BMS for LiFePo4 Batteries?

[North of 56]

With all the interest in Lithium Iron Phosphate batteries lately, it would be great to have a BMS for these batteries that integrates with Midnite Charge controllers, and possibly inverters as well?

The problem with current BMS's is they are awkward to implement with current charge controllers and inverters. Also, they are generally designed to cut out at the top end at too high a voltage, which over time degrades the batteries. Ideally, it would be desirable to have the BMS cut out, when the voltage of a cell reaches a selectable state of charge (something between 90 and 98%)?

The other issue is that the charge (and discharge) rate may skew the voltage somewhat, giving a less accurate indication of the state of charge. Perhaps integrating a coulomb counter could help in determining battery state of charge (but not necessarily cell state of charge).

Cutting out at the bottom end from an LVD, you would likely want a warning prior to shutting down your system. That way you may be able to avoid a system shut down if it gets to that state.

Balancing can be done manually, but most BMS's are designed to top balance using current shunts.

[mike90045]

From what I've understood, EV's with high charge and discharge rates, top balance the pack to get the most power, but with low power, long duration apps (solar off grid) bottom balance is better, and then terminate charge at 90-95% .  May need weekly manual bottom balance while the batteries  are being formed and settling in, but then they should be good for a year between fine tuning.   After being bottom balanced, the charge controller manages the top half, and the inverters LVD manages the bottom.  Having an alarm output would be very handy, or a controller or inverter monitor to trigger the alarm. AFIK, not many large, household size inverters provide any alarm, they just shut off !

[North of 56]

From what I've understood, EV's with high charge and discharge rates, top balance the pack to get the most power, but with low power, long duration apps (solar off grid) bottom balance is better, and then terminate charge at 90-95% .  May need weekly manual bottom balance while the batteries  are being formed and settling in, but then they should be good for a year between fine tuning.   After being bottom balanced, the charge controller manages the top half, and the inverters LVD manages the bottom.  Having an alarm output would be very handy, or a controller or inverter monitor to trigger the alarm. AFIK, not many large, household size inverters provide any alarm, they just shut off !

Whether one top or bottom balances appears to be a personal choice. My take on it is that most of the time, there is plenty of solar power coming in, and with a Pb battery system, batteries get fully charged on a fairly regular basis. If that's the case, then batteries are being taken close to the top more frequently than to the bottom. I know with my system, it is very rare that I go down below a 50% state of charge. I know that one can get away with a smaller battery storage size with LiFeP04 over lead acid, but I also think it's important to have a couple days worth of storage, in order to reduce generator run time.

Here is an example of battery usage with LiFeP04: operate normally between 90% state of charge and 20%. This would give you 70% capacity, which would work out to 35% depth of discharge per day for 2 days of autonomy. If one were to operate with this kind of thinking, I think it would make more sense to top balance, as you have only a 10% margin to 100%, whereas you're only occasionally going down to 20%, and have a 20% margin to the drop off. Plus you still would have protection from the BMS if a cell did drop down too low.

My understanding is that you begin to see cell degradation starting at around a 90% state of charge, which could compound over time (for that particular cell) if one goes above that on a regular basis. Only would need to charge right up very infrequently, when doing a top balance.

As for the alarm, it could be built into the BMS to give an audible and/or visual indication that there is an LVD/HVD event. Also I have read that some BMS's have malfunctioned, causing the shunt to discharge the very expensive cell, turning it into a brick. A BMS that is rock solid and dependable would be very desirable for one going the LiFeP04 route!

[mike90045]

.....Also I have read that some BMS's have malfunctioned, causing the shunt to discharge the very expensive cell, turning it into a brick. A BMS that is rock solid and dependable would be very desirable for one going the LiFeP04 route!   

As I'm looking and evaluating my NiFe bank, and planning on the electrolyte replacement, I'm thinking all this through, and while the EV crowd loves their BMS and pushing the pack hard, they can always call a cab for a ride.   If I had a bank at the house, I'd be looking at the highest reliability method to keep the pack healthy, and I don't think the BMS is there yet for the home user, too many ways they can go wrong, but bottom balance and tuning the pack for a couple weeks, then monthly checks, seems to be the most reliable (if the actual measurements and charge adjustments are really done). if a cell does check out and die early, a 3V gap in the data should alert the owner.  But that's just my opinion.

[North of 56]

.....Also I have read that some BMS's have malfunctioned, causing the shunt to discharge the very expensive cell, turning it into a brick. A BMS that is rock solid and dependable would be very desirable for one going the LiFeP04 route!   

As I'm looking and evaluating my NiFe bank, and planning on the electrolyte replacement, I'm thinking all this through, and while the EV crowd loves their BMS and pushing the pack hard, they can always call a cab for a ride.   If I had a bank at the house, I'd be looking at the highest reliability method to keep the pack healthy, and I don't think the BMS is there yet for the home user, too many ways they can go wrong

Good point Mike, and why it would be great to have a robust BMS designed for home use. I don't trust those shunts, used for balancing, as they do get warm when used, and can fail. Perhaps if a failure was detected, a backup would disconnect the shunt, but then again your still relying on something that can fail.

The other issue is that the BMS and the charge controller must be able to interpolate what the charge state of the pack is. That is straight forward with a constant known current, but of course with solar, the current flow varies up and down considerably at times. I don't think one can rely on voltage alone in order to do that, unless you bring the pack up to a much lesser charge state, but then you're losing potential pack capacity, in times when you may need it.

[Graham]

Came across this http://electrodacus.com/ on kickstarter the other day, they are using the intersil chip to monitor the cells, it also uses the ESP8266 for Wifi.

https://www.intersil.com/en/products/power-management/battery-management/cell-balancing-and-safety/ISL94203.html

[Westbranch]

I thought I had seen that name before...  he posted on NAWS about a year ago.

Here are some development related comments on NAWS...http://forum.solar-electric.com/forum/solar-electric-power-wind-power-balance-of-system/advanced-solar-electric-technical-forum/21654-my-programable-solar-bms-for-any-lithium-cells-and-supercapacitors

[North of 56]

Looks like this fellow has created something that will suite his needs, but likely too small for many house sized off grid systems.

I think that current charge controllers can be used for LiFeP04 charging, and is something that I along with many others already have in place. What is needed is a monitoring system (ie BMS) that integrates with both charge controllers and inverters or load control. IMO, there are 3 parameters useful for determining the charge state of the battery pack and all the cells.

1) individual cell voltages of applicable cells

2) level of current flow into/out of the pack

3) the approximate charge state of the pack (measured with an AH counter)

One modification to current charge controllers that would be very useful for this application, would be the ability to signal it to switch modes, (ie to go to float for example)

[Westbranch]

well according to his site page, http://electrodacus.com/  , he is adding a new model , 100A, that is 20A  bigger than its predecessor. It will handle max 3KwH from an array and 3Kwh load, don't quite follow what he meas by the load statement. Also looks like a VoC of 120A and good for a 400Ah cells 8 cells max.

He has made some god progress over the last year...

[RossW]

What is needed is a monitoring system (ie BMS) that integrates with both charge controllers and inverters or load control. IMO, there are 3 parameters useful for determining the charge state of the battery pack and all the cells.

1) individual cell voltages of applicable cells
2) level of current flow into/out of the pack
3) the approximate charge state of the pack (measured with an AH counter)

18 months ago, I built a prototype "low cost, expandable" universal battery monitor ("universal" in that it's designed for 2V lead-acid and ~3V LFP cells) in virtually any configuration.

Each board monitors (up to) three cells, and the boards simply cascade (daisy-chain) one to the next. One board includes an optically-isolated RS232 interface for configuration and sending data and alarms. Each board will also take 0-3 temperature sensors, so it can monitor individual cell temperatures and voltages.


The boards are intended to be distributed along the battery pack, but some folks wanted to have the electronics in one spot and run cables back, here's a 24-volt board assembly (only partly assembled)


Here's a live 48V nominal system (16 LFP cells) (this is a prototype, and I didn't calibrate it terribly carefully, hence the apparent cell voltage differences)


I've considered how to monitor cell current, and have nutted out what I think will be a simple and effective way, just haven't built or coded one yet. All I really need is some clear time to finish a few tweeks, and some sort of packaging for it.

Basic specs are for it to monitor each cell, and to generate an alarm if anything gets more than a user-defined number of millivolts from the total battery per-cell-average-voltage. (means it can alarm on excess voltage difference while on charge, discharge or static conditions).

[Cniemand]

I haven't seen a real need for Active Balancing on LiFePo4 cells. I have 762 days on my full time, Off Grid battery bank. It is quite boring in how it just works and doesn't have cell drift. :/

I am with Mike. Initial bottom balancing the pack once prior to usage and charging them to 3.5v per cell means you get most of the useful capacity and none of the headaches with runaway cells.


- Cloud