diy solar

diy solar

Mixing LiFePo4 and AGM

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

The internal resistance spec for my 12 V Enersys SLA batteries is 3.2 mΩ. Short circuit current rated at 3990 A.

My 51.2 V LiFePO4 server rack batteries are rated with total impedance <120 mΩ, which is interesting considering the individual 3.2 V cells are rated at <0.68 mΩ. No short circuit current rating listed but the short circuit protection is rated at <30 ms, which I guess is a long time when big currents are flowing.
 
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.

Don't give up on me here! I'm here for a practical argument to dissuade me or, better yet, proof to show me. I travel and camp with my family in this RV and don't want to do anything to intentionally endanger them.

I just don't understand how 4 LFP batteries that can't start my Inverter without going into over-current protection have the potential to be more dangerous, in a dead short situation, than 5 AGMs connected to that shorted battery with 4/0 cable. I've read a research article by Paul Koinek that shows a single VRLA battery can produce upward of 1700 amps in a short circuit situation. With 6 VRLA batteries in an un-fused parallel bank, like they come from RV manufacturers, the potential dead short current flow is immense. Do they just assume that the current surge will destroy whatever is creating the short so quickly that the risk of a thermal event is minimal?

I understand my LFP battery bank has a massive amount of stored energy but its hard to ignore the fact that I'm having to use a 100AH AGM battery to handle the transitional current flow of my inverter because my 1600AH LFP battery bank goes into over-current protection without it. I also feel the need to point out that Signature Solar ships these batteries with 3 gauge wires to connect them to a bus bar. My 2 AGMs that are serving as cranking batteries would turn a 6 gauge wire into a light bulb filament before the Cummins would even think about turning. An electric motor's locked rotor current is basically a short circuit that the amperage has to overcome for the first 1/2 cycle, right?

I want the AGMs torque and the LFPs horsepower because, one without the other is....well... I just want both. I'm not a millennial but, in this case, I want my cake and eat it too.
 
Last edited:
For reference:

The internal resistance spec for my 12 V Enersys SLA batteries is 3.2 mΩ. Short circuit current rated at 3990 A.

My 51.2 V LiFePO4 server rack batteries are rated with total impedance <120 mΩ, which is interesting considering the individual 3.2 V cells are rated at <0.68 mΩ. No short circuit current rating listed but the short circuit protection is rated at <30 ms, which I guess is a long time when big currents are flowing.
It's a good read...


If you want to skip to the good part...

Using Ohm’s law, the potential maximum, zero voltage short circuit current can be calculated by dividing the battery’s nominal open circuit voltage by its resistance (I = V/R). By discharge testing over a wide range of currents and measuring the battery’s voltage response, its internal resistance can be calculated from the slope of the voltage versus current (R = dV/dI). Extrapolating this line back to zero volts yields the resistance-free or zero voltage short circuit current. As the range and magnitude of discharge currents increase, the accuracy of the resistance and short circuit current values increase. In IEC896-2 “Stationary Lead-Acid Batteries, Part 2: Valve Regulated Types”, the estimated short circuit current is obtained by discharging a battery at 4 times and 20 times its rated 10 hour discharge current (I10 at 25o C to 1.75 volts per cell). At the 4X rate, the battery voltage is measured at 20 seconds. After a 5 minute rest without recharge, the battery is discharged at the 20X rate and the voltage is measured after 5 seconds. From these two points, a line is extrapolated back to zero volts to calculate the short circuit current. A rule of thumb estimation method suggests multiplying the 1 minute rate by 10 to obtain the short circuit current; however, there may be a wide variation between estimated and actual values depending on battery design.
 
Don't give up on me here! I'm here for a practical argument to dissuade me or, better yet, proof to show me. I travel and camp with my family in this RV and don't want to do anything to intentionally endanger them.

I just don't understand how 4 LFP batteries that can't start my Inverter without going into over-current protection have the potential to be more dangerous, in a dead short situation, than 5 AGMs connected to that shorted battery with 4/0 cable.

BMSs fail.

I've read a research article by Paul Koinek that shows a single VRLA battery can produce upward of 1700 amps in a short circuit situation. With 6 VRLA batteries in an un-fused parallel bank, like they come from RV manufacturers, the potential dead short current flow is immense. Do they just assume that the current surge will destroy whatever is creating the short so quickly that the risk of a thermal event is minimal?

VRLA can be FLA or AGM. What will happen when they produce that current is a massive voltage drop. If the voltage drop goes below the parallel battery, current will no longer flow any faster beyond that point.

Consider CCA ratings: this is how many amps a 12V battery can supply for 30 seconds at 0°C and stay above 7.2V.

I understand my LFP battery bank has a massive amount of stored energy but its hard to ignore the fact that I'm having to use a 100AH AGM battery to handle the
transitional current flow of my inverter because my 1600AH LFP battery bank goes into over-current protection without it.

We have ignored this, but this shouldn't happen. If you're regularly triggering LVP or HVP setting yourself up for potential OCP, then your system is not properly designed/configured. Your inverters should be shutting down well before the BMSs trigger protection. This prevents discharge of the capacitors and should eliminate the OCP condition.

Your experience provides evidence of the large current potential, otherwise it wouldn't trip.

I also feel the need to point out that Signature Solar ships these batteries with 6 gauge wires to connect them to a bus bar.

That is only evidence of poor decision making on their part. Quite shameful.

My 2 AGMs that are serving as cranking batteries would turn a 6 gauge wire into a light bulb filament before the Cummins would even think about turning. An electric motor's locked rotor current is basically a short circuit that the amperage has to overcome for the first 1/2 cycle, right?

I want the AGMs torque and the LFPs horsepower because, one without the other is....well... I just want both. I'm not a millennial but, in this case, I want my cake and eat it too.

Then I simply refer you back to my previous post...

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.

The LFP has BMS OCP as well as a breaker per battery. Adding a 3rd layer of protection with a class T fuse between the two banks is additional peace of mind.
 
The website wildebus.com has been mentioned previously on this forum. If you go to their web site there is a section called projects that covers the LA/LFP hybrid battery with some pretty interesting test information. I found I quite informative.
 
So I've got a little beginner's system I put 3 x 35aH AGMs on in parallel for a total of 105ah. Panels are 200W for now, but not a factor for this question...

I can use a couple of those AGM's on another application, which would leave me one 35aH AGM. I want to upgrade the system to LiFePo4, so if I bought a nice and cheap 100aH LiFePo4 to replace the 2 AGM's, could I wire the LiFePo4 and the one remaining 35aH AGM just to get the addtional amp hours? Would the AGM drag down the LiFePo4 and ruin it?

Let's go one further: If I left the 3 AGMs inline and added the new LiFePo4 so the mix was more balanced would it have bad affects on the LiFePo4? 105ah AGM to 100ah LiFePo4?
On a sailboat, off grid mostly I have;
six fullriver 2v AGM in series for 12v
16 280ah eve cells in 4s4p with overkill 120 BMS 12v
six 330w solar on an aluminum fabricated hard top 3s2p 120v dc
two midnite solar classic 150's each 150v 100a
I programmed the MPPT controllers for the AGM's years ago to the fullriver tech's specs, they like a good two hours a day at 14.7v if possible
when I added the Lifepo4 to the system, I paralleled them using a victron lynx
with a victron battery protect on the lithium side, set to connect and disconnect the lithium to about 90% of their capacity
so approx 2kw of solar 15kw agm and 15kw of lithium all working as it should,
fulltime monitoring on a cerbo GX and quattro give me enough power for air conditioning, 5 gal domestic water heater fridge (obviously) lights tv's computers etc etc
the fullrivers are about six years old now but show no signs of degradation, after a couple of cloudy or rainy days, when the lithiums have removed themselves from the equation and the voltage is starting to drop, the cebo sends me an alert and I can either turn off the big loads and wait,
or, start the 8k genset and recharge everything.
so if you take your time and make sure your individual chemistries get the volts and amps that suits them
It's like ebony and ivory living together in perfect harmony, on pianos
 
On a sailboat, off grid mostly I have;
six fullriver 2v AGM in series for 12v
16 280ah eve cells in 4s4p with overkill 120 BMS 12v
six 330w solar on an aluminum fabricated hard top 3s2p 120v dc
two midnite solar classic 150's each 150v 100a
I programmed the MPPT controllers for the AGM's years ago to the fullriver tech's specs, they like a good two hours a day at 14.7v if possible
when I added the Lifepo4 to the system, I paralleled them using a victron lynx
with a victron battery protect on the lithium side, set to connect and disconnect the lithium to about 90% of their capacity
so approx 2kw of solar 15kw agm and 15kw of lithium all working as it should,
fulltime monitoring on a cerbo GX and quattro give me enough power for air conditioning, 5 gal domestic water heater fridge (obviously) lights tv's computers etc etc
the fullrivers are about six years old now but show no signs of degradation, after a couple of cloudy or rainy days, when the lithiums have removed themselves from the equation and the voltage is starting to drop, the cebo sends me an alert and I can either turn off the big loads and wait,
or, start the 8k genset and recharge everything.
so if you take your time and make sure your individual chemistries get the volts and amps that suits them
It's like ebony and ivory living together in perfect harmony, on pianos
Hey Joe,

At what voltage do you float the system? When I use the AGM setting on my Magnum, the charger shuts off when they are "full". The charger never turns back on because the Lifepo4 bank holds the voltage high enough that it never feels the need to turn back on.

I know that I'm giving up some top end on my LFPs but feel that not storing them at 100% (14.1-ish) isn't necessarily a bad thing. With the lower resting voltage of the AGMs they do tend to sap a small amount from the LFP when the charger is off, again, not a huge deal. What I'd like to do is CC/CV both banks as high as possible (13.8?) without damaging my AGMs. That way everything is at it's best, understanding there are compromises, when I'm ready to take off on a trip.
 
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.
Well I finally paralleled my 16 - L16 FLAs with my 280 a/h EVE DIY battery. All strings are running through 125 amp breakers. For monitoring I installed a Victron Smartshunt to measure and record the current into and out of the 2 FLA strings. I use the JK BMS to monitor the state of operations of the LFP pack.

To start the parallel operation I set the float rate on the CC to match the voltage of the resting LFP battery so after a full charge on the FLAs their voltage drifted down and when the voltages matched I flipped on the LFP breaker and it was a non event. The peak current flow from the LFP was a couple of amps and soon down to less than an amp. In four days of operation, some smoky and not too sunny the LFP operated between 45 - 90% SOC. The Victron shunt does show some current flow but even when the inverter is pulling 60 amps, the FLAs provide maybe 5%. When the load drops off, the LFP's promptly return the used power to the FLAs. Mostly the shunt shows a net of 0 amps from the FLAs. There will be some cloudy days so I will be able watch the change over from LFP to FLA power consumption but so far so good.

I have the CC set at 56.8 v in bulk and absorb and have turned off float, at least for now. I am allowing 4 hours maximum of absorption time with a cut-off at 8.4 amps. 56.8v is a pretty conservative 3.55 v/c and in four days I have yet to see any cell come even close to that voltage. I'm going to mess around with the CC voltages and settings and keep a close eye on things. I would love to hear your thoughts on optimum charge settings.

Brian
 
Well I finally paralleled my 16 - L16 FLAs with my 280 a/h EVE DIY battery. All strings are running through 125 amp breakers. For monitoring I installed a Victron Smartshunt to measure and record the current into and out of the 2 FLA strings. I use the JK BMS to monitor the state of operations of the LFP pack.

To start the parallel operation I set the float rate on the CC to match the voltage of the resting LFP battery so after a full charge on the FLAs their voltage drifted down and when the voltages matched I flipped on the LFP breaker and it was a non event. The peak current flow from the LFP was a couple of amps and soon down to less than an amp. In four days of operation, some smoky and not too sunny the LFP operated between 45 - 90% SOC. The Victron shunt does show some current flow but even when the inverter is pulling 60 amps, the FLAs provide maybe 5%. When the load drops off, the LFP's promptly return the used power to the FLAs. Mostly the shunt shows a net of 0 amps from the FLAs. There will be some cloudy days so I will be able watch the change over from LFP to FLA power consumption but so far so good.

I have the CC set at 56.8 v in bulk and absorb and have turned off float, at least for now. I am allowing 4 hours maximum of absorption time with a cut-off at 8.4 amps. 56.8v is a pretty conservative 3.55 v/c and in four days I have yet to see any cell come even close to that voltage. I'm going to mess around with the CC voltages and settings and keep a close eye on things. I would love to hear your thoughts on optimum charge settings.

Brian
I would stick with your current settings and watch it for a few weeks. Then you can decide if you should adjust it.
 
At full state of charge most of the load current will be drawn from LFP batteries. As LFP capacity gets discharged then AGM will start to provide some more of load current.
 
Well I finally paralleled my 16 - L16 FLAs with my 280 a/h EVE DIY battery.
Is yours a 17s LiFePO4 set up like you previously mentioned you were going to go with?

Like you, the connection was a total non-event.

In my case I am using SLA, not FLA so I could use regular 16s LiFePO4 set up as the SLA charge voltages are lower than for FLA.

Even if there was a big current, the BMSs in my LiFePO4 batteries have in built current limiters.

When I added my last LiFePO4 server rack battery I didn't care about its SOC or voltage when connecting it. It did absolutely nothing other than begin to self balance with the rest of the battery bank with some modest current flowing.

I treat the whole battery as having the capacity of the LiFePO4 only and charge/discharge accordingly. The SLA is only there for outage backup.
 
Is yours a 17s LiFePO4 set up like you previously mentioned you were going to go with?

Like you, the connection was a total non-event.

In my case I am using SLA, not FLA so I could use regular 16s LiFePO4 set up as the SLA charge voltages are lower than for FLA.

Even if there was a big current, the BMSs in my LiFePO4 batteries have in built current limiters.

When I added my last LiFePO4 server rack battery I didn't care about its SOC or voltage when connecting it. It did absolutely nothing other than begin to self balance with the rest of the battery bank with some modest current flowing.

I treat the whole battery as having the capacity of the LiFePO4 only and charge/discharge accordingly. The SLA is only there for outage backup.
I ended up using the 16s configuration. The voltages were favorable enough to go that way. I think that I will keep the banks in parallel since the FLA will do nothing until the LFP is down to less than 30%. I think that once a month I will disconnect the LFP and run a charge cycle which generally means a discharge to around 70% SOC on the FLA banks.
 
Back
Top