Hence, the beginners corner. Running without BMS

[GerardJacinto]

My LiFepo4 setup is 8 cells of 3.2V on a 12V connection with a 4S Daly BMS. It was top balanced, so they all go up at 3.6, pack voltage of 14.4V. I happily charge them using my benchtop power supply everytime they go low ( made it a habit of charging them back up once they hit 12.6V ) as I dont have my solar panels yet. They came in handy when 2 typhoons ran over our country in just a matter 2 weeks. Using a battery monitor during my usage from full until BMS decided to disconnect them from the inverter, they came up as a 700-ish AH battery. They were advertised as 2nd hand, 90AH cells. Paid dirt cheap for them, around $28 per cell, so I have no complaints. They work. I love them.

But with 8 cells, I was curious to get a 24V system instead, so I bought a 24V inverter. A 2000W rated / 4000W peak. I also ordered an 8S Daly BMS with sufficient Amp rating. A few days ago, I received the inverter. I was keen on testing it, but held back because I don't have an 8S BMS yet. Its still on its way. Curiosity won though, and so I reconfigured my batteries in series, making a 24V pack and wired up the inverter and bam! Works like a charm. Plugged in a 2000W electric kettle, boiled a liter of water for 3 mins. Voltage was 25.7, dipped down to 25.5 during the test, then settled to 25.6 I think. I was happy. Kept the 24V setup like that. Used the batteries more while on 24V config. Without BMS. Big mistake. Dammit.

As I used it even more, pack voltage went down to 23V. I thought it was still ok, as that meant each cell still had 2.8V on them. Or did they? Turned out, one cell was already lower than 2.8V. Battery dipped even more, this time at 22V. I prepared to turn off my inverter to let the battery pack settle, and prepared to reconfig them to 12V so I can charge them all up again. To my horror, my once perfectly aligned Sinopoly batteries, now has a grand canyon-wide gap between one of them. I am exagerrating of course. Maybe half an inch gap. Didn't bother to measure. So I quickly turned off my inverter. Removed all connections, grabbed my multimeter and tested each. All of them were around the 2.7-2.9V ball park, except for one. The preggy battery clocked in at 1.9V. Damn. Will that cell still work? I don't know. I read somewhere that going below 2.5V is bad. And it visibly did bad on the battery.

So, I ordered another one through our online store here in our country. Lots of sellers selling these Sinopoly batteries. I'll just top balance them again once the new one arrives. For the mean time, I wired everything up again in 12V, connected my BMS, and charged the pack again using my benchtop power supply. They are still charging for almost 4 hours now, and is now sitting at 13V. Haven't check each cell yet. I am hoping I can still use it though. If not, backup is on its way.

This all happened a few hours after I finally sealed the deal with a local solar panel distributor that happens to be a neighbor. Imagine my heartbreak if 2 or more cells went pregnant. I am on a tight budget, and could not afford buying 2 or more replacements.

Lesson learned, do not run those dirt cheap, 2nd hand batteries without a BMS. And don't go for cheaps if you can. I did a lot of research about building this setup, but maybe not enough. So research first. Learn, and ask. Don't be like Gerard.

 

[Samsonite801]

Some people here say it isn't good to run mismatched battery cells, as they might recommend for battery shoppers to request for sales vendors to do battery internal resistance testing and capacity testing each cell, so they can match cells of like characteristics into a batch for a bank before they ship them out.

Now I only know how this is relevant in theory as I have not had to replace a single cell myself yet, but I had contemplated buying me a battery internal resistance tester to have in my tool chest just to learn a bit about checking it. And I suppose some could say if you replace a single cell, then it could be a good thing to check the new cell and existing cells to confirm it's not way out of wack from the batch.

Now I don't really know the reality of how important this might be, as it could just be one of those really 'blueprinting' type steps that are just for people who want maximum performance and efficiency from their bank, maybe if people are only running 20-80% SOC charge cycle then they might never notice the difference if they had a new cell out of whack from the batch spec, but I just thought I point it out, and maybe others more experienced in this could comment on whether this really is that big of deal to worry about in the real world of normal use folks.