Power Station UPS Mode & Switchover Time

Here’s the double inversion that catches most buyers off guard. The switchover time printed on the box — that “<20 ms” or “<30 ms” figure — is a marketing ceiling, and independent bench testing has found real units landing well above it. And the always-ready convenience of UPS mode, the thing you’re actually paying for, works by holding your battery pinned at 100% — which is quietly the worst state for long-term lithium health. Most people assume the spec is a measurement and the battery behavior is fine. Both assumptions are backwards.

This guide untangles both problems: what switchover time actually means for your gear, and what running a power station in UPS mode does to the battery over months and years.

The Switchover Number on the Box Isn’t What You Think

When a manufacturer lists “<30 ms transfer time,” they’re stating a ceiling, not a typical result. Any unit they ship can land anywhere below that ceiling — including right up against it. The number tells you the worst case they’ll claim, not the measured average.

The only direct comparison in the public record comes from a bench-testing outlet that put an oscilloscope on several units and measured from AC waveform fall-off to current ramp-up. The results scattered widely:

That last line is the one that matters most. EcoFlow listed the Delta 2 as switching in under 30 ms. Bench testing found 38.5 ms — the unit didn’t even meet its own marketing ceiling. The tester noted it “just barely passed,” meaning the test load survived, but there was no margin.

The lesson isn’t that all power stations fail — the Jackery and Bluetti units performed well, and the dedicated Eaton UPS was the reference standard. The lesson is that the printed spec tells you almost nothing about where your unit will actually land. Trust bench measurements over spec sheets, and treat any number without a scope trace behind it as unverified.

Whether the Switchover Matters Depends on Your Load, Not the Station

Here’s what the spec sheet doesn’t tell you: switchover time only matters in relation to how long your equipment can coast through a power interruption on its own. That’s called holdup time, and for many desktop PC power supplies the design minimum is around 16 ms.

That 16 ms figure is the real pass/fail line — not the marketing number. A station switching in 8–15 ms has comfortable margin for most gear. A station switching in 27–38 ms is gambling against the load’s internal capacitance. The EcoFlow Delta 2’s 38.5 ms is more than double the typical holdup floor, and the test load happened to survive — but that’s not the same as having margin to spare.

A few things tighten the holdup window further:

• Cheaper or older power supplies may hold up for less than the 16 ms design minimum
• The heavier the load relative to the PSU’s rating, the shorter the real holdup
• Sensitive equipment — some servers, certain medical devices — may need a true-online (zero-transfer) UPS where the inverter is always in the circuit, rather than switching at all

Neither number — the station’s switchover time or your PSU’s holdup — appears on the box. “It passed the test” means the test device survived; it doesn’t mean your specific setup will. If you’re protecting gear you can’t afford to reboot unexpectedly, verify the holdup spec for your actual power supply before relying on any station that tests above 15 ms.

UPS Mode’s Hidden Cost: Sitting at 100% Is Hard on Lithium

This is the problem nobody talks about in the marketing copy, because the default behavior is a selling point. UPS mode works by keeping the battery charged to 100% at all times — that’s what makes it “always ready.” It’s also, for a lithium cell, one of the more stressful states to live in indefinitely.

High state of charge accelerates calendar aging in lithium chemistry. The pack isn’t cycling — it’s just sitting there, full, held there continuously. NMC chemistry degrades faster under these conditions than LFP, but LFP is not immune. Heat compounds the problem: a battery that’s both at 100% and running warm ages considerably faster than one that’s partially discharged and cool.

The practical guidance that emerges from this — treat it as sound general lithium practice rather than a bench-confirmed finding — is:

• For permanent standby use: if your unit allows a charge ceiling, set it to around 70–80%. You give up a slice of available backup capacity, but you meaningfully reduce the aging stress.
• For short-term use (a storm, a planned outage): running to 100% for a day or two is fine. The damage accumulates from months and years at full charge, not from an occasional peak.
• For LFP chemistry specifically: you can’t cap it at 70% and forget it. LFP’s voltage curve is very flat, which makes it hard for the battery management system to track the true state of charge. A periodic full charge — roughly weekly — helps the BMS recalibrate. Cap it for daily standby, but let it top off regularly.

Vendors almost never surface the charge-limit setting in their UPS mode documentation, because “always fully charged” is the feature they’re selling. The setting exists on many units — you just have to go looking for it.

The 10-Year Lifespan Claim and What It Actually Means

You’ll see LFP stations advertised with lifespans like “10 years to 80% capacity.” One vendor pairs this with a 5-year warranty. The gap between those two numbers is telling.

No reviewer can independently verify a 10-year projection — no test window is long enough. These figures are lab extrapolations from cycle testing, and they don’t always state the conditions: what temperature, what depth of discharge, what charge rate. When conditions aren’t stated, the number is optimistic by construction.

More practically: that 10-year projection assumes something close to ideal storage and cycling conditions. UPS mode’s default behavior — battery pinned at 100%, continuously — is the opposite of ideal. The lifespan claim and the product’s default operating mode work against each other, and that caveat is absent from the spec sheet.

Treat the warranty period as the firmer commitment. Treat the 10-year figure as a ceiling under favorable conditions you’re unlikely to replicate in a permanent standby application.

Runtime for Real Loads (Not the Cherry-Picked Router Example)

Vendor runtime figures almost always lead with the friendliest possible load. A common example from one manufacturer: a roughly 286 Wh unit running a 3 W router for around 35 hours. That’s technically accurate and nearly useless for sizing decisions.

The same unit running a 600 W desktop runs for roughly 21 minutes. Those two numbers come from the same product page — the headline and the relevant figure are very different numbers, and manufacturers consistently lead with the first one.

Runtime math is simple: capacity in watt-hours divided by load in watts gives you hours. But real runtime is shorter than that calculation suggests, because the inverter carries overhead and surge draws reduce the effective capacity available. Manufacturer figures don’t account for those losses, so treat any vendor runtime spec as an optimistic ceiling.

Where power stations genuinely outperform traditional desktop UPS units is in raw capacity. A small lead-acid UPS gives a 600 W desktop perhaps a few minutes of runtime. A lithium station with even 600 Wh gives you substantially more headroom. The direction of that comparison is correct; the specific minute figures on vendor pages are marketing, not measurement.

Sizing for What You’re Actually Protecting

If your backup goal is keeping a refrigerator safe, the FDA’s guidance gives you a concrete anchor: food stays safe for about 4 hours if the refrigerator door stays closed. That’s your sizing target — bridge 4 hours of fridge load and you’ve covered the food-safety window.

A few things shorten that window in practice:

• Every time you open the door during an outage, the clock resets shorter — warm air replaces cold air
• A near-empty fridge loses temperature faster than a full one
• A hot kitchen shortens the window; a cool one extends it

Treat 4 hours as a best case with the door held closed, not as a guaranteed buffer. Size the station to comfortably exceed that window with your actual fridge’s running wattage, not the vendor’s example load.

The Two Things to Get Right

Strip this down to the decisions that matter: verify the switchover time matters for your specific load (check your PSU’s holdup spec, not just the station’s marketing ceiling), and if you’re running UPS mode long-term, find the charge ceiling setting and use it. The fast-switching number is less reliable than it looks. The always-ready convenience costs battery life you can’t see until it’s gone.