Solgato]

Okay. Crossing my fingers here. Just jumped the battery using parallel connection to the other charged battery. Voltmeter showed +13.4V on Batt #2 (misbehaving battery) and -13.4V on good battery. Within 5 seconds both batteries showed a +V voltage.

Left the connection for a few minutes, removed jumper cables and Batt #1 (good) measured 13.5V at rest and Batt #2 measured 13.52V.

Applied 12V LED dome light to Batt #2 to apply a load, and the light fully lit up and V remained steady at 13.52V.

Put Batt #2 on charger, and it started drawing just under 2A with a voltage of 13.6V.

So at this point it appears the BMS had disconnected in Batt #2 and needed to be woken up.

Still would like to understand the strange behavior and document it incase others experience something similar.

I’m not in the clear yet of course as I could have a faulty BMS or bad cell(s), but I’ll separately charge each battery and lower my Absorbtion V a bit and if everything behaves and measures normal after a few daily solar charge cycles once wired back up, I’ll install a Battery Balancer and hope that the two incidents I’ve had didn’t cause any permanent damage.

 

[Travis]

If you draw a couple of batteries in series, add a load resister and measure the voltages is makes sense. Now mentally replace a battery with a high resistance (BMS open) and the voltage measured on the terminals of the open battery will be equal to voltage of the good battery but the polarity is reversed.

Yes I believe you are correct. One of the batteries is open circuit and that's why he's getting the negative result. Now why is it open circuit? Is the BMS failing or is it in protection mode? Just as a side note Some bms's can protect one way as in if they're overcharged it will allow discharge but not any charge or if they're undercharged it'll allow charge but not discharge.