Mechanisms that decrease the Lifespan of Lithium-Ion batteries and how to avoid them

[wb9k]

Like I said lets see what you got to show. You keep threatening to show it, play your hand. A123 battery may certainly have something inside. But there is nothing in the docs to indicate so. But even if it does, does not change the fact you kind find dozen just like it that do not have anything other than cells inside the box. At 4S is silly to think you could get any one cell so far off without the pack voltage being way below what any equipment can operate with just 7 to 9 volts on a 12 volt system.

And I was being so nice until you showed up to make trouble.

I rest my case. Goodbye.

[Addendum for the benefit of confused onlookers]

P2250001.JPG

Above is a photo of the BMS pcb for the A123 gen II starter battery. Note that there are no relays here (or anywhere else in the battery), switching instead being done by the two banks of large FETs. Current can be cut off for only charge, only discharge, or both. 550 CCA rated, every unit is tested to deliver that amount of current. Cells are Amp 20's in 4S4P or 4S3P. The main microprocessor holds thousands of lines of code. The product specification is 93 pages long, and includes several state diagrams, some of them quite large and complex. A few features:

1) Low volts cutoff. If pack voltage falls below 10 Volts, discharge current is interrupted until battery is charged back up over 10 Volts. If any cell falls below 0.5 Volts, the pack is permanently disabled, since it is no longer safe to charge. THIS IS A REQUIREMENT FOR EVERY OEM.

2) High volts cutoff. Charge current is shut off if any cell reaches 3.9 Volts. THIS IS A REQUIREMENT FOR EVERY OEM.

3) Cell balancing. All four cell voltages are monitored directly. Above 3.6 Volts, balancers are turned on full for 300 mA of balance current. Below this, balancers are PWM'd for lower current until the cell voltage reaches 3.585 Volts. (Error margin in the voltage measurement dictates a number just below 3.600.) Safety and reliability requirements make this a de facto OEM requirement.

4) Temperature sensing. Two thermistors in the module and two on the pcb provide thermal protection for the cells and FETs respectively. Charge current can be limited at extreme low temps, and the pack can be completely disabled if too hot to guard against permanent failure. Safety and reliability requirements make this a de facto OEM requirement.

5) Communication. In OEM applications, the pack communicates with the car over LIN (Local Interface Network). Alternator output will be adjusted to optimize battery performance based on data streaming from the pack to the car's main controller. Error/service messages for the driver are also supported.

6) Data logging. There is an extensive list of histograms, event counters, and other data capturing tools that provide all kinds of historical data for the unit in case it is needed for troubleshooting or failure analysis. When not in use, the battery goes into a sleep mode, but still wakes up periodically to make sure nothing has gone awry and log data.

There are indeed LFP 12 Volt batteries out there with no electronics in them. Most are sold as motorcycle/racing batteries. They can do this because there are no real requirements for the aftermarket--until you manage to kill a few people. Then, beancounters who recognize the cost of a wrongful death lawsuit, or government regulators step in to stop the irresponsible behavior of the manufacturer/designers. As stated in the thread at ES on Ballistic's motorcycle batteries, this has been a disaster for users of these packs, who are experiencing huge failure rates and a shorter life span than the LA batteries they replaced. I've seen others that have balancers and nothing else inside. There is nothing preventing gross overcharge or damage by overdischarge in those batteries. It's a recipe for abysmal performance or worse.

Our friend SK apparently believes his knowledge is so great that he can divine operational details of very complex systems simply by reading the advertising copy for them, or from media reports that give the most scant technical details. Reality strongly suggests otherwise. You have been warned.

[reed cundiff]

What is mechanism that causes LFP cell degradation in charging below 0 C?

To return to the question of mechanisms that decrease lifespan of LFP.

It has been stated quite often that LFP should not be charged below 0 C. Is this true for the quite low C charge rate of solar. Have not seen this discussed elsewhere.

What is the mechanism that would cause degradation in CALB cells?

[wb9k]

To return to the question of mechanisms that decrease lifespan of LFP.

It has been stated quite often that LFP should not be charged below 0 C. Is this true for the quite low C charge rate of solar. Have not seen this discussed elsewhere.

What is the mechanism that would cause degradation in CALB cells?

The bottom limit is really around 300 mV. A123 states 500 mV for additional safety margin. Below this voltage, the Cu anode plates begin to dissolve. The Cu goes into solution in the electrolyte. It happens even faster when the cell is driven negative. When you recharge such a cell, Cu plates onto the cathodes, forming dendrites that can eventually puncture the separator, causing a short inside the cell. The result is a cell that holds no capacity. Kept in place, this usually results in a hot cell in the stack that increases series impedance of the whole pack. Occasionally, the damaged cell will go out in a "blaze of glory", and that's the worst case you're trying to avoid, and the reason OEM's do not tolerate no protection against this scenario.

At low (discharge) C-rates these limits are actually "harder" than with higher currents where a cell might fall this far under load, but then spring back to 2 Volts (for example) once the load is removed. This is a safer condition than slowly pulling down to, say 200 mV and the cell only bouncing back to 300 mV when the load is released. For two cells that show the same loaded voltage but are loaded differently, the cell with the lighter load is more completely discharged than the heavier loaded cell. Charge rate makes no difference to the outcome in this case. Make sense?


[EDIT] I just reread your post and realize you are talking about temperature, not SOC, sorry. The mechanism for damage from charging at very low temperatures is Li plating permanently onto the cathodes, causing rapid and permanent capacity loss. Slower charge rates ARE safer at low temps, so it may or may not be an issue for you. 0 degrees C isn't really that cold....safe charge rates for A123 LFP don't fall off much until you're down to -10 or colder, falling off really far around -20 and below. By the time you hit -40 degrees C, safe charge current is down to a trickle. Bear in mind that at warmer temps, A123 cells are rated to handle 4C of continuous charge current. Ideally, you should be able to get data from CALB showing what the safe charge rates are as temperature falls. If your solar rig can never source more than the worst-case safe current, then you're OK.

Sorry for the confusion.

dh

[Sunking]

The bottom limit is really around 300 mV. A123 states 500 mV for additional safety margin. Below this voltage, the Cu anode plates begin to dissolve.

So on a 4S. 8S pr even a 16S battery with LVD set to 12, 24, and 48 how would it ever be remotely possible to get a cell down to .5 volts when you set the cut off at 3 vpc? You are are biased manufacture and only have one side, otherwise you are in conflict. Just so everyone is properly warned you are a Shill and came here to just make trouble and generate sales. You would also have to explain whey many EV and gizmo manufactures do not monitor at the cell level or 1SxP. They monitor and balance at 1S to 4S. That tells me and the world BMS is not required on lower voltage systems. If one stops and think about it chew on this. A 12 volt system is 4S, 24 volts is 8S or 4 cells in series at 12 volts and 8 cells in series at 24 volts. If you have an Inverter, it has a Low Voltage Disconnect aka LVD. Default is 10.5 volts or 2.625 vpc. The operational range of a LFP battery is 3.0 to 3.4 volts on 12 volts. Critical voltage on a lithium cell is 2 volts or 8 volts on a 12 volt system. The chances of ever over discharging a 4S, 8S, or even a 16S system is so remote you have a better chance of winning the lottery. Even without a LVD your battery operated equipment would quit working long before you ever get the voltage low enough. Now on something like even a 45S 144 volt EV system a BMS is nice to have and takes some workload off the driver. A difference of 3 or 6 volts (1 or 2 bad cells) can go noticed. That is not the case with 12, 24, and 48 volt systems used in solar. Nor do you take all your cells, cram them into a tin can, and set them out in the sun to bake while you charge and discharge them. So in a properly designed Solar System do you have BMS. You most certainly do have a BMS, it is called the LVD in your Inverter, and Set Point voltage in your Charge Controller. You can add more automation if you wish and afford to, but not necessary with a sound strategy of passive fail safes. If you know little about LFP batteries and do not want to learn anything or just the conveyance or piece of mind a BMS may afford you, get a BMS. But just remember Balance Boards are known to destroy cells. I say get rid of them.

[reed cundiff]

Extreme cold degradation while charging

wb9k

Thanks for response. The solar system on 5th wheel is 1420 W of panels which three panels are in series then both strings are parallel to provide 90 V to MPPT and a 48 V nominal battery suite of 16 CALB cells (Manzanita Micro fabrication and BMS). The capacity is 8.6 kW-hr (48 V nominal and 54.4 actual)

We left rig at son's place two winters when we flew down and spent the winters in Guatemala/Honduras and then Ecuador/Peru. The temperatures went down to a minimum of -20 C and battery suite was not disconnected. Inverter was disconnected since it has a parasitic draw of around 50 W. The other parasitic draws are about 20 W so that the daily discharge would have been around 500 W and perhaps 300 overnight. Early morning charging rates would have been around 200 to 400 W, or about 0.02 C. Son checked on this and battery was fully charged at these low rates by mid-morning We have discerned no loss in capacity.

We have had charging rates of over 1400 W in mid-June and high elevation (2700 m) according to Tri-Star MPPT-45 and Manzanita Micro BMS monitors.

Have always appreciated the responses of Karnak and PNJunction as well.

Reed

[wb9k]

wb9k

Thanks for response. The solar system on 5th wheel is 1420 W of panels which three panels are in series then both strings are parallel to provide 90 V to MPPT and a 48 V nominal battery suite of 16 CALB cells (Manzanita Micro fabrication and BMS). The capacity is 8.6 kW-hr (48 V nominal and 54.4 actual)

We left rig at son's place two winters when we flew down and spent the winters in Guatemala/Honduras and then Ecuador/Peru. The temperatures went down to a minimum of -20 C and battery suite was not disconnected. Inverter was disconnected since it has a parasitic draw of around 50 W. The other parasitic draws are about 20 W so that the daily discharge would have been around 500 W and perhaps 300 overnight. Early morning charging rates would have been around 200 to 400 W, or about 0.02 C. Son checked on this and battery was fully charged at these low rates by mid-morning We have discerned no loss in capacity.

We have had charging rates of over 1400 W in mid-June and high elevation (2700 m) according to Tri-Star MPPT-45 and Manzanita Micro BMS monitors.

Have always appreciated the responses of Karnak and PNJunction as well.

Reed

Sounds like you're safe. .02C is a very low charge rate. I could print the A123 numbers here, but CALB is probably different, so it's tough to make an apples-to-apples comparison, even though both cells in question are LFP. I seem to remember seeing the figures for CALB at one point...may be worth a web search.

I usually like to state charge/discharge/C rates in Amps rather than Watts....keeps things a whole lot simpler.

I'm envious of your extensive travel...happy trails!

dh

[Mike90250]

Gentlemen,
I am pruning this thread, and editing out name calling. STOP IT.

If it's off-topic, it's going away.

You can disagree, but can't claim better safer bigger longer than the other. You can use ENGINEERING TERMS, as it relates to safety and best practices, but it's got to be true and verifiable.

You can say I Feel 300.54 is an unrealistic number. You can say why you feel, but you can't say someone else is stupid, fat, wears army boots, too tall...... Get it ?

[reed cundiff]

db

Thanks for response. Googled temperature limits again and found the information you mentioned :"....02C is a very low charge rate..." under Battery University where it noted that "... According to research papers, the allowable charge rate at –30°C (–22°F) is 0.02C..." Always try to investigate before I reply on a thread. I was a research physicist at Army Research Labs on White Sands Missile Range.

Have placed a remote thermometer (Walmart special at about $15) in battery compartment to monitor temperature. Plan to install two 12 V receptacles for 12 V fans (12 W curiously enough = 1 amp at 12 V) to provide cross ventilation. The 4.0 kW Magnum PSWI inverter is in same large compartment as the battery suite which opens to the front under the king pins of the fifth wheel. It can emit a lot of heat when running the 13,500 BTU a/c off solar panels/battery suite (1750 W). Chris Dunphy of Technomadia candidly discussed the mistakes he made with his installment of LFP 4 years ago. He installed his battery in the same cramped and unventilated compartment as his inverter. He compounded this by having his rig parked on tarmac in Phoenix in summer. His mistakes and successes should be studied by anyone planning to install LFP in an RV.

Plan to install two x 30 W halogen lamps in the front compartment as well to keep the temperature above 0 C on cold nights. We can run such for quite a few hours from the battery suite.

The BMS (which came with the Manzanita fabricated batteries - fairly expensive) keeps the 16 individual cells below 3.4 V and balances them every several seconds so we are not worried about over charging. However, I may be a bit OCD and check the three monitors numerous times during the day: TriStar MPPT-45, Magnum inverter, and Manzanita BMS)

Since we have only plugged into line power once in two years, we do not have the problem of keeping the cells at to high a SOC. We generally wake up with a -2 to 3 kW-hr deficit (60 to 70% SOC).

If things finally get worked out with Mexican customs (totaled rig and tow vehicle two years ago between Orizaba and Puebla in 70 vehicle pileup), we shall go back to Yucatan, Belize and Guatemala and not worry one whit about extreme cold
Reed and Elaine

[PNjunction]

Sounds like you're safe. .02C is a very low charge rate. I could print the A123 numbers here, but CALB is probably different, so it's tough to make an apples-to-apples comparison, even though both cells in question are LFP. I seem to remember seeing the figures for CALB at one point...may be worth a web search.

This following might best be put into another thread, but wanted to grab your attention just in case...

Getting back to degradation itself, I have always wondered if any degradation studies have been done for LiFePo4 prismatics (or A123 cells if you prefer) in regards to the fact that as solar users, our charge controllers use PWM in the "absorb" phase (what little there is when fed by decent current!).

In other words, we don't REALLY use CC/CV, but CC/PWM. Typically the pwm is done at about 300hz or so. If looked at on a waveform, this simply means that our controllers just close the circuit during bulk, but once a setpoint has been reached, instead of CV, pwm is actually used. Ie, the voltage can actually shoot up to 4.5 volts per cell! - BUT of course at 300hz, the averaging takes place.

What I noticed when using both prismatics, and my prized A123's from Braille and Antigravity brand batteries was that unlike CV which stops current when the first cell is fully charged, with pwm, they tend to "drift together" - and not an exact balance. We've covered balance enough, but my main interest was how lifepo4 reacts to pwm, since that is what we use in the field. (be it a low-end pwm controller, or an mppt which uses pwm during absorb too actually).

The big fish run things like Outback or Midnite Solar controllers, where us small fry might use Genasun, or Pb-controlers carefully tweaked (no temp comp etc) for voltage - still pwm is present.

Maybe some of the EV chargers ARE using pwm - I don't know. Just wondering if we are hurting or helping by not using a more linear approach to CC's with lifepo4 in solar applications....

[PNjunction]

What I noticed when using both prismatics, and my prized A123's from Braille and Antigravity brand batteries was that unlike CV which stops current when the first cell is fully charged, with pwm, they tend to "drift together" - and not an exact balance.

Ack - quoting myself again. Not a good sign.

I guess I might as well go deeper here - the "drift together" is due to the current reaching the lower cells by virtue of the repetetive leading edge of the pwm pulse. My observation is that it *helps*, but of course is dependent on both capacity and internal resistance.

Egads, I really wanted to avoid balance, so maybe just concentrate on wether pwm is a hurt or helpful charging waveform for lifepo4 ...

[Willy T]

wb9k, I think you have the experience and expertise to answer this question. There are only 3 - 4 people that actually have a functioning solar / solar assisted off-grid system here so it's hard to get facts and not opinions.

The one area I have interest in is the saturation phase of charging. Since charging is a leading voltage and the battery voltage is a lagging voltage, how long should the saturation phase be or to what level. If it is measured by say, ending amps ?? The higher the charge rate, the more disparity I see with shunt counted amp hrs returned. By only using a termination voltage of the charge controller there is a undercharge that accumulates by the cycle.

[wb9k]

This following might best be put into another thread, but wanted to grab your attention just in case...

Getting back to degradation itself, I have always wondered if any degradation studies have been done for LiFePo4 prismatics (or A123 cells if you prefer) in regards to the fact that as solar users, our charge controllers use PWM in the "absorb" phase (what little there is when fed by decent current!).

In other words, we don't REALLY use CC/CV, but CC/PWM. Typically the pwm is done at about 300hz or so. If looked at on a waveform, this simply means that our controllers just close the circuit during bulk, but once a setpoint has been reached, instead of CV, pwm is actually used. Ie, the voltage can actually shoot up to 4.5 volts per cell! - BUT of course at 300hz, the averaging takes place.

This is an interesting question. I don't believe I've ever seen data on pwm-based charging, but I have no reason to believe it makes much if any difference at all to the cells. Many automotive applications drain the cells with high-current PWM at frequencies in the kHz range and nobody considers that a problem. Charging should be no different. When you say the voltage can shoot up to 4.5 Volts per cell, I'm curious where exactly in the system you are measuring the number. When we talk about what voltages are safe for cells, we need to take those numbers RIGHT AT THE CELL with a dedicated v sense line. Measurements made right at the charging supply output terminals are pretty useless as there is typically a significant voltage drop between the supply and the cells. If you're really seeing spikes of 4.5 at the cells, I would worry that the voltage setpoint is too high during the pwm phase. Terminal voltage spikes of about 3.8 are around the upper limit for LFP, and you don't want to stay there any length of time--A123 specs a max of 10 seconds at 3.8 Volts. Do you have any screenshots of the voltage measured at the pack terminals during the PWM phase?

I've done some characterization of commercial chargers and diagnostic stations for professional mechanics/dealerships. The output of the chargers runs the gamut from nice smooth DC power to raw rectified AC. It doesn't seem to matter much. Special techniques for "recovering" LA or NiCd cells don't apply to Li at all as far as I know. If a Li cell has been damaged, it can't really be recovered.

What I noticed when using both prismatics, and my prized A123's from Braille and Antigravity brand batteries was that unlike CV which stops current when the first cell is fully charged, with pwm, they tend to "drift together" - and not an exact balance. We've covered balance enough, but my main interest was how lifepo4 reacts to pwm, since that is what we use in the field. (be it a low-end pwm controller, or an mppt which uses pwm during absorb too actually).

What you are really seeing with cells "drifting together" is the expression of different DC resistances of the series cell-busbar combination. The higher the current, the greater the spread. When you switch from CC to CV(or PWM) the main difference as far as the cell is concerned, is that charge current goes down. The voltage rise (vs. rested voltage) across the whole pack decreases, and so do the differences between cell voltages. Less current, less "Peukert-type" losses, less voltage delta. Most chargers I'm familiar with (including large cyclers that mimic drive cycles) will CC until one cell hits 3.6 or just above, depending on the CC current level. At that point, current is usually cut in half until somebody again hits 3.6 Volts. This may happen several times before getting to actual CV, it depends on the charger. The ideal CV phase, in my opinion, sources exactly 3.6V for each series element (assuming LFP), and allows balancers to trim voltages until every cell is at exactly 3.6 Volts, and charge current is very near to zero. Sometimes the CV voltage will be turned on and off (much too slowly to be called "PWM") so that the balancers can continue to trim without having to fight the charge current 100% of the time. This speeds the final balancing process a bit.

The big fish run things like Outback or Midnite Solar controllers, where us small fry might use Genasun, or Pb-controlers carefully tweaked (no temp comp etc) for voltage - still pwm is present.

Maybe some of the EV chargers ARE using pwm - I don't know. Just wondering if we are hurting or helping by not using a more linear approach to CC's with lifepo4 in solar applications....

I guess I might as well go deeper here - the "drift together" is due to the current reaching the lower cells by virtue of the repetetive leading edge of the pwm pulse. My observation is that it *helps*, but of course is dependent on both capacity and internal resistance.

The second to last sentence above concerns me a bit, and SK has made similar comments to the effect of balancers "shunting off power to weaker cells". This is not how things work. Resistive balancers take excess power from a cell and dissipate 100% of it as heat. The energy cannot be sent to other cells without a much more complex switching network, something that is rarely done because of the high hardware cost (and very low cost of power thrown away during balancing). Amps into the pack puts equal amps through each series cell...some cells may waste more of that power than others as heat with or without the help of balancers, but that excess energy is never passed on to other cells in a passive/shunt balancing system. There is no return path to facilitate that transfer. I think this is important to understand.

Hopefully that is a good start toward answering your questions. A screenshot of what is happening right at the pack terminals (or discrete cell terminals within the pack) would enable me to give a more complete answer, but the short answer is that I wouldn't worry much about how exactly the charge current is meted out to your LFP pack. PWM should be just fine, within reasonable limits.

dh

[wb9k]

wb9k, I think you have the experience and expertise to answer this question. There are only 3 - 4 people that actually have a functioning solar / solar assisted off-grid system here so it's hard to get facts and not opinions.

The one area I have interest in is the saturation phase of charging. Since charging is a leading voltage and the battery voltage is a lagging voltage, how long should the saturation phase be or to what level. If it is measured by say, ending amps ?? The higher the charge rate, the more disparity I see with shunt counted amp hrs returned. By only using a termination voltage of the charge controller there is a undercharge that accumulates by the cycle.

Let me know if my reply to PNJunction doesn't answer your question as well....I think it should come close.

dh

[Sunking]

The second to last sentence above concerns me a bit, and SK has made similar comments to the effect of balancers "shunting off power to weaker cells". This is not how things work. Resistive balances take excess power from a cell and dissipate 100% of it as heat.

Very incorrect statement, a Shunt is a Shunt period. Balance boards put a fixed resistance across the cell to shunt a fixed amount of current AROUND the cell to be passed onto lower level cells. That is the whole concept of Balance boards.The Flaw in that is it only bypasses a small fraction of the current flowing through the cell. Typical Balance Boards only shunt anywhere from 50 ma up to 2 amps in some units. .5 to 1 amp is typical. That does no good when the charger is still pushing 10 amps. You bypass 1 amp around the cell, bu tstill left with 9 amps flowing through a full charged cell. Does the bypass resistor burn of power, sure they do Ohm's law still applies. If you have a 1 amp shunt you have roughly a 3 Ohm bypass shunt resistor burning off 3 watts as 1 amp of current flows through it, and that 1 amp is fed to all lower cells. It recombines with the current still flowing through the cell. This is basic electronic and DC principles 101 a student knows.

[wb9k]

Very incorrect statement, a Shunt is a Shunt period. Balance boards put a fixed resistance across the cell to shunt a fixed amount of current AROUND the cell to be passed onto lower level cells. That is the whole concept of Balance boards.The Flaw in that is it only bypasses a small fraction of the current flowing through the cell. Typical Balance Boards only shunt anywhere from 50 ma up to 2 amps in some units. .5 to 1 amp is typical. That does no good when the charger is still pushing 10 amps. You bypass 1 amp around the cell, bu tstill left with 9 amps flowing through a full charged cell. Does the bypass resistor burn of power, sure they do Ohm's law still applies. If you have a 1 amp shunt you have roughly a 3 Ohm bypass shunt resistor burning off 3 watts as 1 amp of current flows through it, and that 1 amp is fed to all lower cells. It recombines with the current still flowing through the cell. This is basic electronic and DC principles 101 a student knows.

The power that isn't burned off in the resistors stays in the cell. This is why you need to cut charger current back during the balance phase--if the balancers can't keep up, the high cell is still getting overcharged. The energy you don't want to keep in the high cell has nowhere to go but the balance resistor, or into the cell itself. The other cells in the series aren't getting any more power than they were before the balancers kicked on.

dh