DC-UPS or AC-UPS? Choosing the Right Backup Architecture for Critical Loads

When a site loses mains power, the question we are asked is rarely "do we need a UPS?" It is "which kind, and at what point in the power chain?" Engineers specifying control cabinets, telecom shelters, medical equipment and industrial automation face this decision constantly, and the answer is less obvious than product datasheets suggest. Here is how we frame it.

The difference is where the conversion happens

An AC-UPS sits between the mains and an AC load. It rectifies incoming AC to charge a battery and inverts DC back to AC at the output. A DC-UPS sits downstream of an AC/DC power supply and holds up a DC bus directly — typically 12 V, 24 V or 48 V. The battery lives on that bus, and when mains fails there is no inversion step at all, only a transfer of the bus from the supply to the battery.

That single structural difference drives almost every practical trade-off that follows.

Where a DC-UPS is the better engineering choice

Most modern equipment is DC internally. PLCs and their I/O, routers, radios, access control, sensors and instrumentation typically run on 24 V or 48 V DC. Backing those loads with an AC-UPS means the energy travels DC to AC and back to DC again — two extra conversion stages, each with its own losses and each an additional failure point. A DC-UPS removes both.

Efficiency is the second argument. Inverter-based systems are least efficient at light loading, which is exactly where a control cabinet or a small telecom node usually sits — often at a fraction of the UPS nameplate rating. A DC-UPS carries no inverter overhead, so its efficiency curve is far flatter across a lightly loaded day.

Transfer behaviour matters too. There is no output waveform to synchronise or switch; the bus is simply held while the source changes. For sensitive digital equipment this removes a whole category of transfer-related nuisance trips. At the low power levels typical of automation and telecom, the DC approach is also smaller and less expensive to install.

Where an AC-UPS is still the right answer

Some loads are genuinely AC: motors, compressors, HVAC, pumps, imaging equipment and standard IT hardware with fixed AC power supplies. Re-engineering those to a DC bus is rarely justified. A mixed load list with a handful of AC-only items usually points to an AC-UPS as well, simply because splitting the backup across two architectures adds complexity that has to be maintained for years.

There is also a physics limit. Power is voltage times current, so as the load grows, a low-voltage DC bus needs a great deal of copper. A 5 kW load on a 24 V bus draws more than 200 A — conductor sizing, volt-drop and connection quality quickly dominate the design. Moving to 48 V halves that, but beyond a few kilowatts, AC distribution with an AC-UPS is usually the more economical structure.

Size for hold-up time, not for amp-hours

The most common sizing mistake we see is starting from a battery that is already on the shelf. Start instead from the load profile in watts and the autonomy the process actually requires — is it thirty seconds to ride through a transfer to a generator, ten minutes for an orderly shutdown, or eight hours of unattended operation? Those are three different systems.

Then check the window, not just the capacity. A 24 V bus that sags below the minimum input voltage of the equipment has ended the backup event, even if the battery still holds charge. Add margin for temperature — capacity falls in the cold — and for ageing; designing to a usable fraction of the rated capacity rather than the full nameplate figure is what keeps a system meeting its specification in year five rather than only in year one. Finally, account for inrush and for the worst case where every load starts simultaneously after the outage.

Chemistry, and the failure mode that actually costs money

Sealed lead-acid remains attractive on first cost and is well understood, but its float life is strongly temperature-dependent and its usable depth of discharge is limited. LiFePO4 offers longer cycle life, a wider usable window and far better behaviour in warm cabinets, at a higher purchase price and with a mandatory requirement: a proper battery management system and the correct charge profile. Neither is universally right — the deciding factors are ambient temperature, expected number of discharge events and the service interval you can realistically commit to.

Whichever chemistry you choose, the expensive failure is almost never a dramatic one. It is the battery that had quietly degraded and was flat on the one morning it was needed. Continuous monitoring of voltage, current and temperature, with alarms and periodic capacity verification, is what converts a backup system from an assumption into a measurement. If the architecture debate above has a single conclusion, it is this: specify the monitoring at the same time as the battery, not after the first outage.

If you are specifying backup power for a critical load and would like a second opinion on architecture, sizing or chemistry, write to us at info@lmelectro.com. We reply to every enquiry within 24 hours.