Why this comparison matters
A facilities manager called me on a Friday around 4 p.m. in March 2024. His generator had failed its load test. Monday at 8 a.m. is when his customer SLA audit was scheduled, and he'd planned on using that test recording as evidence.
What was actually wrong: the start battery.
He'd bought a dual-fuel unit — gasoline or propane — and the charger was running over the weekend. The battery was two years old. The green light on the charger was on, which pretty much only tells you the charger has power. We hit it with a multimeter. Resting voltage came back 12.29V — no, 12.31V, I'm mixing it up with a different call that week. Then load sag: 9.4V the moment the unit cranked. That battery was done in under two seconds.
We rushed a replacement start battery out of the nearest depot, booked a weekend service window, and paid the extra freight. The unit passed by Sunday night. The audit happened.
That's the job. I coordinate rush service for critical power systems — about 400 calls so far. Maybe 420, I'd have to pull the dispatch logs. A large share of them follow the same pattern: something breaks, a deadline is already moving, and the conversation becomes about certainty rather than price.
So when people ask "do I need a UPS or a generator," I think the question is framed wrong. The real comparison has three paths: the UPS path (Schneider/APC, sealed cells, modular), the generator path (a unit like a Champion dual-fuel running on a diesel battery charger and a start battery), and the repair path you're already in when something fails. Here's how I'd compare them.
Switching time: the gap nobody budgets for
UPS: Online double-conversion units — Galaxy, most of the Smart-UPS line — hold the load with zero transfer time because the inverter is always running. Line-interactive sits around 4–6 milliseconds. Even a basic standby unit is in the 10ms range.
Generator: When utility power drops, the transfer switch sends a start signal, the unit cranks, spins up, stabilizes. Ten to thirty seconds from outage to clean power. Thirty seconds is forever to a server.
On paper this looks like a clean UPS win. It isn't.
A UPS with typical battery strings buys you maybe 10–15 minutes at full load. A generator with fuel and a healthy start battery runs for days.
The part that surprises people: these aren't competing products. The UPS carries the load through the 20 seconds the generator needs to come online. The generator is what recharges the UPS. Buy only one and you've mostly insured yourself against the wrong failure mode.
The "a generator covers everything" thinking comes from an era when UPS wasn't standard on most loads and equipment could tolerate a power gap. That's changed. Any load with a switching power supply — which is basically everything now — needs the gap covered.
Maintenance burden: where the actual money hides
UPS side: Sealed cells, marketed as maintenance-free. Mostly fair — you're not watering anything. But VRLA batteries still need replacement every 3–5 years, and capacity testing matters. Modular designs help here. You swap one module instead of taking a whole row offline. Every unit we ship has to meet UL 1778, the safety standard for UPS equipment, so the hardware side is generally well understood.
Generator side: The engine is rarely the problem. The start battery is.
Here's what I ask a caller to do on the phone before I commit a truck:
- Resting voltage first. Disconnect or isolate the 12V battery, let it sit a few minutes, then probe it with the multimeter. Above 12.6V is healthy. 12.4V is about 75% charged. 12.2V is roughly half. 12.0V or below and you're looking at a battery on its way out. These are approximate, but approximate beats guessing.
- Then the crank test. Leave the meter connected and have someone hit the start. If voltage sags below about 9.6V while cranking, the battery can't reliably start the unit — replace it, don't "keep an eye on it."
- Check the charger output. A 12V battery on float should sit around 13.2–13.8V. Lower means undercharging. Higher means cooking the battery. Both shorten life, and both are invisible with a green indicator light.
(Should mention: NFPA 110 expects emergency power supply systems to be load-tested monthly, not just observed. A recording of a unit that starts is not the same as a recording of a unit that carries load.)
One more detail on fuel. Propane stores for a very long time in a tank. Gasoline doesn't — it starts degrading in a few months and needs stabilizer to stretch toward a year. If the generator is a backup for rare outages, don't fill it with gasoline and let it sit.
In my experience, most "generator started and then died" emergency calls trace back to the battery or the charger. Mechanical failures are the minority.
Rush service and the price of uncertainty
In 2023 a client skipped $400 of preventive maintenance on a standby unit. They ran two fake load tests — no load, just confirm it cranked — on a battery that should have been retired. Monday at 6 a.m., the UPS went to battery, and utility power didn't come back. At minute 11, the floor started dropping.
We had a tech on site in nine hours, a weekend emergency rate, a start battery trucked in from out of state, and a full load sweep before the system was stable. The invoice was just under $12,000. The skipped maintenance was $400.
The thing about rush fees is that you're not paying for speed. You're paying for certainty. That's the position I'll defend every time — a guaranteed window beats a cheaper maybe. If you have customers depending on a delivery deadline, an audit, or a live event, uncertain-but-cheap is the most expensive option on the table.
Same logic in both directions:
- If a Schneider UPS needs emergency service, skip the comparison shopping. Call the rush channel with the model and serial number ready.
- If a generator needs a start battery before a test window, don't wait for a convenient delivery date either. The gap between "next Tuesday" and "tomorrow" is usually the gap between passing and rescheduling.
How I'd actually decide
You have IT load or SLA customers: Both. UPS first, sized for enough runtime to cover generator start and a clean shutdown if the generator doesn't. The generator handles long outages. The Schneider/APC ecosystem is a reasonable choice here mostly because management software and firmware updates exist as a coherent set.
You have office gear, lighting, a few workstations: A decent tower UPS is enough. Skip the generator. Don't overbuild for a failure mode you don't have.
You already own a generator: Your first move is a multimeter and a charger check, not a service contract. The start battery is the single most likely point of failure and the cheapest part to verify.
You're mid-failure right now: Pay for certainty. A standard-turnaround repair that lands three days late isn't cheaper. It's just later.
Bottom line: the expensive part of power protection was never the hardware. It's the afternoon somebody didn't test a battery. The UPS, the charger, the multimeter — those are all ordinary components. What gets skipped is the five minutes it takes to use them.
