diy solar

diy solar

Debugging new LFP battery checklist (sticky?)

fafrd

Solar Wizard
Joined
Aug 11, 2020
Messages
4,188
I on the cusp of assembling my second LFP battery (8S 280Ah) after bungling the assembly of my first ‘learner’ 8S 90Ah LFP battery before I knew about this Forum.

From the incredible amount of information and knowledgeable DIYers with experience, I am feeling much more confident of this second build.

There are many useful sticky posts here summarizing important information for neophytes, but one I have not seen is a checklist of things to check for / debug a new battery build.

I’m hoping those with more experience can flesh this list out but thought I would start the ball rolling by summarizing my current plan to debug this new 280Ah battery based on things I’ve learned here from various threads:

i/ it is assumed that all cells have been properly top-balanced and are at 100% SOC.

1/ make terminal connections: assure terminals are free of residue and clean just before assembly (with alcahol?). Assure busbar surfaces are flat (no burrs) and clean as well. Assemble busbars onto terminals tightening to 4Nm (35inch-lbs) [this is assuming standard M6 threaded holes in aluminum terminals]. BMS harness and any other harnesses for monitors and/balancers should also be assembled as part of this step.

2/ attach BMS: Attach ground path between BMS and ground terminal, assuring clean and flat connections as above (and that wires used are sufficient AWG, meaning AWG at least as big as busbars and probably a notch higher). Likewise attach ground wires to BMS and battery ground terminal. Attach harness connector and verify BMS is connecting battery by measuring voltage between ground terminal and positive cell terminal.

3/ Attach fuse and positive terminal. I build my fuses directly into the battery + path so that aspect may be optional for others who use external fuses for their battery, but assemble properly-sized conductors from positive cell terminal to fuse and from fuse to battery positive terminal.

The battery is now assembled and here is the checklist of things I plan to check:

A/ DISCHARGE TEST: connect whatever load(s) are available to discharge battery at maximum rate of interest (I’ll be using 80A or 0.29C) as my basic discharge). Measure and record terminal voltages with DMM then begin discharge and measure and record new terminal voltages. If any cells drop much more than the others, either that cell has higher internal resistance and is likely defective or there is a high-resistance busbar connection on that terminal (will safe debug steps for another post/thread). Assuming all cells are behaving similarly, continue discharge towards 0% SOC.

B/ HOTSPOT CHECK: Record temperatures at various points on battery including all busbars, BMS and FUSE in and out pads, + and - battery terminals, and each cell (near both terminals as well as on top and sides). Any spots which are at a much higher temperature than their brethren indicate a high-resistance connection to be debugged. If one cell is increasing in temperature much more than the others, that likely indicates a defective cell (high IR?).

C/ DISCHARGE STRESS TEST: I plan to stress my battery to the highest discharge rate it will ever see in use (224A or 0.8C in my case) at which point I will record voltages and temperatures, and let it run for 30-60 minutes checking temps at regular intervals.

D/ LOW VOLTAGE DISCONNECT TEST: after the battery has been drained to minimum voltage of whatever equipment it is connected to, I plan to connect the battery to my 150A capacity tester and discharge all the way to 0% SOC. I will monitor the weakest cell all the way down to the LVD of my BMS at 2.1V to confirm that the BMS disconnects the battery under this condition.

E/ CHARGE / TOP-BALANCE TEST: recharge battery to 100% SOC and check quality of top-balance once charge complete. If one cell hits High Voltage Disconnect begore the others, it indicates either a poorly-executed top-balance or a problem cell (save debug discussion for another thread/post).

F/ HIGH-VOLTAGE DISCONNECT: if battery is well-balanced, LFP charger will stop charging before any cells have hit HVD (my 28.6V 8S LFP charger will stop at 3.575V/cell if my top balance is perfect and my cells are well-matched). So to test the BMSs HVD function, I will disconnect the LFP charger and continue charging with my bench supply set at 30V / 3.75V per cell and continue charging. My BMS has HVD hardwired to 3.7V so your values may vary accordingly but I want to verify that when one or all cells hit 3.7V that the BMS disconnects the battery as it should. I will probably adjust the current limit on my bench supply to a low value of <1A to protect against the possibility of a BMS failure resulting in damaged cells. Since the channel precision of my BMS is 5mV, I’ll probably let the cells go all the way to 3.706V before stopping the test.

G/ OTHER BMS TESTS; Low Voltage Reconnect and High Voltage Reconnect functions can be tested in similar manner but since they are unrelated to protecting the cells, I don’t see the need to test them.

H/ HIGH-CURRENT CHARGE TEST: I don’t have any ability to charge at greater than 20A at the moment and so cannot perform a high-current charge test (at 80A / 0.29C in my case) until my PV-based SCC is built later this year. Because my discharge test was performed at this same current-level, I believe any connection issues should be exposed by the discharge test and see it as unlikely that some new issue will be exposed by charging at full current.

I/ FULL-BATTERY CAPACITY TEST: because I don’t have a shunt and assessing capacity through AC consumption isn’t straightforward, I will use my little 150W capacity tester to measure full-battery capacity once the other battery tests are complete (leaving the battery at 100% SOC). At 5A, that test will take ~56 hours but I’m not in a rush. Once the discharge test is complete, I will recharge the battery to ~50% SOC to measure cell width and start planning for final installation in my 300kgf enclosure/jig (at which point all hotspot/connection tests will need to be repested).

So in in summary, I plan to check for any poor connections using voltage and heat, as well as to characterize cell performance under load and confirm BMS functionality (all steps I skipped in my first ‘learner’ battery build)). Are there any other important tests to perform before a battery is put into service? Are there other means to check for properly-made connections other than using voltage and heat?

Any advice to complement this list and/or criticisms appreciated before I start the build (cells just reached 3.65V and I’m going to let them rest overnight before beginning the build tomorrow...).
 
Based on what I've done, it sounds well thought out to me. May I ask what went wrong in your first build, your "gaps" and consequences?
 
D/ LOW VOLTAGE DISCONNECT TEST: after the battery has been drained to minimum voltage of whatever equipment it is connected to, I plan to connect the battery to my 150A capacity tester and discharge all the way to 0% SOC. I will monitor the weakest cell all the way down to the LVD of my BMS at 2.1V to confirm that the BMS disconnects the battery under this condition.

2.1v disconnect is lower than what I've seen. 2.5v seems more common and that's also what the spec sheet on my lishen says for floor of the discharge test. I assume it might vary a bit depending the cell specs, but 2.1v sounds low.
 
Based on what I've done, it sounds well thought out to me. May I ask what went wrong in your first build, your "gaps" and consequences?
The list is too long to detail.

I only checked for similar voltages and did not too-balance to start.

I didn’t capacity-test the cells before building.

I wasn’t careful about cleaning terminals and checking resistivity of connections.

I did run a capacity test on the full battery and confirmed all functions of the BMS, so that is one thing I did right, but that set of cells has a weak cell that needs to be replaced and I did not discover that until faaar to late into the process.


I’ll eventually circle-back and rebuild that ‘learner’ battery (which largely served its purpose ;).
 
2.1v disconnect is lower than what I've seen. 2.5v seems more common and that's also what the spec sheet on my lishen says for floor of the discharge test. I assume it might vary a bit depending the cell specs, but 2.1v sounds low.
I agree, but probably manufacturer specific... Anything under 12 volts and most rv's and inverters go bonkers...
 
2.1v disconnect is lower than what I've seen. 2.5v seems more common and that's also what the spec sheet on my lishen says for floor of the discharge test. I assume it might vary a bit depending the cell specs, but 2.1v sounds low.
I’ll be programming my inverter to disconnect at 22.5V, so ~2.8V if all cells are well-balanced, but the Heltec BMS I have is hard-wired for 2.1V LVD. I’m comfortable this will never be exercised in actual use unless there is a failure but I want to assure the LVD function operates correctly.
 
I agree, but probably manufacturer specific... Anything under 12 volts and most rv's and inverters go bonkers...
You guys did see that I specifically stated I would be making that LVD test with my capacity tester, right?

In actual operation, the inverter will have disconnected long before that.

The BMS LVD is just there as a fail-safe and figuring how to check that that protection functions is performing as expected is non-obvious.
 
Back
Top