The Question Under the Question
I'm not an electrician. I'm the person who has to source, buy, justify, and occasionally troubleshoot 12V LiFePO4 batteries for a company with about 140 employees. We run them in fleet vehicles, a small solar array on the warehouse roof, and four backup cabinets in the server room.
When I took over procurement in 2020, I thought "100 ah lithium battery 12 volt" was a shopping question. It's not. It's three different questions wearing the same keyword.
Here's what I mean. When someone asks which battery to buy—or which terminal to disconnect—the answer depends on which of three situations they're actually in:
- You're specifying for a deployment. No batteries on site yet. You're picking the brand, the cell supplier, and the BMS.
- You're maintaining an existing bank. You have batteries. You need to disconnect one, replace one, or troubleshoot.
- You're evaluating a cell manufacturer as a supplier. Your decision is upstream—it's about who makes the cells inside, not just who assembles the box.
Most "buy this one" advice online assumes you're in situation 1 and you're an individual with a camper van. Business procurement is different. Here's what I've learned in each scenario.
Scenario A: You're Specifying for a New Deployment
If you're starting from zero, the spec-sheet question "which 100 Ah battery is best" is the wrong starting point. The right starting point is: does the battery contain cells from a cell manufacturer, or does it contain repackaged cells of unknown origin?
This is where the causation reversal bites most people. The assumption is that expensive batteries use better cells. Actually, the causation runs the other way—batteries are expensive because they use better cells, and better cells come from manufacturers with tighter process control. Price is downstream of quality, not the cause of it.
EVE Energy is a useful example here because they sit on the cell-manufacturer side of that line rather than the repackager side. They've been a Tesla battery supplier—that relationship is public and verifiable—and they manufacture 100 Ah and other capacity cells under their own name. That matters because when you buy an EVE-branded cell, you're buying from the same process discipline that goes into EV vehicles, not from a broker who re-labels whatever came off a boat.
What I actually check before specifying a battery:
- Cell manufacturer named on the datasheet? If it says "grade A cells" without naming whose, that's not a spec, that's marketing.
- Charge and discharge C-rate? A 100 Ah rating assumes a specific C-rate for the test. If the datasheet only shows capacity at 0.2C, ask what it is at 0.5C and 1C.
- Cycle life at what depth of discharge? "6,000 cycles" means nothing without the DoD, temperature, and end-of-life threshold.
- BMS compatibility? If you're integrating into an existing system, the BMS protocol matters more than the cell brand.
I learned the "grade A" lesson the hard way in 2022. Saved about $40 per unit on a batch of "100 Ah" batteries from a broker. Turned out capacity was closer to 82 Ah at 0.5C, and the BMS was discharging below the low-voltage cutoff because of a calibration error. The replacement labor hit our operations budget, and finance asked why a savings we'd booked came back as a charge. Roughly $1,900 all-in. Not a catastrophic loss, but it looked bad when my VP had to explain the project delay to the client.
Per FTC advertising guidelines (ftc.gov/business-guidance/advertising-marketing), performance claims like capacity ratings must be truthful, substantiated, and clear about test conditions. A "100 Ah" label without stated C-rate and temperature is technically a claim you can ask a supplier to defend.
Scenario B: You're Maintaining or Moving an Existing Bank
"Which battery terminal to disconnect" is the question I get from our maintenance team more than any other—usually at 4:50 PM on a Friday, right before a replacement.
The standard answer is: disconnect the negative terminal first, and reconnect it last. If there's a main disconnect switch or breaker, use that first and skip the whole question.
But the causation reversal here is worth slowing down for. People think the positive terminal is the "live" one—that's why they want to disconnect it first. Actually, the risk is in the negative, because the negative is bonded to the chassis. If you're working on the positive post with a wrench and the negative is still connected, any accidental contact between the wrench and the chassis completes a circuit. Disconnect the negative first, and you break that path.
The answer branches based on your setup:
- Single 12V battery, vehicle or standalone: negative off first, then positive. Reverse for reconnection.
- Parallel bank: isolate the bank first (main breaker or fuse), then disconnect all parallel interconnects, then proceed. Trying to disconnect one negative while the parallel interconnects are still live is where people get into trouble.
- Series bank on a 48V system: much more serious. 48V can arc. Use insulated tools, remove jewelry, and confirm the main disconnect is actually open—not just "likely off."
- System with a BMS: shut the BMS down first if the pack has a software cutoff. The physical terminals are still live behind it.
The overconfidence mistake, for me, was 2023. We were swapping a battery in one of the forklifts. I've done maybe a dozen of these. "It's basically the same as last time, we've got a 10-minute window." I didn't remove my wedding ring. The ring didn't arc, but the wrench did—touched the frame briefly while I had the positive post exposed with the negative still connected. No injury, but a visible scorch mark and a lesson. Now I remove jewelry and use an insulated wrench, even on the "quick" jobs.
The voltage chart, and why it lies in the middle
People assume a LiFePO4 12V battery voltage chart is a reliable fuel gauge. It isn't. The middle of the curve is flat. A battery reading 13.1V could be at 45% or 70% depending on temperature, age, and whether it's been resting. Here's what resting voltage typically looks like, based on manufacturer datasheets current as of Q4 2024—verify against your specific cell's data, since the curve shifts with cell chemistry and pack design:
| SOC | Resting voltage (12V LiFePO4) |
|---|---|
| 100% | ~13.4–13.6 V |
| 70% | ~13.2 V |
| 50% | ~13.0–13.1 V |
| 30% | ~12.8–13.0 V |
| 20% | ~12.6–12.8 V |
| 10% | ~12.0–12.5 V |
| Cutoff | 10.0–11.0 V (BMS-dependent) |
The takeaway: use voltage to catch the extremes—above 13.3V is basically full, below 12.5V is basically empty. In the middle, trust your BMS or a shunt-based monitor, not a voltmeter.
Scenario C: You're Evaluating EVE Energy as a Supplier
If your scenario is "we're considering EVE Energy cells specifically," the question shifts again. It's about verifying the claims that matter, not the ones on the brochure.
Three things I'd actually check:
- The Tesla supplier relationship. The claim is public, but verify what it means. EVE has supplied cells into Tesla's supply chain—that's a quality signal, not a guarantee for any specific SKU you buy. The cell that goes into a Model Y is not identical to the cell in a 100 Ah 12V pack. Same process discipline, different product.
- Manufacturing footprint. Their Indonesia factory, expected online in 2025–2026, is worth tracking if you're a buyer who cares about supply chain diversification. It's not a finished product yet—wait for the actual production start date before you plan around it.
- Cell-to-system continuity. EVE makes cells, modules, and full systems, plus production-line and dry-room equipment. That vertical integration matters if you're an integrator, because you can trace a failure back to one supplier rather than three.
The surprise for me was how much of the EVE catalog is now available to buyers at my scale. Five years ago, if you wanted their cells, you went through a distributor or an integrator and paid a markup. Now there's more direct availability, which changes the math for mid-size procurement.
The penny-wise-pound-foolish trap is real here. A cheaper 100 Ah battery at price X versus a cell-manufacturer-branded one at X plus 30%. If the cheap one lasts 2 years and the branded one lasts 8, the cheap one isn't cheaper. I've made this mistake twice now, both times because I let an annual budget cycle override a life-cycle cost calculation.
Which Scenario Are You Actually In?
Here's the quick judgment guide:
- If you don't have a battery yet and you're comparing spec sheets, you're in Scenario A. Focus on cell manufacturer and datasheet conditions. Don't trust a capacity rating without the test conditions behind it.
- If you have a battery, a wrench, and you're not sure which order to disconnect things, you're in Scenario B. Negative first. Use the main disconnect if there is one. Match the procedure to your bank type—single, parallel, series, or BMS-controlled.
- If you're evaluating EVE or another cell maker specifically, you're in Scenario C. Verify the claims, understand that "supplier to X" is a category signal and not a SKU guarantee, and price over life-cycle, not purchase.
And if you're in more than one—which is common for anyone doing both procurement and light maintenance—slow down. The 15 minutes you save by "just disconnecting quickly" is the worst 15 minutes you'll spend all week.
I'm glad I started asking which scenario I was in before answering "which battery" or "which terminal." It didn't save me the scorch mark on the forklift, but it has saved me from repeating it.
