Sojourner1
Itinerant
If you just started using your lfp setup I would hope there would be no diminished capacity, especially at 13.6v
Top Balance Questions
- Thread starter xc2
- Start date
MisterSandals
Participation Medalist
Agreed!
But it you look at where the default charge profiles for LiFePO4 are set to, and where most folks charge to (14v+ !!), this is more of an exception to only charge to here. I don't know anybody that only charges to 13.6 (maybe only Ampster, but before he came along recently, i think i was a minority in this thought).
There was a thread a month ago that asked folks what there charge profiles were and quite a few folks floated at 13.8v and some banged their batteries off the ceiling over 14v everyday. I was astonished, still am.
MisterSandals
Participation Medalist
This thread is still open for additions... i'd love to see more here.
diysolarforum.com
Charge Profile Settings - What do you run?
Im trying to gather data around how people have their charge profiles set for different charging sources. There is such vast information out there, and i think itd be good to get real world examples. I suggest copying the following and putting your information in! ----------- Charging Source...
diysolarforum.com
MisterSandals
Participation Medalist
And this continually boggles my mind. The official stance of this website?
12V LiFePO4 Battery w/ BMS:
12V LiFePO4 Battery w/ BMS:
- Absorption: 14.5V
- Float: 13.6V
- Inverter Cut-off: 10.7V-12V (depending on size of load and voltage drop etc)
Design Your Own LiFePO4 Solar Power System
This is a complete system solution complete with: BMS Protection Features (over/under voltage/over-current/balancing) Low Temperature Cut-off Programmable Charge/Discharge Bandwidth Settings Matched...
www.mobile-solarpower.com
Sojourner1
Itinerant
I'm over 4 years now of everyday use with various loads up to 180a everday. Never turned inverter off with a constant load of 2a (25w) this entire time. I've charged to 14.1 or 14.2v (depending on time of year) and "float" everyday at 13.6v with no discernable lose but again there is only 525+ combined full cycles.
Inverter lvd is set at 12v and never has disconnected even at 200a+ load at 25% SOC.
MisterSandals
Participation Medalist
I am seeing this as a recurring theme, that folks that have a continual load have a higher charge and higher float level. @Ped is a full-time RVer i believe and he has an inverter and mini-fridge going 24/7 (i was paying attention to what you were saying!)
My RV sit in the driveway more than i'd like. If i were to float at 13.6, it would be micro-cycling all day long. My cells settle to 3.35v very quickly so my system would start a charge cycle from 13.6v to my max charge voltage several times an hour, all day long, everyday. I have my float value at 13.2v (3.3v) and it may be weeks between charge cycles (2 amp tender to engine battery and small misc drains). I might bump this up higher if i need to maximize battery capacity while traveling for some reason, but i've oversized my battery by design to avoid this scenario.
Ampster
Renewable Energy Hobbyist
Diminished capacity? Are we talking about memory again?
@xc2 I would be interested to know once you were done with the Top Balance and your system goes through some charge/discharge cycles if cell 1 is still in line with the other 3 batteries. If so great, if not then I think that cell may have a different resistance value (poorly matched from the seller).
Folks... correct me if I am wrong but I think a cell's internal resistance can not be fixed with any amount of manual Top/Bottom balancing and you would need to compensate if you continue to use it with on a tighter charge/discharge window, aggressive active/passive balancing or just replace the cell.
Sojourner1
Itinerant
Your situation I probably wouldn't float no need.
I'm in an rv park right now, with my Magnum hybrid inverter I set the load share to 50a (equals pedestal power), that passes powers through to everything in the 5th wheel. I have turned off the charger (stand by) so my batteries power only the 12v loads. The solar tops off what little that is used during the night (35-50ah) then floats the balance of day. This way if power is lost in the park the inverter takes over and life is still good.
Ampster
Renewable Energy Hobbyist
Yes, that has been my experience. It is the cell voltage differences that matter as one charges and discharges. I have a 18P 12S Nissan Leaf pack that consists of 54 modules and I see cell differences under load or charging of 18 to 34 millivolts. Those are voltage differences when the pack is near the top or the bottom of a cycle. The Amperage draw or current is 30 Amps. During times when the just the inverter overhead of 150 Watts (6 Amps) is on that delta can be as small as 6 millivolts.
I admit I did a sloppy job of balancing before assembling the pack. However my pack consists of at least three different vintages of modules some of which are G1 and G2.
Yes that charge profile works great for long term use. I verified it with a battery engineer at battleborn. There is a reason that nearly every battery on market uses this charge profile. It is not special.And this continually boggles my mind. The official stance of this website?
12V LiFePO4 Battery w/ BMS:
- Absorption: 14.5V
- Float: 13.6V
- Inverter Cut-off: 10.7V-12V (depending on size of load and voltage drop etc)
Design Your Own LiFePO4 Solar Power System
This is a complete system solution complete with: BMS Protection Features (over/under voltage/over-current/balancing) Low Temperature Cut-off Programmable Charge/Discharge Bandwidth Settings Matched...www.mobile-solarpower.com
Well the voltage settling of this chemistry does not let this happen. If you try to absorb at 13.6V, you will reach 60-80% SOC. Do not use the SOC chart for absorption settings! Everyone makes that mistake when first using LiFePO4. If you want to modify cycle bandwidth to increase longevity, use a shunt and calculate it yourself for your pack. Do not listen to anyone online about this and do the math at home.
Also keeping at high SOC for prolonged duration will hurt these packs but not much. The lattice has high thermal stability and electrolyte can handle up to 4.2V. Don't worry about it. Everyone tries to modify their profiles to increase life cycle, but this should be done manually with a shunt. Ambient temperature is also another huge consideration here. In my opinion, calendar aging will kill these cells before cycling done for solar. I wouldn't worry about it at all.
And yes great article, but you will notice that the problems occur when you hold it at high SOC for months or push it past 3.6V. Yes upnorthandpersonal is correct about 3.6V, it is plenty. I crank mine so I can top them off fast, but thats just me. Logging each cell and charging them all the way individually is ideal if you can. But putting them in parallel for a top balance wont hurt them. 3.6V is plenty.
Just make sure you are at high SOC and this whole process will take only a couple hours regardless of what method you use.
Last edited:
MisterSandals
Participation Medalist
With my prismatic cells, my experience is that no matter how high i charge them, whether it be 13.6v or 14.4v, they immediately settle down to 13.4v.
So i am not seeing any sign of only reaching 60-80% SOC charging at 13.6v.
I've been banging on 5 batteries for 5 months with hundreds of charge cycles. I don't think i can be convinced not to believe my data.
Ampster
Renewable Energy Hobbyist
That profile may be fine for drop in replacements where one is following the manufacturers specs and has that 8 or 10 year guaranty.
I can tell you that for this DIYer, who wants to get optimal life from some generic batteries from China, I am going to continue to charge my LFP cells to 3.4 per cell.
I am doing that because I want to get more than the 500 cycles that charging them to the maximum will give me per the attached.
See this was a huge mistake I made when I first started charging lifepo4. The settling of voltage can be quite decieving. I actually made a post about it on the community section of my YouTube channel:
Important Update for LiFePO4 Battery Users (scroll to the bottom for a summarized version):
In a couple of my videos, I mentioned charging a LiFePO4 battery to 90% SOC to increase the charge cycle life. I used a SOC chart and charge curves to determine that 13.3 volts for a 4s battery (12v battery), was 90% SOC if charged with CC/CV. Well after doing a week worth of capacity tests, this is wrong.
I have seen a lot of people in the forums argue about the "ideal" charge voltage for this chemistry, which is typically: 3.325/3.45/3.65 volts per cell (most BMS cut off at 3.65v). The problem that I discovered after doing multiple capacity tests, is that LiFePO4 battery voltage will settle a lot after it is done charging. This is not the same with NMC packs (such as the tesla battery).
So you need to charge your LiFePO4 up to 14.0-14.6 if you want 100% capacity (yes, 14.0 volts gave me 100% capacity in my tests). If you charge to 13.3v with CC/CV, you will only have 47% capacity! Not fun.
I was going to do some testing to figure out what absorption voltage would charge to 90% SOC, but then I started thinking about how LiFePO4 can be charged to 4.2v per cell, or 16.8v for a 12v battery (until electrolyte degradation occurs). So charging up to 14.0-14.6 will not cause any noticeable decrease in charge cycle life. The only way it could is if you charge it to 100% and keep the battery in an extremely hot environment. I have a video that covers that as well, with some battery studies as reference.
Also, you can treat these batteries pretty badly and they will still give you 5,000 charge cycles. That is a long time. Calendar aging of your battery may be more of a concern here (especially if batteries are in a hot environment etc).
So I don't see a real need to charge to 90% anymore. These batteries can last for ages and I think the system components around it may break before the battery cells do.
If you wish to actually charge to 90% SOC for your battery, you need a capacity monitor. Once you have charged to 90% SOC, find the voltage while it's charging (before the pack settles) and set your controller.
--------------------------------------------------------
So Long Story Short:
Set your solar charge controller to the LiFePO4 charge setting and do not bother charging to 90% SOC to extend the life of a LiFePO4 pack.
If you want your battery to last a long time, keep it in a cool location. Heat is the enemy! Also, try to size your battery so that the charge rate is less than .4C (which would be 40 amps for a 100ah battery). The larger your bank, the slower each individual battery is charged, and the longer your battery will last.
If you have manual control of your charge profile parameters, it is wise to use Victron's recommendation. This charge profile gives me 100% capacity in all of my tests and none of my cells voltages spike at all:
Absorption: 14.1V
Float: 13.5V
Equalization: Disabled
Temp Compensation: Disabled
Low Temp Cut-Off: 5 degrees Celsius
I also mentioned this below the charge profile recommendation. Did everyone miss that:
Ampster
Renewable Energy Hobbyist
Has anybody seen any of Jeff Dahn's videos? He is the consultant that Tesla hired to improve the life of their cells. He is a professor at Dalhousie University in Canada, and long before he was engaged by Tesla as a consultant he and his graduate students did a lot of research on the life cycle of various Lithium chemistries.
There is a reason Tesla keeps a hidden reserve in their packs that the user cannot access. In addition to that Tesla recommends and has a default setting of 90% for charging a Tesla. Yes that is NMC chemistry which can be charged to 4.2 volts. However from the videos that I have seen it is likely that the cells in a Tesla ever see more than 4.05 volts.
After all of the above and after spending years on diyelecticcar forum that is why this DIYer never charges his LFP cells above 3.4 volts except to run a capacity test and to top balance them. And when I do either of those I discharge them 5% as soon as convenient. Your mileage may vary. Maybe it is time for a Wiki citing some of the Academic and OEM articles on the subject. I had a discussion about this several years ago on another forum with @nebster and I believe he shares the above view. I know in particular he can explain better than I why it is not good to leave a Lithium battery at 100% for very long.
There is a reason Tesla keeps a hidden reserve in their packs that the user cannot access. In addition to that Tesla recommends and has a default setting of 90% for charging a Tesla. Yes that is NMC chemistry which can be charged to 4.2 volts. However from the videos that I have seen it is likely that the cells in a Tesla ever see more than 4.05 volts.
After all of the above and after spending years on diyelecticcar forum that is why this DIYer never charges his LFP cells above 3.4 volts except to run a capacity test and to top balance them. And when I do either of those I discharge them 5% as soon as convenient. Your mileage may vary. Maybe it is time for a Wiki citing some of the Academic and OEM articles on the subject. I had a discussion about this several years ago on another forum with @nebster and I believe he shares the above view. I know in particular he can explain better than I why it is not good to leave a Lithium battery at 100% for very long.
By then, we should have solid state batteries ? and what are your current capacity test figures? What's your low SOC cut off and what did you calculate for absorption?
Yes true in many nmc studies. But lifepo4 has lower degradation in extreme SOC. I haven't found many studies on lifepo4 cycle bandwidth and capacity degradation over time. What study are you referencing? All lithium ion degrade relatively the same w/ some exceptions, but it is true that modifying cycle bandwidth will increase life for most. But in the instance of lto, calendar aging will kill the battery long before capacity loss due to cycling. Thermal stability of lifepo4 needs to be consider here and I don't have much data on it.
Also, not charging up to 100% may not trigger top balancing on some bms. And if cell drift occurs excessively, may have some problems. More an issue on mismatched cell packs.
Need to make sure your absorption voltage is higher than balancing trigger voltage.
Ampster
Renewable Energy Hobbyist
Let us hope, but I follow the research Tesla has done and I do not see them changing their procurement strategies with the exception of trying to reduce Cobalt. Their Maxwell acquisition was for super capacitor capability to enhance acceleration and do regenerative braking when the battery was full or cold.
My four year old Tesla with 100,000 miles has 93% of its original capacity. I have owned 3 lithium packs and going on my fourth.
My Nissan Leaf (NMC) pack is limping along at about 60% of capacity so I never take it below 3.5 volts and don't charge it past 4.05 volts.
Yes but we do not know what that acquisition is for yet. And for stationary storage, they have a partnership with CATL, a LiFePO4 manufacturer. So yes, their procurement strategy can change. They are still using NCA variation for cars. But for solar use? Would be illogical. And their powerwall has a warranty for 5-10 years depending on some factors in the contract and if it has communication with internet.
Yes but thats a tesla. My first tesla had 70K miles with similar capacity figures. Nothing special at all considering cycle life required to achieve that mileage, with that chemistry. We have plenty of data on that. My post was talking about lifepo4 studies and the lack of data. I am not talking about NMC or NCA. I have read most of the NCA studies showing cycling bandwidth modifications and degradation changes. Very straightforward. Did you read my posts? I was not talking about tesla packs.
What chemistry are your other packs?
Yes smart idea to treat those used nissan leaf packs with care.
Last edited:
Ampster
Renewable Energy Hobbyist
Are you talking about dendrite growth at lower or higher SOCs?
According to manufacturers charts Life cycles are 500 when the usage is between 95-100% and 3000 cycles when used at 80%. Given that they will be optimistic I am hedging my bets. I think you would call this lowering my cycle bandwidth.
I am referring to the manufacturers table that show cycle life improves with lower percentage uses. I posted one earlier. My approach to modifying cycle bandwidth is to use a lower absorb voltage. I also discharge to a more conservative value. I assume we agree on the point that lower cycle bandwidth increases life.
Those are all important issues. Ones strategy could be different for different packs and users. For someone who wants to get the most out of every cycle I would suggest higher voltages and let your BMS manage it and balance all the time.
For me with a large pack $3000 to $6000) of perhaps 3 or 4 280 Ahr cells in parallel balancing with less than an Amp isn't going to do much. That is why I have a BMS that disables the balancing. If I properly top balance my pack and with new fairly matched cells I hope to not see much drift especially managing it conservatively. If I do have some imbalance I can adjust the balancing set point on my BMS and do a low Amp charge to see if it will come into balance.
That is actually a strategy used by EVs. If you charge to 80 or 90% every day your EV won't hit the balancing point. But if you take it to 100% ocassionally some drivers have reported small increases in range. I know one of the chargers I bought for an EV conversion had multple charge profiles so the user can determine whether he wanted a maintenance charge or a balancing charge that might take overnight. A lot of my knowledge and philosophy comes from experience in the EV conversion environment for a few years. Those packs get pushed harder than the usual stationary storage pack that most users here are experiencing. However the cost of those packs is why there is more emphasis on getting more cycles. As I often say, it all depends on where you are standing.
Ampster
Renewable Energy Hobbyist
The rumors are that the only way Tesla can reduce Cobalt is by going with LiFePO4. CATL and BYD have significant experience with LiFePO4 so we may see some longitudinal studies over longer time periods for LiFePO4 from those guys if they ever publish. I have seen electric BYD busses in China for the past 7 years. . Who knows, the million mile Tesla may be LifePO4. They may have found otherways to save weight on the cars and maybe LiFePO4 has become a little more energy dense or a least more kWhrs per cubic meter with the new pack assembly. Maybe Maxwell will provide the power density that LifeP04 can't deliver. One thing is certain change is ineveitable.
I have remarked on other threads that I hope this trickles down to lower prices for us hobbyists. I am still amazed that I could buy 28 kWhs of storage for $116 delivered.
My VW conversion was powered by Winstons. The first pack powering my Outback Radian was Thunderskys that I bought slightly used. Those incurred a discharge accident but half of them survived enough to sell them to another EV builder. Then I bought the Nissan Leaf pack and powered the Radian for a couple of years. Sold the Radian and rented that house and moved to the new house. Installed solar and had a Powerwall ordered. Cancelled the Powerwall and installed an Outback Skybox powered by the Leaf pack plus some additional modules from other vintage Leafs. I soon realized that pack was limited. I just bought 28kWhr of 280 Ahr cells that are scheduled to arrive next month.
The data are sparse, but it's very clear that carrying higher temperatures AND holding high SOC will dramatically accelerate capacity fade in LFP. Interestingly, NCA seems to do about as well, and NMC totally falls off of a cliff.
The best paper I've come across is Keil, 2016, JECS. It would be nice to see more tests with newer cell designs. We've had a renaissance in LFP density in the last few years, to the point where the latest prismatic LFPs are close to the (quite a bit older) Tesla cylindrical NCAs. It's possible some of these changes may change the behavior in some way.
I agree that calendar aging could dominate in many real-world settings. (On the other hand, we don't have much aging data on modern cells, either. Who's to say the manufacturing tolerances haven't reduced deterioration?)
LFP is becoming cheap enough now that other factors start to drive the economic rationale for various packs. Not unlike the situation with PV.
Update - My balancing project stalled at 3.38 volts for the last 24 hours. The power supply fell to under 2 amps at 3.65 volts. I suspected the small aligater clips that came with it. So I got larger wires with ring tememals and now I'm pushing almost 5 amps into the pack.
Similar threads
