The call that started it
At 6:12 on a Thursday in March 2024, my phone rang. The caller ID said Dan—a veteran installer I’ve coordinated field support with for six years. He wasn’t the type to panic. But when I answered, he skipped the greeting. “We’ve got a dead system, and it needs to be online by Saturday morning.”
Dan was standing in a remote training center for volunteer fire departments in the Cascade foothills. The building ran on an SMA off-grid system, and the grant that funded the center was tied to a live demonstration that weekend. If the power stayed down, the organization stood to lose a $70,000 grant. So the pressure wasn’t just technical—it was financial. In my role as technical support lead, I’ve seen this movie before. Based on our internal data from 130+ emergency site calls, roughly one in four inverter “failures” turns out to be a connection issue. Panic doesn’t help. Method does.
The site wasn’t exotic: four 12V AGM batteries, a battery disconnect, and an SMA Sunny Island 6048-US. Nothing about the SMA-inverter setup was unusual. That made the failure more suspicious.
The multimeter test that changed the direction of the day
Dan’s first clue was a DC undervoltage alarm on the Sunny Island display. When he tried to reconnect the system, the inverter would start, then drop back to the same fault. He had a multimeter in one hand and a coffee in the other. “I don’t need a lecture,” he said. “I need to know how to test a 12v battery with a multimeter before I buy a whole new battery bank.”
That’s the right question. Here’s the short version I gave him:
- Turn the multimeter dial to 20V DC.
- Put the red lead on the positive terminal, black on negative.
- At rest, a 12V AGM should read about 12.6–12.7V.
- If it reads below 12.4V, the battery is roughly 50% charged.
- Then check under load—a multimeter alone can miss a bad cell that collapses when current flows.
Dan tested all four batteries. At idle, they all looked fine: 12.5 to 12.7V. But when he hit one with a load tester, it sagged to 9.1V. We found the dead cell. He phoned a distributor, paid $180 in expedited freight, and by 2:30 p.m. the replacement battery was on-site.
The surprise wasn’t the bad cell. It was that the new battery didn’t fix the fault.
When the obvious fix isn’t the fix
Dan installed the new AGM, cleaned the terminal posts, and powered the system back up. The Sunny Island’s display lit up—then dropped straight back into the same DC undervoltage fault. That’s when I pulled up the SMA inverter data sheet again and went through the troubleshooting tree line by line.
The SMA inverter data sheet was clear about alarm thresholds. It was not clear about what could cause that alarm when the batteries were healthy. That info was in the installation manual, buried in a section about torque checks on field connections. “Check all torque values after initial installation” is one of those lines you read a hundred times and ignore once.
I can hear Dan’s response over the phone: we already checked the connection torque. But I asked him to check again, specifically at the disconnect switch terminals—not just the battery posts. In the background, I could hear the inverter generator Harbor Freight had sold him the year before, humming along with a work light and a GoPro battery charger. He had set up a video camera so he could document the troubleshooting; that footage ended up being useful later.
When Dan touched the lower-negative terminal at the disconnect switch, the lug rotated under his screwdriver. The bolt felt snug until he applied real force. He pulled the lug off and found green corrosion underneath. The connection was making contact at rest, but as soon as current flowed, it turned into a high-resistance bottleneck. The battery voltage looked good at the battery. At the inverter terminals, it was dropping to 10.8V. The inverter was doing exactly what it should have: shutting down before it damaged itself or the battery.
Twenty minutes later, the connections were cleaned, reterminated, and torqued to the spec printed in the manual. The inverter fired up and stayed up. The demonstration happened on time, and the grant was saved. We paid $180 in freight for a battery we didn’t strictly need, but that test helped us find the actual problem. Sometimes the wrong path is still part of the right path.
The lesson: prevention beats panic
To be fair, Dan had done the right thing by testing the battery first. Most installers focus on the inverter display and skip the 12V system entirely. The question everyone asks is, “Is the inverter broken?” The question they should ask is, “What changed in the last month, or the last season?” In this case, the change was a lug that had gradually loosened through years of heating and cooling cycles.
This is why I keep a checklist taped to my laptop. Before swapping any SMA product, check every battery connection under load, verify torque on both ends of each cable, and confirm the voltage at the inverter bus—not just at the battery terminal. Five minutes of verification beats five days of correction. That checklist, created after that call, has saved us an estimated $14,000 in rework trips. Note to self: I still need to turn it into a printed card for every installer in our network.
The SMA Solar Technology 2023 inverter output sold GW figure, by the way, was 19.6 GW (Source: SMA Solar Technology AG, 2024). With that many inverters in the field, loose terminals and poor connections are going to happen. Plan for them.
Dan’s system is still running today. And his first action on the next site? He tests the 12V battery with a multimeter before he ever calls me.
