WHY UPS SYSTEMS ARE NON-NEGOTIABLE
Even with the best backup generator installed, there is always a gap between when the grid fails and when the generator reaches full synchronised output — typically 10 to 30 seconds. For a server rack, 10 seconds is long enough to corrupt databases, lose in-flight transactions, trigger emergency shutdowns, and damage hardware. For a hospital life-support system, it is simply unacceptable.
The UPS system solves this problem — but its role extends well beyond bridging the failover gap. A high-quality UPS continuously conditions the power supplied to connected loads, protecting against the full spectrum of power quality issues that cause equipment damage and data corruption even when the grid never formally fails.
THE THREE UPS TOPOLOGIES: WHICH IS RIGHT FOR YOU?
- Online Double Conversion — the gold standard for mission-critical applications. Mains power is continuously converted to DC and back to AC, meaning connected loads run entirely from the UPS output at all times. Grid events — surges, sags, frequency variations, complete failure — are completely isolated. Zero transfer time. Required for data centres, hospitals, and financial infrastructure.
- Line Interactive — the UPS monitors input voltage and uses an autotransformer to regulate it within acceptable bounds. Only transfers to battery on outright failure, with a transfer time of 2–4 milliseconds. More efficient than double conversion but provides less complete power conditioning. Suitable for mid-tier IT environments and commercial applications.
- Offline / Standby — passes mains power directly to the load under normal conditions and only activates on failure, with a transfer time of 4–8 milliseconds. Lowest cost and lowest protection. Appropriate only for non-critical consumer or light commercial loads where power quality is reliably good.
For mission-critical facilities, online double-conversion is the only appropriate topology. The efficiency premium of line-interactive designs does not justify the reduction in protection for facilities where downtime or equipment damage carries severe consequences.
POWER QUALITY: THE SILENT EQUIPMENT KILLER
Most facility managers focus on outages as the primary power threat. In reality, power quality issues — which occur on the grid continuously without triggering formal outages — cause an enormous amount of cumulative damage to sensitive equipment:
Voltage Sags
Momentary drops in voltage, typically caused by large motor starts or faults elsewhere on the grid. Can cause IT equipment to reset or crash without triggering UPS battery mode in offline designs.
Harmonics
Non-linear loads (switching power supplies, VFDs, LED drivers) inject harmonic distortion into the supply. Causes transformer overheating, capacitor failure, and interference with sensitive electronics.
Transients / Surges
High-energy, short-duration spikes caused by lightning, capacitor switching, or grid switching events. Can destroy unprotected electronics instantly.
Frequency Variation
Deviations from nominal frequency (50/60 Hz) affect motors, clocks, and synchronised systems. More common on isolated grids or during heavy generation events.
Online double-conversion UPS systems isolate connected loads from all of these issues simultaneously, because the load never actually connects to the mains — it connects to the UPS output, which is generated fresh from DC regardless of what is happening upstream.
SIZING AND SCALABILITY: PLANNING FOR GROWTH
A common and costly mistake is sizing UPS systems purely for current load. This creates two problems: first, UPS systems are most efficient at 40–80% of rated capacity, so an over-sized system wastes energy; second, an under-sized system cannot support load growth without replacement — which means downtime and additional capital expenditure.
Modular UPS architectures solve both problems. The base frame is installed with sufficient modules for current load at the target efficiency window, with additional module bays available for expansion. As load grows, modules are added online without system interruption. N+1 redundancy is built in by including one additional module above the load requirement.
BATTERY TECHNOLOGIES: VRLA, LITHIUM-ION, AND BEYOND
The energy storage element of a UPS is critical to its performance and total cost of ownership. Three main technologies are in widespread use:
- VRLA (Valve Regulated Lead Acid) — the traditional UPS battery. Lower upfront cost, proven reliability, wide service network. Requires temperature control (optimal at 20–25°C), has a 5–10 year lifespan, and is heavy relative to energy capacity.
- Lithium-Ion — higher upfront cost but significantly longer lifespan (10–15 years), smaller footprint, lighter weight, faster charging, and better performance at elevated temperatures. Increasingly the preferred choice for new data centre installations despite higher initial cost.
- Nickel-Cadmium (NiCd) — extremely robust across temperature ranges, very long lifespan, but high cost and regulatory complexity around cadmium disposal. Used in specialist industrial and outdoor applications.
"The total cost of ownership calculation over a 15-year period often favours lithium-ion, even at a significantly higher upfront price — fewer battery replacement cycles, lower cooling costs, and reduced rack space all contribute to the equation."
FREQUENTLY ASKED QUESTIONS
Runtime depends on the battery capacity relative to the connected load. Most data centre UPS systems are designed for 5–15 minutes of runtime — sufficient to bridge the generator start-up period with significant margin. Longer runtime (30 minutes to several hours) requires larger battery strings and is used in applications where generators are not immediately available.
Battery health degrades over time — capacity reduces and internal resistance increases. UPS monitoring systems track battery state of health and should alert when capacity drops below a threshold (typically 80% of rated capacity). Batteries should be replaced before they reach end of life to ensure the system maintains its specified runtime. The Power Vault Group includes battery health monitoring and replacement in all maintenance programmes.
UPS sizing requires a full load audit: the nameplate or measured kVA of every connected device, plus headroom for growth, inrush current during startup, and N+1 module redundancy. The Power Vault Group conducts comprehensive load assessments as part of every UPS specification project, ensuring the system is correctly sized for both today and the foreseeable future.
An online double-conversion UPS provides comprehensive protection against virtually all power quality issues — outages, sags, surges, harmonics, and frequency variation. Line-interactive and offline UPS systems provide partial protection. For complete isolation of sensitive loads from the grid, online double-conversion is the only topology that achieves this.
UPS systems require regular preventive maintenance including battery capacity testing, visual inspection, connection torque checking, filter cleaning, firmware updates, and full load testing. Annual maintenance by a qualified engineer is standard practice for mission-critical installations. The Power Vault Group provides manufacturer-authorised maintenance across all major UPS brands.
Yes — and this integration is critical. The UPS must be designed to tolerate the generator's output characteristics, which differ from clean mains power (typically wider voltage and frequency tolerances during startup and load acceptance). The transfer time between mains failure and UPS battery mode must be shorter than the generator start time, and the generator must have sufficient capacity for the UPS input current on load acceptance. The Power Vault Group engineers these integrations as complete systems, not individual components.