Welcome to HG Group official website – your trusted supplier of data center power solutions
ABOUT US

Industry Knowledge

What Is Transfer Time and Why Zero Is Best?

UPS transfer time is the interval between a utility failure and the moment the load is fed from battery, and a zero-millisecond transfer is best because the protected equipment never experiences the interruption at all. Only online double-conversion topology delivers a true zero transfer; line-interactive and standby designs always spend a few milliseconds switching, which matters for sensitive loads even when it is invisible to a desktop.

What "Transfer Time" Actually Measures

Transfer time is the gap between the input supply dropping below the acceptable window and the inverter taking over the output. During that gap the load is either briefly on a raw, unconditioned source or momentarily open. The figure is quoted in milliseconds because the events that cause it, a voltage sag or a grid switch, unfold faster than a human blink. A spec of 0 ms means the output is already being synthesized from the battery-backed DC bus before the fault arrives, so there is nothing to bridge.

Where the Gap Comes From Inside the Unit

Every topology pays a switching cost unless it is already running on the inverter full time. A standby unit must detect the loss, release the bypass relay and connect the inverter, which takes several milliseconds. A line-interactive unit must open the mains path and engage its buck-boost or inverter stage, a smaller but still real delay. An online unit never leaves the inverter path, so the moment the rectifier loses input, the DC bus simply draws from the battery with no electromechanical change at the output. The difference is architectural, not a tuning setting.

Why Even Milliseconds Matter for Some Loads

A few milliseconds is harmless to a fan or a monitor, but it can reset a programmable logic controller, drop a network session, or trip a medical device that expects a continuous supply. Servers with wide-input power supplies often ride through tens of milliseconds, yet storage arrays and industrial controllers are far less forgiving. The safe engineering rule is to assume the worst-case load, not the average one: if any single device on the circuit cannot tolerate a gap, the whole circuit needs zero transfer.

Online Double-Conversion and the Zero-Transfer Design

In an online unit the load is permanently powered by the inverter, which in turn is fed by the rectifier when mains is healthy and by the battery the instant it is not. Because the output stage never changes state, the transfer is genuinely 0 ms and the output voltage and frequency stay regulated throughout. This is why hospitals, data halls and process lines standardize on online topology: the question of transfer time disappears from the risk register rather than being managed down to an acceptable number.

Line-Interactive and the Buck-Boost Switch

Line-interactive designs are popular for small server rooms because they save energy and cost while still conditioning most sags. They sit on the mains and only break to the inverter when the supply leaves tolerance, so their transfer is typically two to four milliseconds. That is enough for general IT but too much for the sensitive equipment noted above. Treat line-interactive as the right answer for resilient office loads and the wrong answer for anything described as critical or continuous in its own standard.

Standby Units and the Longest Gap

Standby, or offline, units are the cheapest and switch last. They pass mains straight through until a fault is detected, then transfer to the inverter, a gap of four to ten milliseconds. They protect against a full blackout and give time for a graceful shutdown, but they do nothing for the sub-cycle sags that actually damage control gear. Use them only for single-user equipment where a brief flicker is an annoyance, not an incident.

How to Read a Vendor's Transfer Specification

A datasheet line of "0 ms" is meaningful only for online topology; on a line-interactive or standby unit it usually refers to the transfer between inverter and bypass under normal conditions, not to a mains failure. Always read the spec next to the topology name, and ask for the switchover behaviour under a deep sag versus a hard break. Independent test reports that show output voltage during a simulated fault are worth more than a marketing number, and our engineers can walk through the test method on the https://www.upsboss.com/faq/ page.

Matching Transfer Time to the Load

Build the choice from the load list. Desktop and point-of-sale equipment tolerates standby or line-interactive. Shared servers and storage want line-interactive at minimum and online for the demanding arrays. Life-safety, process and data-center loads require online with zero transfer by default. Size the decision on the single most sensitive device on the circuit, then select the whole bank to that standard. The full hardware range, organized by topology and transfer behaviour, is on the https://www.upsboss.com/products/ listing.

If you are unsure whether your load tolerates a switching gap, send us the equipment make and model and we will confirm the tolerated interruption and recommend the lowest-cost topology that meets it. Review the topology comparison on the https://www.upsboss.com/products/ page before you decide, and raise any edge case through the https://www.upsboss.com/faq/ section so the choice is documented.

Key takeaway: transfer time is the gap between a power fault and battery takeover, and zero is best because the load never sees the cut. Choose online topology for any sensitive or critical device, and treat line-interactive or standby only for loads that can genuinely ride through a few milliseconds.

Frequently Asked Questions

Is a 0 ms transfer possible on a line-interactive UPS?
No. A line-interactive unit must change its power path when mains leaves tolerance, which takes a few milliseconds. Only an online double-conversion design, which runs the inverter continuously, reaches a true zero transfer.

Do servers need zero transfer time?
Many servers ride through tens of milliseconds, but storage arrays and controllers often do not. If any device on the circuit is sensitive, the whole circuit should be on online topology so the weakest link sets the standard.

Why does a datasheet say 0 ms on a cheap unit?
That figure usually describes the bypass-to-inverter transfer under normal conditions, not a mains failure. Read the spec beside the topology name and ask for output behaviour during a simulated sag before trusting the number.