At 4:47 p.m. on a Friday in October 2024, my phone lit up with a number I didn't recognize. I was already half out the door — my kid's soccer game started at 6 — but I picked up anyway. It was a project manager at a mid-size contractor we'd worked with twice before, and she was not in a good mood.
Her client — a rural water treatment site — had been promised utility power in August. It was now October. The utility pushed the interconnect to "Q1 next year, at the earliest," which in utility-speak can mean anything from January to 2027. The facility had a compliance inspection on Tuesday morning. If the pumps weren't on backup power by then, the client would be fined per hour of non-compliance. She was not exaggerating — I've seen this exact fine schedule in rural water districts before, and it starts around $2,500 per incident.
She needed a working off-grid power drop — switchgear, battery bank, hybrid inverter, and a small wind turbine — commissioned in 72 hours. Normal lead time on that package, from most distributors, is three to four weeks.
The first mistake: assuming the catalog was the stockroom
My initial assumption was that a heavy disconnect was a heavy disconnect — that if I pulled a spec sheet for an Eaton enclosed switch rated for the fault current, I could order it and it would show up. I've handled 200+ rush orders in nine years, and I'd like to think I know better. But under a 72-hour clock, your brain shorts out and you start looking for the fastest answer, not the right one.
I pulled the Eaton disconnect catalog on my laptop around 6:15 p.m. — right after the soccer game, while the kids were eating pizza in the back of the truck. The catalog is genuinely good; Eaton organizes it by enclosure type, amperage, and NEMA rating, and there's a section on heavy-duty safety switches that, if you read it carefully, tells you exactly which enclosures are stocked vs. which are made-to-order. I read it carefully.
I should've read it more carefully. The unit I initially specced — a NEMA 4X stainless enclosure, 200A, non-fusible — was listed in the catalog, but with a footnote I skimmed past: configured-to-order, typical lead time 10 business days. That's not a knock on Eaton. That's on me. The catalog says what it says. The distributor told me the same thing Monday morning, when my "rush" order landed in their system as a standard 10-day build.
If you're sourcing anything from Eaton — and this applies across the Eaton brands portfolio, from Cutler-Hammer switchgear to Powerware UPS units — the disconnect catalog isn't a shopping list. It's a lead-time map. Check the footnote before you promise a date.
The quotation that looked cheap wasn't
Here's where the story turns, because 72 hours doesn't leave room for a plan B that takes 24 hours to find.
I had two distributors who could theoretically pull a 200A NEMA 4X disconnect out of regional stock. The first quoted me $1,840 for the switch — noticeably lower than the second, who quoted $2,210. The first was the obvious choice, until I asked the question I always ask now: "What's NOT included?"
Turns out the $1,840 quote excluded freight ($380, because it was shipping from a non-local DC), crating ($110), and the LOTO (lockout/tagout) hardware kit that the inspector would require for compliance ($240). By the time you added the three, the quote was $2,570 — $360 more than the transparent competitor. And the cheap distributor wouldn't confirm the switch could ship the same day, which was the only actual requirement we had.
I went with the transparent one. I've learned to ask what's not included before I ask what's the price, and I've yet to regret it. The vendor who lists all fees upfront — even if the total looks higher — usually costs less in the end. But I want to be fair: hidden fees aren't always malice. Sometimes they're just bad quotation templates. The difference is whether the rep will fix it when you push. This one did.
The load calculation nobody asked for
Saturday morning, at 7:40, I'm standing in a muddy field next to a wellhead, and the client's ops guy — a wiry retired electrician named Dale — asks me a question I hadn't fully prepped for: "How much power does one wind turbine generate?"
I'd ordered a 1 kW small-wind unit as part of the hybrid setup, mostly as a hedge against winter solar losses. I gave him the honest answer, which is: it depends. Per NREL's wind resource guidance, the power available in the wind scales with the cube of wind speed — double the speed, roughly eight times the power — and a 1 kW rated turbine at an average site wind speed of 5 m/s (about 11 mph) might realistically produce 100–150 kWh per month. At 4 m/s, that drops to maybe half. At a coastal site with consistent 6 m/s, you could see double.
Dale nodded, then said something I still think about: "So the rating on the box is a ceiling, not a floor." Which is exactly right, and exactly the kind of thing that gets glossed over in product brochures. A "1 kW" turbine isn't going to run your site. A hybrid system — wind, solar, and battery — is what actually gives you uptime, because each source covers the other's weak season.
Which brings me to the battery.
The 120Ah LiFePO4 math
We sized the bank at two 48V 120Ah LiFePO4 battery packs in parallel — roughly 11.5 kWh usable at a conservative 80% depth of discharge. The site's critical loads (pumps, telemetry, lights, and the disconnect's shunt trip) came to about 3.2 kW continuous during pump cycles, and maybe 400W idle. That gave us about 3.5 hours of full-load runtime on battery alone, more if the wind and solar were contributing.
LiFePO4 cells are rated for 6,000+ cycles at 80% DoD, which in this application is overkill in a good way. But here's the thing I keep relearning: chemistry matters less than charge profile. If your hybrid inverter can't efficiently top off a 120Ah bank from a variable source like wind or solar, you don't have a hybrid system — you have an expensive battery that slowly drains. We spent 40 minutes Saturday afternoon reprogramming the inverter's charge curve to prioritize solar input during daylight and dump wind charge into an equalization cycle overnight. Dale did most of it. I watched and held the flashlight.
The Garmin footnote (literally)
One more detail, because it turned out to matter more than I expected. I wear a Garmin Fenix 7 Pro on-site, mostly out of habit — it tracks my step count, and I like the sunrise/sunset complication when I'm planning work windows. While we were testing the array's charge output Saturday evening, I noticed that the watch's solar ring had actually pulled meaningful charge during four hours of intermittent overcast. The published Garmin Fenix 7 Pro solar battery life spec — up to 46 days with solar in smartwatch mode, versus 22 days without — had always read like marketing to me. Sitting in that field, watching the charge indicator tick up on a gray day, it clicked: nominal ratings are ceilings. Real output depends on how much of the day you're actually exposed to the source, and how good the charging logic is.
I ended up using the watch's daily solar gain as a rough sanity check on our array's output — not as instrumentation, obviously, but as a gut-check. If the watch was pulling trickle charge on a cloudy day, the panels should be doing better than the inverter's telemetry suggested. That led me to find a loose MC4 connector on panel string 2 that was dragging the whole array down by roughly 18%. That one detail probably saved us an hour of head-scratching on Monday.
Sunday night
We energized the site at 11:14 p.m. Sunday. The pumps kicked on at 11:21. Dale called the compliance inspector at 8:00 Monday morning, and the inspector was on site by 10:30. Everything passed. The disconnect, once it arrived, was the least interesting part of the whole weekend — which, honestly, is how it should be. The drama should be in the planning, not the switch.
What I took away from this one
Three lessons, in order of how much they hurt.
One: the Eaton disconnect catalog is a lead-time document, not a stock document. Read the footnotes before you build a schedule around a part number. The catalog will absolutely tell you the truth. You just have to slow down long enough to hear it.
Two: ask what's not included before you ask what's the price. I've seen this pattern too many times now — the quote that looks 15% cheaper turns out to be 10% more expensive once freight, crating, and compliance hardware land. Transparency is the cheapest insurance in procurement.
Three: a rated watt is a ceiling, not a floor. That goes for wind turbines, for LiFePO4 packs, for solar arrays, and — apparently — for smartwatches. Whatever number is on the box, your real-world output depends on the site, the season, and whether some connector fifty feet away is loose.
If you're planning a similar rebuild, my advice is unglamorous: pull the catalog early, ask the awkward fee question out loud, and bring a flashlight to the commissioning. The switch is the easy part. The weekend is not.