LiFeP04 Batteries for Solar & BMS

[Living Large]

So far so good.

I'm not so sure about that. I think we have a different vision. Since I am not sure what you suggest can be implemented with the equipment I am looking at, see if what I am suggesting stands any chance, or similarly you don't think it will work. I was thinking the BMS thresholds would be never to exceed high and low values that indicate one or more cells has gone out of balance, started to malfunction, etc. The charger/inverter and CC settings window would be fully contained inside of the BMS window. I was going to use the BMS triggers, both high and low, to shut down the system. The programmable thresholds of the CC and inverter/charger would be the ones "normally" used, both low and high. Your vision is the high BMS threshold is where I would stop charging. You say below that the low BMS voltage would mean it needs to be "recharged," but in my vision this may mean re-balancing might be due because that cell fell lower than the others. We bottom balanced the pack. Recharging is needed at whatever the low threshold is set at in the charger and CC, which I was going to set above the "never to exceed" low voltage.

Note: The input on the XW is a SYSTEM POWER OFF signal - shuts the XW off. I don't know that it will function as "charge is complete" signal as you suggest. It shuts off both the inverter and charger.

OK this is where you start to go off track. During charge cycle be it from solar or generator the BMS is monitoring all cell voltages. If the pack has been bottom Balanced properly, the weakest cell reaches 90 to 95% SOC before any other cell. When that condition is met the BMS sends a signal to the charger to terminate the charge. Nothing gets disconnected, thus nothing is at High Voltage, you are just done charging and everything is normal.

In my vision, something IS high voltage - the cell has exceeded the highest voltage we want it to see.

OK this is where the wheels fly off the track. There is nothing to repair or replace. The LVD is a fail safe to prevent any one cell from being over discharged. If a LVD is issued means you have discharged one or more cells to the SAFE LIMIT, and it disconnects to prevent damage. All it means is you need to recharge before it reconnects, there is no damage.

What I meant by that is, either the cell has degraded or it has gone out of balance. That's all. Instead of letting things get worse, shut down. Make the monkey (me) come along and figure out what cell has gone out of bounds, for whatever reason. Rebalancing is supposed to be rare, since LFP cells tend to stay in balance.

So the real question is how do we take equipment made for Pb and make it work on LFP. {snip} So from this point forward let's call it Charge Terminate or CT. OK?{snip} So in theory the INVERTER should trip before the BMS.

I want to terminate charge normally at whatever high threshold is set in the CC and charger for the entire pack. We bottom balanced, so I wouldn't have the exact same 90% SOC on each cell, but do I care? The BMS high threshold would be above that, and represent that I detected a high voltage on a single cell, not waiting to see it on the entire pack, which may mask the individual exceedance.

Let's go one step further. What if we buy an Integrated Inverter with built-in genny support and charger?

The XW is that, isn't it?

[Sunking]

LL is it possible you are stuck inside a Pb box mentality. It is easy to do. With Pb we work on system battery voltages as a whole. Example charge at 56 volts, and disconnect @ 42 volts on 48 volt system. With LFP we work or control at the individual cell level voltages without much regard for total cell voltages as that lacks the level of detail we need to know.

As discussed we can either balance at the top of bottom. Perhaps Balance is not the right way to think of it because it is really a reference point voltage. A 100 AH LFP cell actual capacity will range from 101 to 115 AH. If we balance at the top, all cells will be 100% SOC voltage of 3.43 volts when rested. As we discharge and we hit say 101 AH discharged, the weakest cell will already be down to 2.0 volts and falling into death zone while all the other cells still have acceptable voltages and pack voltage would still appear to be OK when it is not

If we Bottom Balance we are setting SOC voltage at 0% where all cell voltages are equal. When we charge and monitor for the first cell to hits 100%, it is going to be the weakest cell at 101 AH, and its resting voltage is 3.43 volts and the other 15 are at a lower SOC voltage. However all 16 cells have the same 101 AH capacity which is dictated by the lowest cell.

Sorry if that is repetitive, but very important for you to understand.

As of today there is no solar charge controller made to work with Lithium. When it does I promise will be designed for Top Balance. However we can use what controllers are on the market today if we bottom balance because all we need is the Constant Current or Bulk mode with a BMS to achieve that goal. To do that all we have to do is set Bulk to something equal to or higher than 57.6 volts (16 x 3.6 volts). Once that first cell reaches say 3.4 volts we want the controller to do one of two things. 1. Turn off until the next morning, or lower voltage and go into float mode just below the pack total voltage, so charge current stops until the batteries drain down to Float Voltage level so we can have the power from the panels to be as much as possible while the Sun is up.

[tasman]

What I meant by that is, either the cell has degraded or it has gone out of balance. That's all. Instead of letting things get worse, shut down. Make the monkey (me) come along and figure out what cell has gone out of bounds, for whatever reason. Rebalancing is supposed to be rare, since LFP cells tend to stay in balance.

I want to terminate charge normally at whatever high threshold is set in the CC and charger for the entire pack. We bottom balanced, so I wouldn't have the exact same 90% SOC on each cell, but do I care? The BMS high threshold would be above that, and represent that I detected a high voltage on a single cell, not waiting to see it on the entire pack, which may mask the individual exceedance.

The XW is that, isn't it?

A Lifpeo4 battery pack doesn't necessarily stay in balance, it depends on a number of things I believe. Size of pack, number of cells in lines, where in the pack you connect load and charge to, makes a big difference. Unlike LA, where you connect to dedicated terminals, with lifepo4 there are a number of options. That's why we use active cell balances to compensate for uneven energy distribution, cell acceptance and capacities.

My packs are connected diagonally, which allows for better distribution across the pack compared to connecting to the front terminals. Small packs seem to stay in balance better than large ones because of less discrepancy in distributing charge across all the cell lines and individual cell blocks.

Been thinking of connecting to all 4 points of the pack to get a better more even distribution, but don't know the technicalities and have yet to look at them, so don't know if it will work or be detrimental. My gut feeling is connecting both negatives at each end of the pack and each positive, would allow for an even better distribution and make balancing more stable. Maybe someone can enlighten me on that score.

[Living Large]

LL is it possible you are stuck inside a Pb box mentality. It is easy to do. With Pb we work on system battery voltages as a whole. Example charge at 56 volts, and disconnect @ 42 volts on 48 volt system. With LFP we work or control at the individual cell level voltages without much regard for total cell voltages as that lacks the level of detail we need to know.

As discussed we can either balance at the top of bottom. Perhaps Balance is not the right way to think of it because it is really a reference point voltage. A 100 AH LFP cell actual capacity will range from 101 to 115 AH. If we balance at the top, all cells will be 100% SOC voltage of 3.43 volts when rested. As we discharge and we hit say 101 AH discharged, the weakest cell will already be down to 2.0 volts and falling into death zone while all the other cells still have acceptable voltages and pack voltage would still appear to be OK when it is not

If we Bottom Balance we are setting SOC voltage at 0% where all cell voltages are equal. When we charge and monitor for the first cell to hits 100%, it is going to be the weakest cell at 101 AH, and its resting voltage is 3.43 volts and the other 15 are at a lower SOC voltage. However all 16 cells have the same 101 AH capacity which is dictated by the lowest cell.

Sorry if that is repetitive, but very important for you to understand.

As of today there is no solar charge controller made to work with Lithium. When it does I promise will be designed for Top Balance. However we can use what controllers are on the market today if we bottom balance because all we need is the Constant Current or Bulk mode with a BMS to achieve that goal. To do that all we have to do is set Bulk to something equal to or higher than 57.6 volts (16 x 3.6 volts). Once that first cell reaches say 3.4 volts we want the controller to do one of two things. 1. Turn off until the next morning, or lower voltage and go into float mode just below the pack total voltage, so charge current stops until the batteries drain down to Float Voltage level so we can have the power from the panels to be as much as possible while the Sun is up.

I am not stuck in Pb mentality, because remember, I am brand new to this. What I am stuck with is the current state of the art, and that is why I am looking at it the way I am. While a BMS can monitor to the cell level, the question is what can be done with that information. That was the angle I was coming from.

You didn't specifically critique what won't work in what I proposed. I told you one potential issue with your proposed solution - the XW, which appears to be a pretty flexible unit, will shut down using the Aux input. It's definition is "REMOTE POWER OFF". It doesn't suspend activity and wait for the input to change state. So I am stuck with equipment function limitations, and how to operate within the confines. You're the one who taught me this.

I sense this is not the route you would go, but I ask you, IF I have a CC and inverter/charger that shut down using their Aux inputs, would it be [not possible], [no added benefit], [acceptable but not an optimal use of resources], [acceptable] to add a BMS to an LFP system with gen backup as follows:

- use the BMS to only flag crossing dangerous low and undesired high thresholds (with a timer to account for transients, if needed) and shut down the offending operation (charge, load). User intervention would be required to resume.
- rely on only the built in CC and inverter/charger thresholds and operating modes for normal operation.

We started the LFP conversation a month or more ago with your suggestion that you wouldn't suggest an LFP bank without a BMS. At the extreme, one could have an LFP bank with no BMS, monitor it frequently to assure proper operation. My idea, since I am constrained by equipment, is add a BMS to prevent out of bounds conditions that normally would not be seen assuming the cells haven't fallen out of balance. Is this not the kind of protection I would want to add? Does what I suggest not provide such protection, or is your principal objection much more could be done having added a BMS?

[Living Large]

A Lifpeo4 battery pack doesn't necessarily stay in balance, it depends on a number of things I believe.

What I have been hearing and seeing in videos of users is it tends to stay in balance for extended periods of time (months).

[Sunking]

A Lifpeo4 battery pack doesn't necessarily stay in balance, it depends on a number of things I believe.

I do not think that is true. Large format cells are not paralleled. If they are they are ladder connected. I know a lot of EV guys who run EV's Bottom Balanced and do not have BMS and running fine for 20 months. With everything in series and bottom balanced capacity remains equal.

[Mike90250]

You didn't specifically critique what won't work in what I proposed. I told you one potential issue with your proposed solution - the XW, which appears to be a pretty flexible unit, will shut down using the Aux input. It's definition is "REMOTE POWER OFF". It doesn't suspend activity and wait for the input to change state.

While the XW does not have a good way to easily do remote shutdown or restart (without manual button pushing on a control box) it does have a programmable LVD both time and voltage settings.

If you bottom balance a pack, and set the LVD for something you are comfortable with, it should stop inverting till the battery recovers and then resume.

The top voltage for the charge controllers can also usually be set for 1 minute absorb and then drop back to Float voltage.

Or the BMS can manage a Contactor or Solid State relay and disconnect the battery bank in case of a violation. That would require manual intervention to restart the system.

[Sunking]

I am not stuck in Pb mentality, because remember, I am brand new to this. What I am stuck with is the current state of the art, and that is why I am looking at it the way I am. While a BMS can monitor to the cell level, the question is what can be done with that information. That was the angle I was coming from.

You didn't specifically critique what won't work in what I proposed. I told you one potential issue with your proposed solution - the XW, which appears to be a pretty flexible unit, will shut down using the Aux input. It's definition is "REMOTE POWER OFF". It doesn't suspend activity and wait for the input to change state. So I am stuck with equipment function limitations, and how to operate within the confines. You're the one who taught me this.

I think where we are stuck is you need to make some decisions. So here is my thoughts.

If you are going to use LFP batteries, I highly recommend using some sort of BMS. That BMS will control the charger and a LVD device. The challenge is how to do that with what equipment is on the market. Unfortunately there is nothing really usable on the Solar side for LFP except some toy equipment, but none for serious energy management. Everything for solar is geared for Pb and operates on total battery voltage and not Cell Voltages where LFP has to be controlled.

I assume from this point forward you are going to Bottom Balance as that is the easiest and least expensive way. For charging you are going to have to use the BMS to terminate the charge. For me it is real easy I modified a telecom 48 volt rectifier to accept a logic 1 signal from the BMS when the first cell hits set point. So now we, more specifically you have to determine a way to terminate the charge. I am not familiar with Xantrex aux inputs if that can be done or not. I know MS can be controlled. One work around would be to use a 48 volt 50 amp DPDT relay to open the panels from the input of the controller, and shorting out the panels. You don't even need to short the panels, just open them up and fool the controller thinking it is night time. I do not care much for mechanical operations and one could use an electronic relay. Bottom line is you gotta figure out how you want to handle it.

On the disconnect side you need to figure out how to protect those batteries. Since you are Bottom Balancing and assume you reference at say 2.5 vpc you could set the inverter built-in disconnect to operate at 2.5 volts x 16 = 40 volts. Then reconnect at 48 volts. Personally I would not do that. I would have something redundant. For example what I do on my EV is I have two fail safes and you can employ the same idea. Install a 200 to 400 amp contactor between the battery and inverter. The contactor operates from the BMS signal at say 2.6 vpc first cell which should be a pack voltage of around or 42 volts. Then have the Inverter built in LVD operate at a lower voltage of say 40 volts.

Bottom line here is LFP batteries are expensive. Fortunately LFP are fairly tolerant of Over Charge so that operation is not real critical. Just keep in mind to maximize cycle life you really do not want to go to 100% SOC 90% on the weakest cell is good enough because every LFP cell is somewhat under rated so if you buy 100 AH cells you get something of 101 to 115 AH. What LFP and all lithium is very sensitive to is over discharge and you have to guard and never let it happen. The best and easiest way is to Bottom Balance so all cells reach 2.5 volts at the same time.

[Living Large]

The Aux 2 input on the Classic turns out to suspend charging until its state changes. The active polarity can be set.

[Sunking]

The Aux 2 input on the Classic turns out to suspend charging until its state changes. The active polarity can be set.

Perfect.

[northerner]

The Aux 2 input on the Classic turns out to suspend charging until its state changes. The active polarity can be set.

One could easily build a circuit to have that latch on once HVD triggers, and then have it latch back off again when battery pack drops below a preset level (ie rebulk voltage). If you don't latch it, the charging voltage may get stuck there (at the HVD trigger point) for a while. It's too bad you couldn't trigger the Classic to go to Float mode.

[tasman]

I do not think that is true. Large format cells are not paralleled. If they are they are ladder connected. I know a lot of EV guys who run EV's Bottom Balanced and do not have BMS and running fine for 20 months. With everything in series and bottom balanced capacity remains equal.

You'll have to explain what you mean by paralleled, in my meagre knowledge thought they were all parallel series connections, otherwise you'd never get above 3.x volts out put. Do you mean by ladder connected, that all parts of the cell lines are connected series parallel, instead of just one end and how it would have a different effect for charging.

I believe I've set my MH pack up as full series parallel connections, as it was easier and meant the pack was secure, nice and tight for being in the MH. We've come a cross a number of people with big lifepo4 packs on the road, they don't use a BMS and just meters, manually control their charger rates and have problems balancing no matter how they are connected. That's why I'm considering connecting both negatives and positives ends pf the pack together to get a better distibution of charger and discharge.

[Sunking]

You'll have to explain what you mean by paralleled, in my meagre knowledge thought they were all parallel series connections,

Look at this Stickie, Alternative is ladder connected series/parallel

[tasman]

Look at this Stickie, Alternative is ladder connected series/parallel

Thanks for the links, but it hasn't enlightened me much as can't see much difference between the conventional way and alternative. Here's a photo of my MH pack when I was putting it together, is this what you mean by ladder connections and is this a good or bad way to connect lifepo4 cells. My current setup is of smaller ampergage, found didn't need 600amps in the MH, so dropped it to 480ah and used the other 120ah for a second battery in our sahara which has heaps of elctronics to run.

[Sunking]

The Alternative method you looked at is LADDER connected. That is the proper way to connect lithium cells in parallel as it forces each paring of 2 or more cells in parallel to equalize. It also makes battery management much less complicated and less expensive to implement. For example in the Stickie is shown a 3S2P or 3 cells in series, 2 in parallel. From a BMS point of view only requires 3 monitor points. If you were to use conventional method would require 6 cells to monitor and control. In addition in a conventional method both charge and discharge currents would be unequal forcing the string with lower Ri to do the majority of the work. Connecting them in a ladder style eliminates all that trouble and complexity. For example the Nissan Leaf battery is 96S2P made up of 33 AH cells. Together makes a 355 volt 66 AH battery with a 24 Kwh capacity.

Your picture is quite unique and difficult to make out what you have. It appears you have 64 cells arranged S4P16 12 volts or most likely S16P4 for 48 volts I hope. Either way they should be Ladder Connected as shown in the Stickie.