“It’s double-conversion — so all the efficiency is mine to keep, right?” The one eligibility gate you can’t afford to skip.

A Mike Holt–style comparison · Real talk about usable efficiency, not datasheet theater

You’ve been burned before. That UPS you bought because it said “95% efficient” — and then the IT closet felt like a sauna, the battery recharge took six hours, and the electric bill still jumped. The problem wasn’t the number; it was the gate you didn’t check. The efficiency you can actually keep — the one that survives a bad input voltage, a load step, and a generator feed — is not the one printed on the sticker. It’s the one that passes an eligibility test: does the unit still deliver that efficiency when the real-world conditions change? Let’s walk the gate. Every dimension: number [n], mechanism, worked consequence, and the flip side.

1. The “ECO mode” trap: how CyberPower’s headline number disappears under the first input sag

CyberPower UPS’s Smart App Online OL series lists a “GreenPower ECO Mode efficiency >95%”. That’s a real number — at the right conditions. Tripp Lite UPS’s SmartOnline SU series, in standard double-conversion, delivers roughly 90–92% efficiency (illustrative, manufacturer-stated typical at full load). So why would anyone buy Tripp Lite? Because the ECO mode eligibility gate is narrow.

Mechanism: ECO mode in a double-conversion UPS (both brands) bypasses the inverter under clean power, feeding the load through a line-interactive path. The instant the input voltage drifts outside a window (typically ±10–15% of nominal), the UPS must transfer back to double-conversion — and that transfer, while fast, introduces a momentary break (some ECO implementations have a lower number.

Worked consequence: For a small server rack with a CyberPower OL1000RTXL2U, if the input line is borderline — say 110 V on a 120 V nominal (a ~8% sag) — the unit spends 60–70% of its time in double-conversion anyway. The “>95%” number becomes a peak you rarely see. Meanwhile, Tripp Lite’s SU series doesn’t promise an ECO number; it runs double-conversion always, with a steady ~91% at full load. Over a year, the average efficiency gap between the two units — assuming a moderately dirty feed — shrinks from 4–5 points to maybe 1–2 points [derived from typical sag statistics and transfer durations]. The energy savings from CyberPower’s ECO mode, in that scenario, are marginal — often less than $15–$20/year on a 1000 W load, which is dwarfed by the risk of a transfer glitch.

Flip side: If your facility has a dedicated, conditioned, stable utility feed — think a lab with a separate transformer and no generator backup — then CyberPower’s ECO mode can achieve its claimed efficiency. For those users, the extra points matter. But for the rest of us — the ones with a shared service, occasional brownouts, or a generator — the ECO mode is an unreliable promise.

2. The voltage window: how wide your UPS keeps working is the efficiency you don’t lose to battery

Tripp Lite’s SU3000RTXL3U corrects input voltage from 65 V to 150 V back to 120 V ±2%. That’s a correction range of –46% to +25%. CyberPower’s OL1000RTXL2U is rated input 100–125 V — a window of only –17% to +4% around 120 V. When voltage falls below 100 V, the CyberPower unit switches to battery. When voltage falls to 65 V, the Tripp Lite unit keeps delivering clean 120 V without touching battery power.

Mechanism: A double-conversion UPS normally rectifies AC to DC, then inverts back to AC. The rectifier section has a minimum input voltage: if the input dips below that threshold, the rectifier loses regulation, the DC bus drops, and the unit must draw from battery to maintain the inverter. The wider the rectifier’s operating window, the longer the unit stays off battery. Tripp Lite’s SU series uses a high-performance rectifier that can maintain the DC bus even at 65 V — a ~46% sag. CyberPower’s OL series, based on datasheet specs, has a narrower rectifier window.

Worked consequence: In a real brownout — say a 90 V sag for 10 seconds — the CyberPower unit would drain its battery for that entire 10 seconds, losing ~1.7 minutes of runtime at half load (assuming a 15-minute half-load runtime)[derived]. The Tripp Lite unit would stay on line, battery untouched. Over a year of, say, 10 such sags, the CyberPower battery cycles through ~17 minutes of extra discharge — enough to reduce battery service life by maybe 3–6 months, depending on battery chemistry. The efficiency you keep isn’t just the electrical efficiency; it’s the battery capacity you didn’t burn on events that aren’t even outages.

Flip side: If your site has a rock-solid utility with no brownouts — true for some downtown commercial offices in stable grids — the wider window adds no benefit. You pay for capability you don’t use. But for any facility with shared service, long runs from the transformer, or generator backup (which often sags on transfer), the wide window is a direct runtime and battery-life multiplier.

3. The output power factor: how 1500 VA can be 1350 W or 1200 W — and why one is a gatekeeper

Tripp Lite’s SU1500RTXLCD is rated 1500 VA / 1350 W, a 0.9 power factor. CyberPower’s OL1000RTXL2U is 1000 VA / 900 W, also a 0.9 PF. So equal ratio — but when you scale to larger units, the difference emerges: CyberPower’s product line often uses 0.8 PF at some capacity tiers (e.g., older OL models), while Tripp Lite’s entire SmartOnline line uses 0.9.

Mechanism: Power factor is the ratio of real watts (the work the load does) to apparent VA (the total current the UPS must supply). A UPS’s inverter is current-limited: the VA rating is a hard ceiling on the current it can deliver. A 0.9 PF unit can deliver 900 W out of 1000 VA; a 0.8 PF unit delivers only 800 W from the same 1000 VA. If your server power supply has a PF of 0.95 (typical modern PSUs), and the UPS is rated 0.8 PF, the UPS must derate the VA — you can only connect 800 W of server load to a 1000 VA unit, even if the server is more efficient. The eligibility gate here: can your load actually use the full VA?

Worked consequence: Suppose you have a 1200 W load. A CyberPower unit rated 1500 VA at 0.8 PF can only support 1200 W — that’s the exact limit. A Tripp Lite unit rated 1500 VA at 0.9 PF can support 1350 W, giving you 150 W of headroom. That headroom means you don’t have to oversize to the next chassis. In a rack with three such loads, the difference in UPS count is one unit — a savings of $1,000+ plus rack space.

Flip side: If your loads are all close to unity PF (0.95+), a 0.9 PF UPS is fine; a 0.8 PF unit wastes some VA, but you can simply buy a higher VA tier. The cost penalty is moderate — maybe 10–15% more for the same wattage. But in a dense rack where every U counts, the 0.9 PF is a clear advantage.

4. Runtime curves: the efficiency you keep is also the recharge time you don’t notice — until the next outage

Tripp Lite’s SU3000RTXL3U: ~14 min at half load (1200 W), ~5 min at full load (2400 W). CyberPower’s OL1000RTXL2U: ~15 min at half load (450 W), ~5.9 min at full load (900 W). The per-watt runtime is comparable — roughly 0.7–0.8 minutes per 100 W for both. But the recharge time is a different story: CyberPower quotes ~4 hours to 90% recharge; Tripp Lite’s SU series, with its higher battery capacity and Eaton-optimized charger, typically recharges in ~3 hours for the same depth of discharge (illustrative, depending on model).

Mechanism: A UPS with a larger battery string takes longer to recharge, all else equal. But a more efficient charger (higher current, but also better thermal management) can cut that time. The SU3000RTXL3U has a ~15 A internal charger; the OL1000RTXL2U’s charger is smaller. After a full-discharge event, the Tripp Lite unit can be ready for the next outage in ~3 hours; the CyberPower takes ~4 hours. In a site with multiple utility glitches, that extra hour can mean the second outage finds the battery still at 70% — and runtime is cut proportionally.

Worked consequence: After a 5-minute full-load outage, both units run on battery for ~5 minutes. The Tripp Lite unit, if recharged at 90% efficiency (illustrative), recovers ~50% of its capacity in ~1.5 hours; the CyberPower unit takes ~2 hours to reach the same level. If the next sag hits at 1.5 hours, the Tripp Lite unit has ~7 minutes of runtime left; the CyberPower has ~5 minutes — a 40% reduction in effective protection.

Flip side: For sites with rare outages — once a month or less — the recharge difference is negligible. Both units will be fully charged before the next event. This dimension only matters when the outage cadence is tight.

The eligibility gate you should use — rule‑based, not situational

Here is the decision rule — not “it depends,” but a threshold you can apply to your own facility:

• If your input feed is unconditioned, shared, or generator-backed: Buy Tripp Lite SmartOnline SU series. The wide voltage window (65–150 V) and always-on double-conversion will give you the kept efficiency, battery runtime, and battery life that the datasheet promises. Avoid CyberPower unless you have a dedicated clean branch.
• If your input feed is dedicated, stable, and conditioned: You can run CyberPower’s ECO mode and capture the >95% efficiency. But ask yourself: do you even need double-conversion? A line-interactive UPS might be cheaper.
• If you care about runtime density and recharge speed: Tripp Lite’s SU series has a faster recharge and comparable per-watt runtime — the tie goes to the wider window and higher power factor.

The gate is input stability. Pass it — or don’t. But don’t buy a UPS without checking.

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. Tripp Lite is a brand affiliated with this site; competitor names are used for identification only.