Mixing LiFePo4 and AGM

[sunshine_eggo]

48V LFP is 16S. Stick with that.

Concept generally not recommended unless the FLA is in a very healthy state (SG of all cells at 1.277+) and meticulously maintained. Even then, no one would actually recommend it, but it's less insane.

Recommend you put a current meter between the banks so you can monitor how much is flowing between them. It could serve as an early warning if an FLA cell fails and starts pulling current from the LFP.

Wouldn't hurt to have voltage monitoring on each 6V group to confirm nothing is getting wonky.

Float for LFP is 3.375V/cell not 3.2. Cell resting or floating at 3.2V is a low SoC (15-20%).

 

[bfitzgerald]

48V LFP is 16S. Stick with that.

Concept generally not recommended unless the FLA is in a very healthy state (SG of all cells at 1.277+) and meticulously maintained. Even then, no one would actually recommend it, but it's less insane.

Recommend you put a current meter between the banks so you can monitor how much is flowing between them. It could serve as an early warning if an FLA cell fails and starts pulling current from the LFP.

Wouldn't hurt to have voltage monitoring on each 6V group to confirm nothing is getting wonky.

Float for LFP is 3.375V/cell not 3.2. Cell resting or floating at 3.2V is a low SoC (15-20%).

Thanks for your reply. Super fast. Yes, I was using the nominal voltage for a LFP as a fully charged, resting voltage and that isn't correct. I believe that a fully charged cell will, over a short period of time, drop to around 3.3v (resting) and depending on the discharge rate, slowly drop to 3.2v like you said at around 15 - 20% SOC. My mistake and it seems that the best configuration would be 16s (as you also said). That configuration would solve the high float rate for the FLA bank but require at least a 3.6 v/c charge rate for the LFP bank to get a high enough voltage to charge the FLA banks. That would require the occasional separation of the LFP bank for an equalization charge of the FLA bank. The separation is easy but putting them back together would require some care with bank voltages.

I have a question on the state of health of the FLA batteries. My batteries, even brand new wouldn't reach 1.277. I'm happy these days at 1.26. I may be undercharging them with solar since the maximum my array will charge at is 60 amps. The inverter charger will do about 85 amps so I can charge at 10% of C. But I don't do it very often. Anyway, assuming that the voltage of all the battery banks are the same, why would it matter if the s.g. of the FLA was in the lower range, as it will always be when discharged? I think that the FLA bank will follow the LFP bank around voltage wise and most of the time provide very little power. I would, of course, monitor the current between the LFP and FLA bank.

I am not trying blow anything up or start a fire so I am curious about the possible bad scenarios that I can control through both monitoring and current protection devices such as breakers, fuses, etc.

Thanks again for your advice.

 

[sunshine_eggo]

Thanks for your reply. Super fast. Yes, I was using the nominal voltage for a LFP as a fully charged, resting voltage and that isn't correct. I believe that a fully charged cell will, over a short period of time, drop to around 3.3v (resting)

depends on cell size. Smaller cells will drop pretty fast. I have 70+ CALB 40Ah cells that drop to 3.30 very quickly, but they are high resistance compared to the 280Ah cells. I let 9 EVE 280Ah cells sit full charged for 5 months. 6 of them were 3.40-3.45, 1 was 3.38, another 3.39 and one 3.30. The 3.30 cell only lost 1.2% SoC in 5 months.

and depending on the discharge rate, slowly drop to 3.2v like you said at around 15 - 20% SOC.

Aforementioned CALB cells sat for a full year and none measured less than 3.30V and retained > 90% charge.

My mistake and it seems that the best configuration would be 16s (as you also said). That configuration would solve the high float rate for the FLA bank but require at least a 3.6 v/c charge rate for the LFP bank to get a high enough voltage to charge the FLA banks.

Not sure what you mean. You list 56.4V as the low end of the FLA bank. That's 3.525.

That would require the occasional separation of the LFP bank for an equalization charge of the FLA bank.

This will be required in almost all cases. Maybe more frequently at the lower absorption.

The separation is easy but putting them back together would require some care with bank voltages.

Yep.

Both of the above are further reasons for not paralleling.

I have a question on the state of health of the FLA batteries. My batteries, even brand new wouldn't reach 1.277. I'm happy these days at 1.26. I may be undercharging them with solar since he maximum my array will charge at is 60 amps. The inverter charger will do about 85 amps so I can charge at 10% of C/20. But I don't do it very often. Anyway, assuming that the voltage of all the battery banks are the same, why would it matter if the s.g. of the FLA was in the lower range, as it will always be when discharged? I think that the FLA bank will follow the LFP bank around voltage wise and most of the time provide very little power. I would, of course, monitor the current between the LFP and FLA bank.

1.277 is Trojan's higher than typical SG value. It generally takes several cycles with 0.1-0.13C charging before full capacity and SG arrive. You might have missed it. Personally, I would be okay with > 1.26 provided all cells within each battery are very consistent.

 

[timselectric]

The only issue I can think of with this combo is whether the agm's not getting it's recommended absorption charge has any effect on it's performance or lifespan but I haven't seen any evidence of that yet after more than a year in service.

I believe that the LFP batteries will provide the absorption charge. Just by being in parallel with the AGM.

 

[bfitzgerald]

depends on cell size. Smaller cells will drop pretty fast. I have 70+ CALB 40Ah cells that drop to 3.30 very quickly, but they are high resistance compared to the 280Ah cells. I let 9 EVE 280Ah cells sit full charged for 5 months. 6 of them were 3.40-3.45, 1 was 3.38, another 3.39 and one 3.30. The 3.30 cell only lost 1.2% SoC in 5 months.

Aforementioned CALB cells sat for a full year and none measured less than 3.30V and retained > 90% charge.

Not sure what you mean. You list 56.4V as the low end of the FLA bank. That's 3.525.

US Battery recommendation for bulk/absorb charging is 56.4 - 58.8 v but my 10 years of experience with this set of batteries tells me that I need to be closer to 60 v, particularly in cooler weather. I currently set the charge controller at the very top of the range and let temperature compensation do it's thing. Spring and fall the compensated charge rate is around 60 v. That was why I was was considering a 17s configuration. I will set the system up as a 16s and monitor it closely, all strings properly fused with the LFP BMS as the final safety cut-off for any shorted cells in the FLA banks.

This will be required in almost all cases. Maybe more frequently at the lower absorption.

Both of the above are further reasons for not paralleling.

1.277 is Trojan's higher than typical SG value. It generally takes several cycles with 0.1-0.13C charging before full capacity and SG arrive. You might have missed it. Personally, I would be okay with > 1.26 provided all cells within each battery are very consistent.

 

[sunshine_eggo]

US Battery recommendation for bulk/absorb charging is 56.4 - 58.8 v but my 10 years of experience with this set of batteries tells me that I need to be closer to 60 v, particularly in cooler weather.

Agree. My T-1275 recommend 14.82V/12V or 59.28V for 48V. I dialed that back mostly because they would boil off a lot of their water at that voltage. Cooler weather has a huge influence. With proper temp compensation, I saw the vicinity of 62V due to extreme cold.

I currently set the charge controller at the very top of the range and let temperature compensation do it's thing. Spring and fall the compensated charge rate is around 60 v. That was why I was was considering a 17s configuration. I will set the system up as a 16s and monitor it closely, all strings properly fused with the LFP BMS as the final safety cut-off for any shorted cells in the FLA banks.

IMHO, if you're charging both banks, go to 58.4V and disable temp comp.

 

[bfitzgerald]

Agree. My T-1275 recommend 14.82V/12V or 59.28V for 48V. I dialed that back mostly because they would boil off a lot of their water at that voltage. Cooler weather has a huge influence. With proper temp compensation, I saw the vicinity of 62V due to extreme cold.



IMHO, if you're charging both banks, go to 58.4V and disable temp comp.

Agree. No temp compensation.

 

[sunshine_eggo]

Agree. No temp compensation.

Something else to consider concerning the health of the FLA... They are barely going to be used at all. the operating voltage are such that they'll hardly discharge much of anything before the LFP is down around 20%. They'll basically be in float mode held at full charge almost all the time.

 

[bfitzgerald]

Something else to consider concerning the health of the FLA... They are barely going to be used at all. the operating voltage are such that they'll hardly discharge much of anything before the LFP is down around 20%. They'll basically be in float mode held at full charge almost all the time.

I agree that the FLA bank will back up the LFP bank. However, I will increase solar power consummation by switching from propane "power" to solar power. During the peak solar 3 months in the summer, we get most of the cabin hot water from excess solar power. Although this is somewhat automated it does require some management. While I will certainly monitor power consumption, I know that if we have more "free" hot water, it will be more convenient. For example, we have a dishwasher which we use whenever the battery state of charge allows it. It is not exactly an off grid friendly appliance or energy efficient, but it's not just the hot water used, it's mostly that the damn thing runs for an hour or longer and endlessly pumps water around. It probably uses .5 or more kwh per load, not even factoring in the hot water production. When we have family around (often) we could, but don't, do 2 loads of dishes a day. That will change. We will be able to use appliances such as slow cookers, sous vide cookers, and induction stove tops instead of propane. I won't start the generator to fill the wood fired hot tub, I plan to dip into the FLA's to insure that they get some exercise and that we enjoy the maximum convenience of having more available power without expanding anything except for adding an LFP bank. In the future the FLA batteries will become a liability, but that could be 10 years down the road. Who knows what wonderful battery technology will become available whenever that day comes.

I know that this seems a bit like getting more power for nothing (except for the cost of the LFP cells) but that is my plan. We currently consume 9 -12 kwh per day depending on how many people are around. We can run for a couple of cloudy days without running the generator. Even cloudy days produce some power. The most power we have ever produced on a single day is 15.5 kwh. With the 2 way efficiency of LPF chemistry, I think that we can get a 20% or better efficiency bonus from our current array/equipment and I plan to spend that bonus. I do hope that I have not set my expectations too high!

 

[sunshine_eggo]

I agree that the FLA bank will back up the LFP bank. However, I will increase solar power consummation by switching from propane "power" to solar power. During the peak solar 3 months in the summer, we get most of the cabin hot water from excess solar power. Although this is somewhat automated it does require some management. While I will certainly monitor power consumption, I know that if we have more "free" hot water, it will be more convenient. For example, we have a dishwasher which we use whenever the battery state of charge allows it. It is not exactly an off grid friendly appliance or energy efficient, but it's not just the hot water used, it's mostly that the damn thing runs for an hour or longer and endlessly pumps water around. It probably uses .5 or more kwh per load, not even factoring in the hot water production. When we have family around (often) we could, but don't, do 2 loads of dishes a day. That will change. We will be able to use appliances such as slow cookers, sous vide cookers, and induction stove tops instead of propane. I won't start the generator to fill the wood fired hot tub, I plan to dip into the FLA's to insure that they get some exercise and that we enjoy the maximum convenience of having more available power without expanding anything except for adding an LFP bank. In the future the FLA batteries will become a liability, but that could be 10 years down the road. Who knows what wonderful battery technology will become available whenever that day comes.

I know that this seems a bit like getting more power for nothing (except for the cost of the LFP cells) but that is my plan. We currently consume 9 -12 kwh per day depending on how many people are around. We can run for a couple of cloudy days without running the generator. Even cloudy days produce some power. The most power we have ever produced on a single day is 15.5 kwh. With the 2 way efficiency of LPF chemistry, I think that we can get a 20% or better efficiency bonus from our current array/equipment and I plan to spend that bonus. I do hope that I have not set my expectations too high!

15% is pretty much the upper limit of the efficiency gains in a complete switch to LFP from FLA. If you actually get into the FLA capacity, you're going to eat away at that efficiency.

IMHO, you have a PV problem, not a battery problem. A typical 9-12kwh/day consumption vs. 15.5kWh best ever is a pretty tight margin. IMHO, 14kWh+ should be typical for a 9-12kWh/day consumption.

PV power and daily harvest determines how much energy you can use per day.
Battery capacity determines how long you can go between charging.

In other words, if you want to use more energy in a 24hr period, you need more PV. If you want to go longer between charges due to inclement PV conditions, you need more battery.

 

[timselectric]

Short and sweet........
You need enough inverter to run what you want, at any given time.
You need enough battery to run what you want, when there is no sun.
You need enough solar to run what you want and recharge the battery, when there is sun.
Keep growing until you have accomplished all three.

 

[EWesch]

Just keep in mind that a huge danger of these types of configurations is if an AGM has a shorted cell (which will happen eventually), and your lifepo4 packs feed into that cell (considering their low internal resistance, can push a lot of current), your AGM will create a lot of heat. It will probably cause the over pressure relief valve to open. Battery could melt, or the lid could pop off, worst case. Really messy, really dangerous.

I tell people not to do it. But if you do, use a fuse for each parallel string of packs.

I'm running 4 server rack batteries from SS and, though happy with their performance, I occasionally get a complete shut down when the system is switching from shore power or generator to inverting. If I have solar or DC/DC from the alternator, it doesn't happen. Logic tells me it is a timing issue with the inverter and BMSs switching because it never happened with the original AGM batteries.

A couple questions...

1. If properly fused, is there a reason that adding a couple AGM batteries to prevent voltage sag when the BMSs are switching from inrush to outrush is a BAD idea?
2. Other than giving up a little top end storage capacity from the LifePO4 batteries, is there a major downside to charging the entire system with an AGM algorithm?
3. How bad is it to keep the LifePO4s topped off by the inverter/charger when not is use?
4. Do the 12v 400AH LifePower4 server rack battery BMSs have a low temperature disconnect that is effective enough for me to just leave it plugged in to shore power all the time?

 

[sunshine_eggo]

1. AGM don't really start to deliver much of anything above ~13V.
2. Provided LFP cell voltages are within spec, and the absorption period isn't too long, no.
3. Depends on temperature and time. Mild/cool temps and short times (a few days), it's not a concern. Warmer temps and long times (weeks/months), the LFP will degrade.
4. What does the battery datasheet say? In most cases, the answer is yes.

 

[EWesch]

1. AGM don't really start to deliver much of anything above ~13V.
2. Provided LFP cell voltages are within spec, and the absorption period isn't too long, no.
3. Depends on temperature and time. Mild/cool temps and short times (a few days), it's not a concern. Warmer temps and long times (weeks/months), the LFP will degrade.
4. What does the battery datasheet say? In most cases, the answer is yes.

1. When I unplug shore power, the batteries and the inverter shut off. If the engine is idling and there is current being supplied by the DC/DC charger, this doesn't happen. Is it logical to believe that supporting the system with battery voltage from a source that doesn't rely on a BMS to switch it from inrush to outrush would prevent the power loss?
2. How long is too long?
3. This is another reason I want to add AGMs to the system. I would prefer to just charge the LifePO4 to 100% then shut them off when not needed/in storage.
4. Would be a moot point if I can do #2 and #3... charge them to 100% with the AGM algorithm then shut them off and let the system sit in storage on shore power with the inverter/charger maintaining the AGMs.

Does this make sense or am I overcomplicating things to preserve some LifePO4 cycles that I will never get near in my application?

 

[sunshine_eggo]

1. I think so.
2. LFP rarely needs more than 15-30 minutes of absorption. 1-2 hours is probably fine.
3. Fully charging them and letting them sit for extended periods still has the same rules.

Most folks want to parallel them because they already have AGM or FLA, and they don't want to waste the old batteries. Some have a small AGM in parallel to provide a voltage source if the BMS disconnects, but not for any "cycle preservation" as you say.

The better solution is to insure the battery BMS never trips by engaging equipment protection inside of BMS limits.

 

[bfitzgerald]

15% is pretty much the upper limit of the efficiency gains in a complete switch to LFP from FLA. If you actually get into the FLA capacity, you're going to eat away at that efficiency.

IMHO, you have a PV problem, not a battery problem. A typical 9-12kwh/day consumption vs. 15.5kWh best ever is a pretty tight margin. IMHO, 14kWh+ should be typical for a 9-12kWh/day consumption.

PV power and daily harvest determines how much energy you can use per day.
Battery capacity determines how long you can go between charging.

In other words, if you want to use more energy in a 24hr period, you need more PV. If you want to go longer between charges due to inclement PV conditions, you need more battery.

Good points for sure and I will take the 15% if that's all I get. But there are many, many afternoons that the charge controller sits on float when I don't have anywhere to use or store more solar. Having more battery storage capacity available is going to equal more power available and it wouldn't matter what kind of batteries I added. I could add more solar but I would need to upgrade my solar charging capacity beyond it's 60 amp limit. And I'd need a location for more panel racking, not impossible but a bit challenging. When I get this up and running I will report the results back the forum. Hopefully good results!

Thanks for your helpful insight and suggestions.

 

[pollenface]

I'm currently using Gels to top up my AGMs with a dc-dc charger through the night. By day they are charged separately and by night the gels scc streetlight function activates and begins giving a small amount of power to the AGMs. The system is working well. I will likely switch the gels to a lifepo4 in the future.

 

[EWesch]

1. I think so.
2. LFP rarely needs more than 15-30 minutes of absorption. 1-2 hours is probably fine.
3. Fully charging them and letting them sit for extended periods still has the same rules.

Most folks want to parallel them because they already have AGM or FLA, and they don't want to waste the old batteries. Some have a small AGM in parallel to provide a voltage source if the BMS disconnects, but not for any "cycle preservation" as you say.

The better solution is to insure the battery BMS never trips by engaging equipment protection inside of BMS limits.

They are Signature Solar LifePower4 12v 400AH Server Rack Batteries and, from what I understand, the BMS settings are not user programmable.

I added 1 AGM battery to my system last night to test my theory and it completely resolved my inrush/outrush shut down condition. I tried every scenario and it functioned flawlessly.

Question 1 ... How is mixing lithium with AGM more dangerous that a bunch of AGMs in parallel? If you have 8 AGM batteries in a factory configuration from an RV (or golf cart) manufacturer, and 1 of them shorts internally, the other 7 will feed into the failed battery at an uncontrolled rate with no fuses or breakers to disconnect them. In a mixed configuration, the BMS in the lithium battery would most certainly detect the short and disconnect. Logic tells me that putting an AGM and a lithium (with a BMS) in parallel would be safer than putting 2 AGMs in parallel.

Question 2 ... Has anyone actually tested a hybrid setup to qualify resting voltage variances, parasitic drains, and diminished performance or is it just theory? My AGM starting batteries rest right around 13.1 when they are in storage and being maintained by a 3 amp battery tender. My lithium bank, when being stored at 60% charge, rests at the same exact voltage. If 2 AGMs requires no more than 3 amps (at 12v) to maintain a 13.1 constant voltage, how long would it take those 2 batteries to drain a bank of lithium batteries down to the accepted 12.8(?) resting voltage of the the AGMs?

I've got room for 6 and I'm really tempted to add 600AH of AGM to my 1600AH of lithium. I really want someone to provide definitive proof that it is more dangerous than paralleling AGMs and/or it will reduce the efficiency of my lithium by a measurable amount.

 

[wattmatters]

Question 1 ... How is mixing lithium with AGM more dangerous that a bunch of AGMs in parallel?

It's a good question. I don't have an answer other than perhaps the much lower internal resistance of the LiFePO4 means a LOT more current can flow in a very short space of time meaning the potential for damage is greater.

I think the answer is simply to have appropriate battery protection no matter what combination of batteries you use.

 

[sunshine_eggo]

They are Signature Solar LifePower4 12v 400AH Server Rack Batteries and, from what I understand, the BMS settings are not user programmable.

I added 1 AGM battery to my system last night to test my theory and it completely resolved my inrush/outrush shut down condition. I tried every scenario and it functioned flawlessly.

Question 1 ... How is mixing lithium with AGM more dangerous that a bunch of AGMs in parallel? If you have 8 AGM batteries in a factory configuration from an RV (or golf cart) manufacturer, and 1 of them shorts internally, the other 7 will feed into the failed battery at an uncontrolled rate with no fuses or breakers to disconnect them. In a mixed configuration, the BMS in the lithium battery would most certainly detect the short and disconnect. Logic tells me that putting an AGM and a lithium (with a BMS) in parallel would be safer than putting 2 AGMs in parallel.

Typical AGM internal resistance is typically 10X that of LFP, and they discharge at a lower voltage.

If you don't perceive the potential for 10X the current flow and subsequent heating of a failed AGM cell as more dangerous, I don't think any practical argument will dissuade you from your desire to do so, and that's fine.

Question 2 ... Has anyone actually tested a hybrid setup to qualify resting voltage variances, parasitic drains, and diminished performance or is it just theory? My AGM starting batteries rest right around 13.1 when they are in storage and being maintained by a 3 amp battery tender. My lithium bank, when being stored at 60% charge, rests at the same exact voltage. If 2 AGMs requires no more than 3 amps (at 12v) to maintain a 13.1 constant voltage, how long would it take those 2 batteries to drain a bank of lithium batteries down to the accepted 12.8(?) resting voltage of the the AGMs?

Maintaining typical AGM at 13.1V is generally fine for storage, but in an operational system, most need to be floated at 13.5-13.8V. There may be a higher draw at that voltage. It's not that big of a concern except in very long term storage.

I've got room for 6 and I'm really tempted to add 600AH of AGM to my 1600AH of lithium. I really want someone to provide definitive proof that it is more dangerous than paralleling AGMs and/or it will reduce the efficiency of my lithium by a measurable amount.

Again, the 10X current potential at a higher voltage alone is enough "proof" to establish that one is more dangerous than the other.

Many folks have done this. I have some spare FLA laying about, and I would consider it if I had LFP, but NMC and FLA/AGM/GEL are a bad match for each other. The most important thing is to understand the implications and put safety measures in place. A Class T fuse between the two banks is probably all that's needed to minimize danger.