How I Learned to Stop Guessing and Start Checking: A Battery Engineer's Confessions

I've been handling battery system orders for about ten years now. I've personally made (and documented) a lot of significant mistakes—enough to fill a small notebook with cautionary tales. My title might be 'Senior Project Lead,' but my real job, the one I've earned through thousands of dollars in wasted budget and missed deadlines, is maintaining our team's checklist. I'm the guy who writes down the things we should have known and makes sure no one repeats them.

Everything I'd read about battery systems said the biggest challenges were chemistry and cost. In practice, the most expensive errors I've seen have nothing to do with cathode material or energy density. They're about assumptions—assuming a spec is standard, assuming a cable is disconnected in the right order, assuming a customer knows what '48V' actually means for their setup.

The Solid-State Bet That Almost Broke Our Timeline

The biggest mistake of my career? That has to be the solid-state pilot line project we kicked off in early 2024. The conventional wisdom in the industry is that solid-state is the future, and Samsung SDI has been leading the R&D. We'd announced our pilot line target for 2027. Everything looked promising.

But I made a classic error: I assumed the manufacturing process would be a simple upgrade from our existing lithium-ion lines. We started specifying equipment based on that assumption. It wasn't until September 2024—after we'd submitted orders for specialized mixing equipment worth roughly $320,000—that our R&D team flagged a critical issue. The solid-state electrolyte required a completely different tack pressure profile. The equipment we'd ordered wouldn't work.

I still kick myself for that one. If I'd run the spec changes past the process engineers before placing the order, we'd have saved a three-month delay and a lot of embarrassment. The lesson I added to our checklist: Never assume a new chemistry behaves like the old one. Verify every single processing parameter.

We caught the error when our lead chemist asked, 'How are you planning to handle the pressure sensitivity?' A simple question. The answer was: we weren't. That $320,000 worth of equipment had to be put on hold (which, honestly, felt like a massive failure at the time). We eventually sourced the right gear, but the timeline slipped. We're back on track for a 2027 pilot line, but only because of the team's scrambling—not my planning.

The 48V Battery That Wasn't (And What I Learned About Customer Education)

Around the same time, we had a project involving our 48V battery systems for a potential client. A European customer—I won't name them—was exploring our 48V solutions for a backup power application. Everything looked straightforward. The customer's engineer said they needed a 'standard 48V system.' So we quoted a standard 48V system.

The problem? Their idea of 'standard' involved a voltage tolerance of ±2%. Our standard 48V battery (which uses a typical lithium-ion stack) has a nominal voltage of 48V, but its open-circuit voltage can swing between ~40V and ~54V depending on the state of charge. It's not a regulated power supply. It's a battery.

The customer rejected our first quote. They rejected the second. After the third rejection in Q1 2024, I finally created a pre-check list for all custom quotes. At the top of the list: Verify the customer's actual voltage requirements, not just their 'standard' assumption.

I'd rather spend 10 minutes explaining that a 48V battery isn't always exactly 48V than deal with mismatched expectations later. An informed customer asks better questions. After that third rejection, we clarified the spec, explained the voltage curve, and offered a DC-DC converter as a solution. The customer bought it. The mistake wasn't the customer's—it was ours for not educating them up front.

The Disconnect Saga: What Cable Do You Disconnect First on a Car Battery?

This one isn't about a huge budget loss, but it's the kind of fundamental mistake that makes you feel like an absolute amateur. A junior engineer on our team—let's call him Tom—was tasked with testing a new 12V monitoring system for a client. He's a smart guy, great with data analysis. But he'd never actually worked on a car battery before.

He called me one afternoon, frustrated. 'I think I shorted the test rig,' he said. 'The ground cable sparked when I was disconnecting it.' I asked him which terminal he disconnected first. There was a pause. 'The positive,' he said. 'Everything I'd read said to disconnect the negative first. In practice, I was in a hurry and just grabbed the red one.'

I wanted to be angry, but I couldn't. Because I'd made the same mistake seven years ago, in 2017. I was testing a prototype monitoring system and had the same spark, the same moment of panic, the same realization that the ground cable (negative) always comes off first to break the circuit safely. It's not about the cable color—it's about safety. Disconnecting the positive first leaves the circuit complete through the chassis ground. A wrench touching the positive terminal and any metal part of the car creates a dead short.

I added a new step to our safety checklist: For any automotive battery work, verify the disconnect sequence: Ground (negative) first, then positive. Reconnect in reverse: positive first, then ground. Tom hasn't repeated the mistake. Neither have I, since 2017.

Medical Systems and the Power Quality Lesson

Our battery systems are used in medical equipment, including nerve monitoring systems and MRI patient monitoring systems. These applications are unforgiving. A power glitch during an MRI can corrupt data or, worse, affect patient monitoring.

We had a project in late 2022 involving UPS batteries for an MRI patient monitoring system. The customer had specified a standard UPS with a battery backup. We quoted our high-capacity ESS solution. On paper, it was fine.

The mistake? We didn't ask about the transition time—the milliseconds between mains power failing and the battery taking over. The customer's existing UPS had a transition time of 10ms, which was fine for their previous equipment. Our system? It had a transition time of 15ms. For a standard PC, that's negligible. For sensitive medical monitoring equipment that samples data every few milliseconds, that 5ms gap caused a glitch. The system reported 'power quality issue' on the first test.

The most frustrating part of this: we had the spec. It was buried in the datasheet, but we hadn't flagged it as a critical parameter. You'd think we'd know to check transition times for medical equipment, but we didn't have a specific checklist item for it. After the Nth time something like this happened—well, the third time in a year—I was ready to consider a career change. What finally helped was creating a dedicated checklist for medical and critical power applications, with explicit notes on transition time, voltage regulation, and ripple tolerance.

Building the Checklist

The fourth or fifth failure—I've lost count—prompted me to create the master checklist I maintain today. We've caught 47 potential errors using this checklist in the past 18 months. 47 errors that could have cost us time, money, and credibility. The checklist isn't a one-size-fits-all document. It's modular:

• Chemistry Verification: Does this spec match the battery chemistry? (Learned from the solid-state pilot line disaster.)
• Voltage Clarification: What is the customer's actual voltage tolerance? Do they understand nominal vs. open-circuit? (Learned from the 48V rejection.)
• Safety Sequence: For any field or test installation, confirm the connect/disconnect order. (Learned from Tom's spark.)
• Critical Power Parameters: For medical or sensitive loads, verify transition time, ripple, and regulation. (Learned from the MRI glitch.)

I don't pretend this checklist is infallible. I'm sure I'll add more items in the future. But it's better than relying on memory, which is what I used to do.

The Bottom Line

The most expensive mistake I ever made cost about $3,200 in redo charges—the solid-state equipment order. The most embarrassing mistake cost a few hundred dollars but damaged our credibility with a client. The lessons, though, are worth more than any of those figures.

An informed customer is the best customer. If they understand the voltage curve of a 48V battery, they won't specify impossible tolerances. If they know the safety sequence for a car battery, they won't short a test rig. If they understand transition times, they won't assume all UPS systems are equal.

And for the engineering side: assumptions are the enemy. Every spec, every parameter, every 'standard' should be questioned. That's why I now maintain our team's checklist. Not because I'm naturally organized, but because I've made enough mistakes to know I need it.

(Prices and examples as of 2025—verify current specs, obviously.)