Eaton vs Tripp Lite UPS: when the load doubles

Wednesday 17th of June 2026 · Jane Smith · Blog
failure‑mode comparison double‑conversion class myth vs reality

You sized a 3 kVA UPS for 1.5 kW of IT gear. Then a second compute node is added — load goes to 2.7 kW. The UPS doesn’t trip immediately, but something will fail first. Which brand’s failure mode costs you less? This is not a runtime‑curve shootout. It’s a stress‑path forecast.

Myth: "3 kVA = 3 kW, always"

The claim: any 3 kVA double‑conversion UPS can deliver 3 kW because “power factor doesn’t matter in data centers.” The reality: the usable watts are set by the output power factor (PF). The Eaton 9PX at 3 kVA is rated 2700 W (PF 0.9). The Tripp Lite SmartOnline SU3000RTXL3U is rated 3000 VA / 2400 W — an implied PF of 0.8. That’s a 300 W gap on the same VA shell.

Mechanism. The inverter + output stage are designed for a maximum current at a given crest factor. A 0.8‑PF load draws more RMS current per watt than a 0.9‑PF load. The Tripp Lite unit’s 2400 W ceiling is a thermal limit: above that the inverter’s IGBT junction temperature rises toward the desaturation threshold. The Eaton 9PX uses a higher‑grade IGBT module and a wider output filter that can sustain 0.9 PF continuously.

Worked consequence. Your original 1.5 kW load is fine on either. After the second node (total 2.7 kW):

  • On the Eaton 9PX (2700 W max) you are right at the nameplate — the UPS will run but with zero headroom; every AC ripple pushes the inverter into current limit. The unit will eventually transfer to bypass due to “overload” if the load crest factor exceeds 3:1.
  • On the Tripp Lite SU3000RTXL3U (2400 W max) you are already 12.5 % over the rated watts. The UPS will not immediately drop — the overload alarm sounds, and after ~2 minutes at 105 % the inverter shuts off and the static switch transfers to bypass.
    Failure mode: the load runs on raw mains with no battery backup until a human intervenes.

When this flips. If your load power factor is closer to unity (e.g. modern rectifier‑based servers with active PFC), the Tripp Lite UPS’s 2400 W limit is still a hard ceiling. But if you never exceed 2000 W, the Tripp Lite’s extra VA gives no advantage — and the Eaton UPS’s 0.9 PF is irrelevant. The myth bites only when you expect 3 kVA to deliver 3 kW.

Myth: "Overload = immediate shutdown"

The claim: any UPS that exceeds its watt rating will just click to bypass and keep the load alive. The reality: the type of overload and the firmware logic determine whether the load stays protected.

The Eaton 9PX handles overloads in three tiers: 105 % for 2 min, 110 % for 30 s, >150 % for 300 ms. After the timer expires it transfers to bypass and sends a Network Shutdown command (via PowerChute) to gracefully shut down the OS. The Tripp Lite SU3000RTXL3U uses a simpler threshold: overload at >105 % of rated VA triggers a gradual transfer to bypass after ~2 min, but the Eaton Brightlayer software (Tripp Lite’s management suite) does not distinguish between “overload” and “critical battery” events in the same way.

Mechanism. The Eaton 9PX uses a digital signal processor (DSP) with a true‑RMS current foldback algorithm: it measures both the RMS current and the crest factor, and extends the overload timer by 40 % if the load is below 0.9 PF. The Tripp Lite unit uses a fixed‑time thermal replica relay in the output stage, which cannot adapt to waveform shape.
Worked. Your 2.7 kW load on the Tripp Lite (2400 W max) is 112.5 % of the watt rating. The UPS sees >105 % on the current transformer and starts the 2‑minute timer. If the load has a slightly high crest factor (2.8 vs typical 2.5), the Eaton would hold another ~40 s before bypass; the Tripp Lite will drop at exactly 2 min regardless. In a real brownout the load may flicker during transfer.

Failure mode. If the load is a pair of dual‑corded servers, the Tripp Lite’s bypass transition may go unnoticed — one power supply briefly sees mains, the other stays on battery. But if the UPS is feeding a single‑corded switch with no graceful‑shutdown agent, the 2‑minute alarm is a scramble. The Eaton at least issues an orderly shutdown sequence before the battery drains.

When this flips. If you run a warm‑standby topology (two UPS units, each sized for the full load), the overload scenario is less likely: load is always below 50 %. Here the finer overload logic gives no benefit — both units will sit in idle.

Myth: "Battery runtime scales linearly"

The claim: double the load, half the runtime. The reality: internal battery packs obey Peukert’s law — at higher discharge rates the effective capacity drops faster than linear. The Tripp Lite SU3000RTXL3U provides ~14 min at half load (1200 W) and only ~5 min at full load (2400 W). That is a ratio of 2.8× less runtime for 2× load — a 40 % penalty beyond linear.

The Eaton 9PX has a similar battery chemistry (VRLA) but the 9PX 3 kVA model uses a slightly larger battery block: 4× 9 Ah vs Tripp Lite’s 4× 7.2 Ah. At 2700 W (full load) the Eaton is rated ~6 min; at half load (1350 W) it is ~16 min. The penalty ratio: 2.67× runtime drop for 2× load — similar but not identical.

Mechanism. The difference is not in the chemistry but in the DC‑bus voltage. The Tripp Lite SU series runs a 96 VDC bus; the Eaton 9PX runs 108 VDC. A higher bus voltage draws lower current for the same wattage, reducing I²R losses in the battery interconnect and reducing the Peukert penalty by roughly 6 %. That’s why the Eaton’s full‑load runtime is 6 min vs Tripp Lite’s 5 min — a 20 % improvement.

Worked consequence. With a doubled load (2.7 kW), the Tripp Lite gives about 5 min of autonomy — enough for a scripted shutdown but not for an extended blip. The Eaton gives 6 min. Neither is enough to ride through a 15‑minute utility outage. Failure mode: the battery voltage collapses before the generator starts; the UPS drops the load hard (no orderly shutdown). The Eaton’s extra minute may allow a second shutdown attempt if the first fails.

When this flips. If you add external battery packs, the Peukert penalty decreases because the load is shared across more cells. With two external packs on the Tripp Lite, total runtime at 2.7 kW can reach 18 min — erasing the Eaton’s advantage. The internal‑battery comparison matters only in a standalone configuration.

Myth: "Double‑conversion fixes everything"

The claim: a VFI UPS (both are double‑conversion) will always output clean 120 V regardless of input. The reality: each unit has a rectifier input window; outside that window the UPS transfers to battery, and if the battery is already depleted, the load drops.

The Tripp Lite SU3000RTXL3U corrects input voltage from 65 V to 150 V back to 120 V ±2 %. That is a very wide window. The Eaton 9PX (240 V model) has a typical input window of 160–276 V, but on the 120 V version the window is narrower: 85–145 V.

Worked consequence. In a severe brownout (e.g. 70 V due to a downed transformer), the Tripp Lite stays online and continues charging the battery while regulating output — the load never sees the disturbance. The Eaton transfers to battery at 85 V, draining the reserve even though the utility is still present. If the battery is already depleted from the earlier overload event, the Eaton’s output drops immediately. Failure mode: the wider input window of the Tripp Lite prevents a premature battery drain, preserving runtime for a later total outage.

When this flips. If your site has a voltage stabilizer or a generator with AVR, the window width is irrelevant. The narrower window of the Eaton becomes a liability only in weak‑grid scenarios.

⚙ Rule‑based wrap‑up (use the failure mode that decides for you)
  1. If your peak load is >85 % of the UPS watt rating → check the output power factor. If the VA‑to‑watt ratio is 0.8 (Tripp Lite), the unit will go to bypass under a doubled load. Only the Eaton 9PX (PF 0.9) can hold the full 2.7 kW without overload.
  2. If your site has brownouts below 85 V → the wider input window (65 V) of the Tripp Lite avoids unnecessary battery transfer. The Eaton’s 85 V threshold will drain reserve prematurely.
  3. If you need >15 min runtime at doubled load → plan for external battery packs on either unit. The internal‑based advantage (Eaton +1 min) disappears with one extra pack.
  4. If you run a graceful‑shutdown agent → Eaton’s tiered overload + PowerChute gives a more predictable shut‑down sequence. Tripp Lite’s overload transfer is simpler but lacks the “pre‑shutdown” phase.

Non‑obvious insight: The dimension that actually kills the UPS under a doubled load is not runtime — it’s the inverter’s current limit at low power factor. The Tripp Lite’s 0.8 PF ceiling is the binding constraint, not the battery capacity. In practice, a 2.7 kW load on the SU3000RTXL3U will hit the overload timer before the battery reaches 50 % depth of discharge. The failure mode is a transfer to bypass, not a battery‑down shutdown. That distinction changes how you size your power distribution: if you can accept a brief bypass transfer, the Tripp Lite still works; if the load is single‑corded and cannot tolerate even 20 ms of mains, the Eaton’s higher watt rating prevents the transfer entirely.

▲ Reversal case: For a load with near‑unity PF (e.g. fully active‑PFC rectifiers), the Tripp Lite’s 2400 W limit is still absolute, but the Eaton’s 2700 W limit is also a ceiling. If the load is 2.5 kW at PF 0.98, the Eaton holds; the Tripp Lite does not. But if the load is 2.0 kW, both units run without stress, and the Tripp Lite’s wider input window gives better brownout tolerance. The “winner” flips with load composition.

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

Leave a Reply