I recently had recall work done to my car that required access to the fuel tank. Guess where the fuel tank is accessed? That’s right, it’s accessed by removing the rear seat! If you’re familiar with my car, then you know that I don’t have a rear seat. Instead, I have a 1280 watt-hour lithium iron phosphate (LiFePo4 or LFP) battery where the rear seat once was. There’s no way that I’d let anyone else work on this battery, much less remove it! Removing it requires the tedious unclipping and disconnecting of dozens of wires, as well as extra electrical safety precautions when working with secondary power sources. Reinstalling it requires even more attention to detail! I postponed removing the battery until I knew that my travel schedule would be clear.
I took about two days to remove everything, label the wires, and bag the loose bundles of wire. The recall service was completed a week later, then I took several days to reinstall everything. I upgraded some of my fuse holders and wiring as I worked, including circuits for the 50A in/out of the DC-DC charger, the 100A feed to the trunk for radio equipment, and a 150A circuit for the 3000-watt pure sine wave inverter (self-limited to ~1500W). This schematic shows an overview of my 12V distribution. A close look reveals that all of my circuits combined can exceed the power limit of my 120A battery management system (BMS). I’m not concerned because I never use all the circuits to their capacity at once; plus, the charger adds up to 50A to my overhead when the engine is running. Have a look at my 12V Power Distribution page for more details.
With the battery reinstalled, I topped-off the charge, parked the car, and performed an annual load test, which usually takes about 10 hours at 10 amps. The battery had measured 100 Ah or more in years past. This year, I was stunned to measure only 79 Ah! I checked my BMS and charger settings, made minor adjustments, recharged the battery, and then ran another test. The result was still just 80 Ah! I was stumped! However, this time I noticed that cell #3 was reaching its low voltage disconnect of 2.5V far before the others at the end of the test, as shown in this image. The BMS stops everything when any cell reaches a disconnect limit, be it voltage, current, or temperature. I consulted with some experts in the DIY Solar Forum. They helped me realize that cell #2’s voltage was higher than the rest when charging, causing the BMS to terminate charging when it reached its high voltage disconnect of 3.65V, all without indicating a problem. I was hung-up on the fact that cell #3 was low enough to terminate the load test early, but I followed their advice about cell #2.
One user recommended manually balancing the individual cell by using a 3.3-ohm resistor and some jumper leads. I hadn’t thought of it before, but manually balancing is basically doing the BMS’s job, only 20 times faster and more deliberately. The forum also shared Overkill Solar’s BMS app, shown in this image. The app is more stable than the Xiaoxiang app, shows me when a disconnect condition has occurred (see the red text in this image, an indication not given by Xiaoxiang), and it allows me to view temperatures in Fahrenheit… less math! 😉 The resistor trick was helpful. I bled cell #2 to match the lower cells, which led me to bleed cell #1 as well. I quickly realized that cell #3 continued to be lower than the rest. I decided to charge the battery and run another load test, but still achieved only 83 Ah. Clearly, cell #3 was the problem. The flaw with manual balancing is that I was bleeding cells to a state that’s somewhere in the middle of a very wide state of charge (SOC) that’s around 3.35V. Is the SOC at 40%, 70%, or somewhere else? I needed a way to be more certain.
I found an alternative top-balancing method in Overkill Solar’s BMS manual. It’s safe to charge each cell to 3.65V individually without disconnecting the cells. I hadn’t thought of that before. Sure, it’s not as effective as charging all cells together in series, but it’s far less laborious than removing and disassembling my battery. I started with a full charge, at least to the point where one cell reached over-voltage cut-off. From there, I charged each cell individually to 3.65V, starting with the lowest, cell #3. Cell #3 took over eight hours to stabilize at 3.65V. Then I charged the remaining cells, which took much less time per cell. Next, I put my power supply across the entire battery to stabilize it at 14.5V. Then I ran another load test. This time, the result was 92 Ah. That’s not as high as I had hoped, but it’s not bad when considering some of the abuse that I may be giving the battery!
Overall, I’m still pleased with with my self-built battery. I’ve used it from 15°F to 110°F with almost no rest, including the nearly full-time use of a refrigerator! Heat is detrimental to battery cells. There’s not much I can do to improve that in my passenger car; cell degradation will happen over time, heat or not. So, I just charge to 100% and run to 0% if it’s necessary. “Use the capacity that you have today,” as recommended in this video by Will Prowse. Will is highly regarded as a battery expert in the DIY solar community. However, I’m not currently following his charging advice due to my cell imbalance.
Instead, I’m using recommendations that are based on a series of comprehensive charging tests by “Off-Grid Garage.” Will Prowse advises to charge to 3.625V per cell, or 14.5V in a 12V system. However, since my #2 or #3 cell still triggers a high-voltage disconnect when I attempt to charge to 14.5V, or even to Victron’s preset of 14.2V, I have set my Absorption voltage to 13.9V, Float voltage to 13.6V, and Storage voltage to 13.4V. The lower charger voltage extends the Absorption stage to allow more time for the BMS to balance the cells, as does the slightly elevated float voltage, all while still charging the battery to nearly 100%. This image shows the battery at 13.8V with a small current still working into the charge and cell #3 being balanced by the BMS. Cell #3 quickly settles to match the other cells when the engine is off or when the charger enters Float, shown below.
I have since done several charge cycles with the lower Absorption voltage and have been pleased. Charging to 13.8V has been successful and repeatable enough that I have raised the Absorption voltage to 13.9V, or 3.475V per cell. This setting still requires some tweaking because the charger enters Float a little sooner than expected, particularly if the car has been parked for less than about two hours with a heavy load. This image shows that the charger is in Storage mode, meaning that it’s holding at ~13.4V with no power in or out of the battery. Radios and other electronics are powered by the charger and the battery cells are within 10mV of each other, which is an excellent balance between the cells.
This challenge revealed a benefit of having a user-assembled battery that features accessible cells and an external BMS with Bluetooth monitoring: SERVICEABILITY. Had this happened to a sealed off-the-shelf (OTS) unit, then the capacity might have been unrecoverable, which would probably hasten the battery’s replacement and trip to a recycler. This experience could cement my belief that component battery builds are the way to go, even though they’re more expensive. Then again, my $800 component battery can be functionally replaced by an OTS battery that costs less than $200. It’s difficult to ignore the “money talks” aspect of a purchase when energy is so cheap to acquire. Is serviceability worth quadruple the price? Are the cheap OTS batteries built to last? Are most OTS batteries destined to suffer from premature cell imbalances?
I’m temped by bargain OTS units, but I don’t know of any with the same slender form that I’ve created to fit inside my shallow “basement” that’s beneath my cargo area, as shown in this photo. With better treatment of my current battery, and perhaps some luck, I hope to see it last long enough that I don’t have to contemplate its replacement until I’m ready to move to another vehicle. When that happens, I’ll probably chose a car that will fit a taller battery and then get a 300-Ah battery of some sort. “Go Big or Go Home!” 😀
This entry is longer than usual. I wanted to share as much detail as I knew about the issue at hand, the troubleshooting and investigation that I tried, and the methods that I used to overcome the challenge. I hope it was worth the read.
Power to Spare!
Scott
