LiFePO4 DIY options

[lfhoward]

Capacity test time!

The best I can do is about 13 amp draw which is closer to 0.1C with the inverter I have. So I might expect to get a little more than rated capacity (156 Ah) at that rate. We will see in about 10 hours. (It’s running 3 lamps and two air filters from the inverter.)

The battery test setup


The loads


The Victron BMV says 13 amps


Where we started on the Overkill BMS — really nice balance!

[saltydawg]

You also piggy back some 12 volt loads on it if you want to do it faster. Have a spare head light bulb, thats another 55 to 85 watts or another 4.25 to 6.5 amps. Break out your fan ( 12 volt roof fan ) and wire that up, depending on fan you could get another 6 amps or so there.

[lfhoward]

Scott, thanks for the idea of adding DC amps to the load. I wired up a pair of light pods that I hadn’t yet installed on my Jeep and we are now up to a total of 16.31 amps draw. The time remaining is down to 7 hours and 20 minutes. Nice!

[lfhoward]

Capacity Test Results:

The inverter shut down at 11.0 volts, and at that point the BMV registered -152.5 Ah. The LED light pods, directly wired to DC, shined on drawing 3.24A. I was able to restart the inverter minus the air filters (lights only) for a 7.7 total amp draw including the light pods. The BMS showed one cell getting a bit lower than the others at this point, and it should shut down the test when a cell reaches 2.5V.

The Overkill BMS shut down the test when cell #2 hit 2.5 volts. At that point, the battery had given up -153.4 amp hours. That’s only 2.6 Ah less than advertised, so I’m still very happy. It’s over 150 Ah.



Next, I need to charge this battery back up. I configured my power supply to mimic the volts and amps of a 200 watt solar panel and fed that to the Victron 75/15 charge controller.


If discharging at 17 amps took 10 hours, charging at 15 means this won’t be ready until morning.

[bdosborn]

Your cells went pretty far out of balance at the end but I think that's the way it works when you top balance versus bottom balance. It's great that the BMS disconnected on low cell voltage instead of low battery voltage. Watch them as you charge them back up and make sure they're balanced at the end of charge. My 400 a-hr battery cells are only 0.003V difference when fully charged and after an hour or so of resting, 0.009v for my 360 a-hr battery.
Bruce
P.S. Ha, you went Victron, so did I. Now you have to get a Raspberry Pi so you can run Victron connect.

[saltydawg]

Probably too late to do it, I would have looked at the individual cells. You might have been able to swap some pairs around and even out the voltage at the low end. It could also just be one cell that pulled its pair low.

If you could find that one low cell your close enough to the seller you might be able to swap it out.

[lfhoward]

Bruce, I’ll let you know what the delta is when we get charged back up to full. The charge level is now in the 70% range so I should know by the end of the day.

Scott, that is a good point. I wish I had thought of that at the time. I could do another run down to zero and see if the #2 pair is low again. At that point I could disconnect everything and look at the individual cell voltages. It could be one cell is lower then the other. How long after uncoupling the parallel connection would the difference be obvious? I wouldn’t want to leave the batteries that discharged for very long.

If I did decide to swap out a cell, I would have to do the top balancing process over again, wouldn’t I? Is it worth it for 3 amp hours? On the other hand if I paired the low cell with a high one to even it out, it is already top balanced.

[lfhoward]

Actually, thinking about last night’s results…



I could swap one of the cells in block #2 & #3 as soon as the pack is fully charged, then do another discharge test. If one of those blocks hits 2.5 volts much earlier than the rest of the pack, then we know we have a bum cell and we know which one it is.

[lfhoward]

Bruce, after recharge the cell profile looks like this:



A delta of only 0.003 is pretty great for top balanced cells.

Scott, I am swapping a cell from block #2 with one from block #3. Once I get it back together I’ll discharge it again and we’ll see if the delta is as high when the first block hits 2.5 volts.

[lfhoward]

Prior to the second discharge test, while swapping the batteries as noted in the previous post, was a good time to install the beehive heaters.

Heaters on black silicone trivet material:


Aluminum floor wrapped in Kapton tape to distribute the heat:


A slot for the 3/16” diameter temperature probe. I put it near the edge because this outside battery is probably the most susceptible to cold.


Cells back in place. I got the height of the box nearly right for the heater sandwich + batteries.

[saltydawg]

Swapping the cells is a good idea and a good way to find if one cell is weak. You could also make 2 packs and test them again and that will let you know which cell is weak.

Personally I would do the 2 pack test, you would not need to to balance them. Just hook em up and let the bms manage the charge. Then drain them and find the weak cell. You had enough of a delta finding it should be easy and not require over the top testing.

[lfhoward]

Swapping the cells is a good idea and a good way to find if one cell is weak. You could also make 2 packs and test them again and that will let you know which cell is weak.

Personally I would do the 2 pack test, you would not need to to balance them. Just hook em up and let the bms manage the charge. Then drain them and find the weak cell. You had enough of a delta finding it should be easy and not require over the top testing.

Thank you for the idea. I already got it set up with 8 cells and wires all tidied up. The end of the test will be in approximately 8 hours. I have 19.3 amps going.



Since the former block #3 had the highest voltage at the end of the test last night, I’m assuming both of those are good. By swapping one of them with a cell in block #2, I can see whether block #2 improves (in which case the bad cell is still in #2), or if it stays the same and #3 gets worse (meaning the bad cell is the one I moved to #3). Another option is everything ends up relatively equally, and swapping cells just averaged everything out. This would be the best case scenario because there would be no really bad cell.

[lfhoward]

My battery has an emergency shutoff switch right here:



It turns the BMS off which cuts the power. It doesn’t replace the physical battery disconnect that is between the battery and positive bus in the trailer. But it’s nice to have two ways to kill power in an emergency. The BMS off switch will also be handy if storing the trailer for long periods without solar (like in a barn) because it eliminates parasitic draw.

[bdosborn]

Swapping cells is what I did to balance the cells better:



http://www.tnttt.com/viewtopic.php?f=30&t=73606&hilit=+lifepo4

But your cells are balanced well at the top, I wouldn't change them unless they go out of balance after a few charge/discharge cycles.

Bruce

[bdosborn]

My battery has an emergency shutoff switch right here:

I'd still put a battery terminal fuse on it.

MRBF

I've looked at the fuse time-current curves and its similar to the class J, fast acting fuses we use in commercial design. The fault current withstand rating is also high enough for a direct LiFePo4 short circuit. The fuses are kind of expensive so don't get careless with a screwdriver and blow them out.

Bruce