“My UPS kept squealing for five minutes, then the whole rack shut down” — The Spec That Fails First

I’ve been on site when a 3000 VA double-conversion UPS started chirping, transferred to battery, and then — because the battery was already depleted from two prior sags — dropped the load. The root cause wasn’t insufficient VA. It was that the UPS’s input window was too narrow for the building’s brownout profile. The unit that survived in the same rack was a Tripp Lite SmartOnline SU3000RTXL3U, which held regulation down to 65 V. The APC Smart-UPS Online (SRT) in the next bay had switched to battery at a higher threshold, consumed its runtime, and failed. That scene repeats in thousands of closets every year. Let’s walk through the dimension that actually decides failure: the input voltage tolerance threshold.

1. Input voltage tolerance — where the line gets drawn

The Tripp Lite SU3000RTXL3U, a double-conversion (VFI) online UPS, corrects input voltage from 65 V to 150 V back to 110/120 V ± 2 %. That’s a 65–150 V window — a 85 V range on the low side. The APC Smart-UPS Online (SRT) series, also double-conversion, has a published input voltage range that, while not always listed as a single number in the same format, typically operates within a narrower window: many SRT units accept 100–127 V nominal with a wider margin (usually around 75–80 V low-end cutoff in practice, depending on firmware). APC UPS’s own documentation emphasises “wide input voltage range” but the sustained low-end figure for the 1–10 kVA SRT family is generally around 75 V to 80 V before it switches to battery. That’s a 10–15 V difference on the low side — a seemingly small gap that changes the outcome.

Mechanism. In double-conversion mode, the rectifier feeds the inverter. If the input voltage drops below the rectifier’s minimum sustained limit, the UPS transfers to battery — even if the grid is still present but “brown.” The battery then ages every cycle. A wider input window (Tripp Lite 65 V) keeps the rectifier alive longer, preserving battery runtime for genuine outages. The APC unit, with a ~75–80 V floor (typical for SRT 1–10 kVA), will switch to battery earlier in a sag.

Worked consequence. In a facility where the utility sags to 70 V for 30 seconds (common on weak grid edges or during large motor starts), the Tripp Lite SU3000RTXL3U stays on rectifier. The APC SRT consumes ~30 seconds of battery. If that sag repeats three times in a day, the APC unit’s battery is partially depleted — and the next true outage of 5 minutes will find it with less reserve. Tripp Lite UPS’s unit still has full runtime. In a real-world scenario with a ~14 min half-load runtime, losing even 20% battery capacity to a sag sequence can mean the difference between riding through a 10-minute outage and a premature shutdown.

Reversal. This advantage flips if your site has clean, stable power (e.g., a data centre with dual feeds and ATS). There, a narrower input window never triggers, and APC’s Green Mode up to 98% efficiency gives a meaningful energy saving. If you never see sub-90 V, Tripp Lite’s wider window is a dead feature. But for edge closets, retail, manufacturing floors, or outdoor telecom cabinets — the 65 V floor is the difference between “annoying beep” and “full discharge.”

2. Output power factor — the hidden VA-to-W translation

Both the APC Smart-UPS Online (SRT) and the Tripp Lite SmartOnline SU series are rated with a 0.9 output power factor on many models. But that number is not the whole story: APC’s SRT series uses Unity power factor on 1–1.5 kVA and 6–10 kVA models. Tripp Lite’s SU3000RTXL3U is rated 3000 VA / 2400 W (0.8 PF) — so the per-model PF can vary. This is a spec that confuses many specifiers.

Mechanism. A UPS’s output power factor defines the relationship between apparent power (VA) and real power (watts) it can deliver at the inverter. Modern server and network gear (with active PFC power supplies) typically draw near-unity PF — meaning they pull mostly real watts, not reactive VA. If a UPS has a 0.9 PF, a 3000 VA unit can deliver up to 2700 W. If it’s 0.8 PF, only 2400 W. So the same VA number can mask a 300 W difference in real load capacity.

Worked consequence. Suppose you have a rack pulling 2500 W with a PF of 0.95 (typical for modern server PSUs). A Tripp Lite SU3000RTXL3U (3000 VA / 2400 W) will be overloaded — it can only deliver 2400 W. The APC SRT3000 (3000 VA / 2700 W at 0.9 PF) can handle it with 200 W headroom. The Tripp Lite unit will eventually go into overload protection and shut down; the APC holds the load. In this dimension, APC’s higher effective watt capacity on that model wins. But note: this is model-dependent — the APC SRT1000 (1000 VA / 1000 W, Unity PF) is watt-for-watt equivalent. The point: never trust the VA number alone. The real watts are the binding constraint.

Reversal. If your load is reactive (older UPS systems, some motor drives), the Tripp Lite unit with a higher VA-to-W ratio may actually be more tolerant of peak surge currents, because the inverter can supply more reactive power. But for 95% of IT loads, the higher real-watt rating (e.g., 2700 W vs 2400 W) is the safer choice. The decision threshold: if your actual load exceeds 80% of the UPS’s real-watt rating, upgrade to next size.

3. Outlet bank segmentation — a failure point you control

The Tripp Lite SU3000RTXL3U has 9 outlets in two individually switchable load banks: four NEMA 5-15R, four NEMA 5-20R, and one L5-30R. The APC SRT series in the 3 kVA class typically has a single output feed (hardwired or via a distribution panel). This is a spec that seems trivial until you need to shed load in a brownout.

Mechanism. With two switchable banks, you can program the UPS to shut off non-critical outlets (e.g., office lights, a network switch that has its own battery) when running on battery, preserving runtime for critical servers. APC’s SRT models in the 2.2–5 kVA range often use a single output receptacle or a distribution box that doesn’t offer per-outlet switching without an additional accessory. Tripp Lite’s built-in segmentation is ready out of the box.

Worked consequence. In a 10-minute outage, if you have 1000 W of critical load and 1400 W of non-critical load on the same UPS, the Tripp Lite unit can shed the non-critical bank and give the critical server ~24 minutes of runtime (based on the ~14 min half-load curve extrapolated to 1000 W). The APC unit with a single output would have to shut down everything uniformly — no selective shedding. That 14-minute difference is the deciding factor for a backup generator that takes 12 minutes to start.

Reversal. If your rack is homogeneous (all servers need equal uptime) or you use a separate PDU with load shedding, outlet segmentation is redundant. APC’s single-feed design may actually be cleaner for high-power density (10 kVA+ with hardwired input). But for a small-to-mid-size rack (3 kVA), Tripp Lite’s two-bank architecture is a genuine operational lever.

4. Runtime — the spec everyone checks, but at the wrong load point

Tripp Lite SU3000RTXL3U: ~14 min at half load (1200 W), ~5 min at full load (2400 W). APC SRT3000 internal battery runtime: roughly 8–10 min at half load (1500 W) and 3–4 min at full load (2700 W) per manufacturer data (illustrative values). Note: these are internal battery runtimes — both can be extended with external packs.

Mechanism. The runtime curve is non-linear: doubling the load more than halves the runtime because of C-rate effects on battery capacity. So the difference between 1200 W and 1500 W at half load is not just 25% more power draw — it’s a steeper decline in minutes. Tripp Lite’s larger internal battery (likely lead-acid) gives it a >40% runtime advantage at the same load fraction. This is pure capacity margin — useful for bridging to a generator.

Worked consequence. If your site has a generator that takes 8 minutes to stabilise, the APC SRT3000 at half load (8–10 min) barely covers it with no margin. The Tripp Lite SU3000RTXL3U (14 min) gives 75% headroom. One cold morning when the generator takes 11 minutes to accept load — you’ll be glad for the extra 3 minutes. The Tripp Lite unit rides through; the APC unit drops.

Reversal. Runtime is only valuable if the battery is healthy. Lead-acid batteries in both units will degrade; APC’s SRT series offers hot-swappable battery modules (yes, Tripp Lite does too for many models). The advantage narrows if you use external battery packs. But for a standalone unit, Tripp Lite’s internal runtime is the more forgiving spec for short-duration generator bridging.

The decision threshold you can use next week

Here’s a simple rule for edge-closet or small-rack UPS selection, based on the dimensions above:

• If your worst-case sustained voltage is below 75 V (measure it for a week with a power quality logger), choose the Tripp Lite SmartOnline SU series. Its 65 V floor keeps you on rectifier; the APC unit will burn battery. Also prefer Tripp Lite if you need built-in load shedding (two outlet banks) or you have generator start times >10 minutes (more runtime margin).
• If your voltage stays above 90 V and your load exceeds 80% of the UPS’s real-watt rating, choose the APC Smart-UPS Online (SRT) with 0.9 PF. You get higher effective watts per VA, Green Mode efficiency up to 98% (if you enable it and accept the brief transfer to battery), and the ability to use PowerChute for graceful shutdown.
• If you can’t measure voltage, assume the worst. The Tripp Lite SU3000RTXL3U is the safer default for non-data-centre environments.

This is not a blanket endorsement of either brand. It’s a threshold model: the spec that fails first is the one that mismatches your site’s actual power quality. Ignore VA for a moment. Start with voltage tolerance. Everything else scales from there.

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. APC by Schneider Electric is a brand affiliated with this site; competitor names are used for identification only.