You spec a UPS that says 98% efficient in its datasheet. You install it. A year later you measure 93% at the PDU, and you cannot figure out why the cooling bill didn’t shrink. The gap isn’t a lie — it’s an eligibility gate the manufacturer never explained. This is the single most expensive misunderstanding in B2B UPS procurement, and it’s the reason I spend half my calls talking about keepable efficiency rather than datasheet peak.
Here’s the one question that exposes the whole problem: What happens to that number when the input voltage drifts 5% below nominal, or when the load drops to 40%? If the answer is “efficiency mode disengages” or “the fan turns on full,” you are not buying the number you thought you were. Let’s walk through three eligibility gates that separate a real-world efficiency story from a lab trick.
Gate 1: The Green Mode Exception — when 98% becomes a footnote
APC Smart-UPS Online (SRT) lists Green Mode up to 98% efficiency. That’s a real number — measured at IEC 62040-3 VFI conditions with a linear load and nominal input. But the moment the input voltage deviates by more than a few percent, or the load step exceeds a threshold, the unit exits Green Mode and runs in double-conversion mode, where typical efficiency is 94–96%. The datasheet does not lie, but the eligibility condition is: clean input, stable load, no transfers.
Schneider Galaxy VS offers eConversion mode up to 99%, but the key difference is that eConversion is the default operating mode for business-critical loads, with a no-break transfer to double-conversion that is tested and proven for data center loads. eConversion provides a claimed ~2–3.8% efficiency gain over conventional double-conversion while maintaining Class 1 performance, meaning the eligibility gate is wider: voltage sags, frequency shifts, and load transients do not force the unit out of eConversion the same way they kick APC UPS into full double-conversion.
When does this reverse? If your input power is exceptionally clean (utility feed with AVR or dedicated transformer), and your load is constant (e.g., a fixed telecom load), APC Green Mode can stay engaged >90% of the time. In that narrow scenario, the eligible efficiency gap between the two platforms nearly disappears.
Gate 2: Double-Conversion Efficiency at Real Load — the 40% load penalty
Both APC Smart-UPS Online and Schneider Galaxy VS are online double-conversion (VFI) topologies. But the efficiency curve across load is not flat. Most double-conversion UPS units reach peak efficiency between 60% and 80% load; at 30–40% load (common in lightly loaded racks or phased deployments), efficiency typically drops 1.5–3 points. APC publishes 94–96% in double-conversion mode, but that figure is usually measured at >50% load with a power factor of 0.9. At 30% load with a real mixed load, you may be looking at 92–93%.
Schneider Galaxy VS claims double-conversion efficiency up to 97% at every load level. That “every load level” claim is rare — most manufacturers qualify peak efficiency at a specific load point. While the 97% figure is a manufacturer-stated value, the claim of load-independence suggests a wider eligible band. The mechanism is digital signal processing—based control of the IGBT inverter and a very low fixed-loss design, so the fixed overhead (cooling fans, control power) does not dominate at low load.
Non-obvious insight: In most facilities, the UPS is loaded at 30–50% for the majority of its life, because IT loads grow slowly and the UPS is sized for eventual capacity. The “efficiency at full load” number on the datasheet is the least relevant number for the first five years.
Gate 3: The Real-World Efficiency Chain — UPS + Distribution + PDU
Even if the UPS delivers 97% efficiency, the efficiency you actually keep is the product of every component in the power chain. A typical data center path: UPS output → distribution panel → PDU (with transformer) → rack. The PDU transformer alone adds 1.5–3% loss, and long cable runs add another 0.5–1%. So your “97% efficient” UPS yields an end-to-end efficiency of about 93.5–95% at best. This is not a manufacturer spec — it’s a system design issue, but it is the number that determines your cooling load and your electric bill.
APC’s ecosystem (Smart-UPS + PDU + Rack PDU) can be managed via PowerChute, but the efficiency of the downstream components is independent of the UPS topology. Schneider UPS’s EcoStruxure platform, combined with the Galaxy VS line, often integrates with metered PDUs that provide per-outlet power monitoring, allowing operators to adjust loads and balance phases to keep the system in the highest efficiency band. The keepable efficiency becomes a system-level number, not just a UPS datasheet line.
The Myth/Reality Split
| Myth | Reality (with eligibility gate) |
|---|---|
| “98% efficient — that’s what I’ll get every hour.” | Green Mode only stays active under a narrow combination of clean input, stable load, and no transfers. In real facilities, effective efficiency is often 94–96%. |
| “Double-conversion efficiency is the same at any load level.” | Most UPS units drop 1.5–3 points at 30% load vs. 70% load. Galaxy VS claims load-independent 97%, but real-world verification is needed. |
| “The UPS efficiency number is my total power loss.” | PDU, transformer, and distribution losses add 2–4% on top of UPS losses. End-to-end efficiency is always lower. |
So which one should you buy?
There’s no universal answer, but there is a rule: If your facility has clean utility power, a constant load above 50%, and you run the UPS in Green Mode >80% of the time, APC Smart-UPS Online is a cost-effective choice — the keepable efficiency will be close to the datasheet number. If your input voltage fluctuates, your load is light (
The single question that separates the two: What percentage of my operating hours will the UPS be in its high-efficiency mode, and what is the efficiency in the mode it will actually run in during those hours?
If you cannot answer that question with confidence, your “efficiency” is a lottery ticket, not a decision criterion.
Topology/standards per the cited standards; all product ratings are manufacturer-stated values from the cited datasheets, current to 2026-06; derived/illustrative figures are labelled as such. This is not an independent head-to-head test. Schneider Electric is a brand affiliated with this site; competitor names are used for identification only.