Skip to content
TechnoGuru — Think Technology, Think TechnoGuru

02 / 03

Case file

05 · Building Management

Online UPS.

Clean power, isolated from grid reality.

Double-conversion online UPS — continuous AC-to-DC-to-AC rectification that delivers regulated sine to your load with zero transfer time and complete isolation from grid sag, surge, harmonics and frequency drift.

Online UPS — representative visual (illustrative scene, not a project photograph)
Online versus line-interactive UPS: design fit
Online versus line-interactive UPS: design fit
AspectLine-interactive UPSDouble-conversion online UPS
Transfer on grid dropA brief switching gapNo transfer event — no switching path
Typical fitOffice loads that tolerate the gapMedical, broadcast, mission-critical loads
Power conditioningPartialFull isolation and regulated sine

Educational comparison — not about any specific installer.

How we approach it

Clean, continuous power — sized and supported

Backup power is only as good as its weakest cell and its last service. We plan UPS provision to the actual load and then keep it dependable through its service life.

Sized to the real load
Capacity is matched to the connected load and the runtime the site genuinely needs, with BOQ / scope alignment kept clear — not a number picked in the abstract.
Designed to be serviced
We plan for access and orderly battery replacement so maintenance is routine rather than disruptive.
Battery health over time
Batteries age whether or not they are used; an AMC scope can fold in periodic UPS and battery health checks so backup is there when the mains is not. Response targets are documented in the AMC scope.

/ The discipline, in detail

How we approach online ups.

Power architecture is where the cost of getting it wrong shows up months later — in compressor failures, drive trips, telco modem reboots, and the slow attrition of equipment that nobody links back to dirty mains. Double-conversion online UPS isolates the load from grid reality completely: the rectifier converts utility AC to clean DC, the inverter rebuilds it as a regulated sine wave, and the battery acts as the kinetic-energy reservoir between the two. There is no transfer event when the grid drops because there was never a switching path; the load has been on the inverter the whole time.

We size the topology to the actual load profile and the actual recovery time the building's diesel set can guarantee, then we right-size the battery accordingly. For data centres and clinical environments we engineer N+1 or 2N redundancy with synchronised parallel modules; for smaller deployments a single right-sized unit is the more honest answer than over-engineered architecture for its own sake.

VRLA versus lithium-ion is the second decision after topology. We almost always recommend lithium for any new deployment above 20 kVA — see our separate BESS service for the energy-asset case lithium unlocks. For sub-20 kVA loads where battery autonomy is short, VRLA remains a reasonable choice and we will say so in writing.

On record

Every online ups engagement is documented end-to-end — design, programming, commissioning, calibration — and handed over with the files our successors would need if we were never to return.

/ Power hierarchy

Continuity, by architecture

Utility, DG, ATS and double-conversion UPS as one power hierarchy — which loads ride through, which transfer, and how runtime is engineered.

Per-load UPS vs shared-bus UPS topologyPer-load UPS architecture (left): each critical support load sits on its own dedicated online double-conversion UPS so a fault on one load cannot cascade across the stack. Shared-bus architecture (right): one mainline UPS feeds every load, so a fault or maintenance event takes the entire stack offline. The per-load architecture costs more capex but enforces fault containment at the power layer rather than the application layer.UPS topology · per-load vs shared-bus · uptime engineering, not power-quality engineeringPer-load enforces fault containment at the power layer · shared-bus collapses on a single eventPower hierarchy · upstreamUtility · DG · ATS · stabilizerUtility feedGrid transformerDG standbyN+1 poolATSAuto transferMain stabilizerSensitive loadsEarthingPlant-boundSurge protectionType 1 + 2Per-load UPS architecture · uptime engineeringEach load class on its own UPS · fault isolation enforced at power layerLoad A UPSDedicated holdoverLoad A loadIsolated supportLoad B UPSDedicated holdoverLoad B loadIsolated supportLoad C UPSDedicated holdoverLoad C loadIsolated supportLoad D UPSDedicated holdoverLoad D loadIsolated supportLoad E UPSDedicated holdoverLoad E loadIsolated supportShared-bus architecture · power-quality engineeringOne mainline UPS · single event takes every load offlineSingle mainline UPSSingle point of failureBattery refresh / fault drops the whole stackLoad A (shared)Load B (shared)Load C (shared)Load D (shared)Load E (shared)Cascading failure across the whole stackSingle UPS fault = perimeter screening offlinePer-load UPS is uptime engineering — a fault on one support load cannot pull adjacent loads offlineShared-bus is power-quality engineering — cleaner sine wave, single point of failure · the brief decides which architecture is right, not the catalogue
Indicative comparison — illustrative pattern only, not a project-specific power design; actual topology follows the load schedule.

Diagrammatic view — a system planning illustration for design discussion, not a project drawing or live interface.

/ Integration with

How online ups talks to the rest.

A serious deployment of this system rarely operates in isolation. The disciplines below most commonly share its cabling pathways, its controller logic, and its cause-and-effect matrix.

/ Plan it right

Online UPS — getting the brief right.

Common mistakes to avoid

  • Sizing the UPS to the nameplate sum of every connected device instead of the measured demand and its inrush — the unit ends up either oversized and inefficient or tripping on startup surges.
  • Choosing battery autonomy as a round number instead of deriving it from the generator's real start-and-stabilise time — the battery exists to bridge exactly that window.
  • Ignoring the battery room — ventilation, ambient temperature and dust decide battery life as much as the chemistry does, and heat is the quiet killer of VRLA banks.
  • Buying the UPS, diesel set and ATS as three separate contracts, so the changeover sequence is never engineered or tested as one chain.
  • Skipping the maintenance-bypass provision, so the first service visit requires a full load shutdown.

What to share before a quotation

  • The load list — what must ride through an outage, its measured or estimated demand, and any motor or compressor inrush on the circuit.
  • Whether a diesel generator exists, and how long it really takes to start and stabilise.
  • The autonomy expectation — minutes to reach generator power, or a longer window for clean shutdown.
  • Room conditions where the UPS and batteries will live — space, ventilation, ambient temperature.
  • Any existing UPS or battery bank being replaced or migrated, with its age and service history.

/ Frequently asked

Online UPS — what buyers ask first.

Why double-conversion online and not line-interactive UPS?

Choose double-conversion online UPS for medical, broadcast and mission-critical loads. Line-interactive UPS uses a transfer switch that introduces a 4–10 ms gap when the grid drops. For most office equipment that gap is invisible; for medical devices, broadcast servers, fire-alarm panels, switching-mode power supplies running tight tolerances and any equipment with sensitive transformers, that gap registers and over time accumulates as failures. Double-conversion online removes the question: there is no transfer event because there was never a switching path. We use line-interactive only where the load can genuinely tolerate it.

How long should the battery autonomy be?

UPS battery autonomy is typically 10–15 minutes for buildings with diesel backup, 30–60 minutes for IT loads without generators, and longer with redundancy for critical-care clinical loads. We size it to the longer of two windows: the time your generator takes to start and stabilise, plus a 50% safety margin; or the time the load needs to shut down cleanly.

Lithium-ion or VRLA batteries?

Choose lithium-ion for new UPS deployments above 20 kVA — 3–4× cycle life, half the floor space and 30–45% lower 8-year lifetime cost than VRLA. 3–4× cycle life, half the floor space and a fraction of the maintenance versus VRLA. Initial cost is ~2× but lifetime cost is typically 30–45% lower over 8 years. For new installations above 20 kVA we almost always recommend lithium. For sub-20 kVA loads with short autonomy requirements, VRLA remains acceptable.

What's the AMC scope for an online UPS?

Online UPS AMC scope includes quarterly preventive checks — capacitor inspection, fan-bearing health, internal-temperature trend, log review and a battery-bank impedance test (or BMS report for lithium) — plus an annual full discharge test, firmware patching on schedule and a spare-board pool for the modules in your rack. Response targets are documented in writing for the Guwahati metropolitan area and structured around scheduled rotations elsewhere.

Do you handle the diesel set and ATS coordination?

Yes — we coordinate the UPS with the building's diesel generator and automatic transfer switch as one engagement: runtime calculations, ATS logic, harmonic-load coordination and the documented start-up sequence. The whole power chain is treated as one system rather than three contracts.

Can you migrate an existing VRLA bank to lithium without replacing the UPS?

Yes in most cases — a careful audit of UPS firmware, charging-curve compatibility, communications interface and physical mounting tells us whether the existing UPS supports lithium without replacement. Many recent generations of Vertiv, Delta and Fuji modules do. We provide the audit report before quoting; if the UPS itself needs replacement we will say so.

Does an online UPS protect equipment against lightning?

It is one layer, not the whole answer. Double-conversion isolates the load from sag, surge, harmonics and frequency drift on the mains, but a direct or nearby lightning strike calls for a coordinated protection design — surge protection at the distribution boards and a sound earthing system — engineered alongside the UPS. In a high-lightning region like the Northeast we treat the two as one conversation.

Why do UPS batteries fail so much earlier than the datasheet says?

Almost always heat. VRLA battery life is rated at a controlled ambient temperature, and a widely used engineering rule of thumb is that service life roughly halves for every ten degrees above it — so a battery bank in a hot, unventilated room quietly loses years. Ventilation, charging settings and periodic impedance testing are what close the gap between datasheet life and real life.

Should the UPS back up the whole building or only the critical loads?

Usually the critical loads. The honest design separates essential circuits at the distribution board — servers, clinical equipment, fire and security panels, the network stack — and sizes the UPS to those, letting the diesel set carry the general building load after its start-up window. Backing everything through the UPS inflates the battery bank for loads that can tolerate a brief gap.

· Begin

Begin a
online ups
brief.

Tell us about the building, the timeline, and what success looks like a year after handover. We will reply within two working days with a written response, not a sales pitch.