STS vs ATS in Data Centers: The Transfer Point Matters More Than the Acronym

A transfer switch can operate exactly as designed and the IT load can still go down.

That sounds contradictory until you stop looking at the switch in isolation.

In a data center, the important question is not simply whether a device can move the load from Source A to Source B.

It is whether the entire power path can make that transition in a way the downstream load can tolerate.

That is where discussions about STS and ATS often become misleading.

People reduce the choice to one sentence:

STS is fast. ATS is slow.

There is some engineering history behind that distinction, but it is not enough to design a critical-power system around.

The better place to start is:

Where is the transfer happening, what is connected downstream, and what must remain powered while the sources change?

STS and ATS Are Not Just Two Speeds of the Same Switch

An STS — Static Transfer System or Static Transfer Switch in common industry usage — uses static switching technology to transfer a load between available AC sources.

IEC 62310-1 covers static transfer systems intended to maintain continuity of power to a load through controlled transfer between two or more independent AC sources.

The standard explicitly allows the transfer to occur with or without interruption.

That matters because an STS should not be described as producing literally zero interruption under every possible operating condition.

In many data-center STS products, semiconductor devices such as SCRs or thyristors perform the switching.

This makes very fast transfer possible in suitable applications.

Specific commercial STS products publish transfer performance in the millisecond or sub-cycle range.

Those values are product-specific examples, not a universal specification for every STS.

ATS is a different equipment category.

IEC 60947-6-1 covers transfer switching equipment used to transfer a load between power supply sources.

The current edition also distinguishes static transfer switches as equipment covered by the IEC 62310 series.

That gives us a useful technical distinction:

STS and ATSE are not simply the same device with different speed settings.

But it still does not mean that every device called ATS takes hundreds of milliseconds to transfer.

The Word “ATS” Covers More Than One Kind of Application

This is where data-center terminology gets awkward.

A facility engineer may hear “ATS” and think of a large transfer switch between utility and generator sources.

An IT engineer may hear the same acronym and think of a rack automatic transfer switch sitting beside servers.

Those are different operating problems.

A facility-level source transfer may need to determine that a source is unacceptable, initiate or wait for an alternate source, confirm that the alternate source is suitable and then move the load.

The total interruption experienced by the system can therefore involve much more than the switching mechanism itself.

At rack level, the situation can be completely different.

There are Rack Automatic Transfer Switch products designed specifically to provide redundant source selection for single-corded IT equipment, including products with transfer times below 10 milliseconds.

That is why I would not write a data-center specification around:

ATS = slow

and

STS = fast.

The actual equipment and architecture matter more than the acronym.

Start With the Load

Before choosing the transfer device, look downstream.

What happens if the incoming waveform disappears briefly?

How much disturbance can the load tolerate?

Does the load have its own energy storage?

Does it have one power input or two?

Does it contain redundant power supplies?

What does its manufacturer specify for source transfer and ride-through?

This is where generic statements about “servers can survive X milliseconds” become dangerous.

Different power supplies, operating loads and equipment designs can behave differently.

If the project depends on a specific transfer interval, the correct value should come from the actual equipment requirement or an approved project specification — not from a generic internet rule saying every server has the same hold-up time.

The design target should therefore be:

transfer performance compatible with the actual downstream load.

Not:

a switch with the smallest number in the brochure.

Dual-Corded and Single-Corded Equipment Create Different Problems

A server with genuinely independent A and B power inputs already gives the electrical designer options that a single-corded device does not have.

If each internal power supply is connected to a genuinely independent path, the equipment itself may be able to continue operating when one path disappears.

The important word there is genuinely.

Two power cords do not create redundancy if both eventually depend on the same upstream breaker, panel, UPS, bus section or transfer point.

Single-corded equipment is different.

It cannot simply receive one independent supply on each input because it only has one input.

That is one reason transfer devices are sometimes used near critical single-corded loads.

An STS can serve a similar architectural purpose at another point in the critical-power distribution system.

But the design question remains the same:

Where should two independent paths become one?

That is the transfer point that deserves the most attention.

Where the Transfer Happens Changes the Design

Imagine two independent UPS output paths.

There are several ways to bring those paths toward the IT load.

One design may maintain A and B separately all the way to dual-corded servers.

Another may use an STS upstream of downstream distribution so that a single output path can select between two protected sources.

Another may preserve two paths further downstream and use a rack-level transfer device only for equipment that has a single cord.

All three can appear in real projects.

They are not interchangeable.

Moving the transfer point upstream can place more downstream loads behind one switching device.

Moving it closer to the rack can reduce the size of the affected load group but increase the number of transfer devices that must be installed, monitored and maintained.

So instead of asking:

Should the data center use STS or ATS?

I would ask:

At what point does the architecture actually need two sources to become one?

That question usually reveals much more.

ATS Before the UPS and STS After the UPS? Sometimes — Not Always

A common simplified diagram shows:

Utility / Generator → ATS → UPS → STS → IT Load

It is useful as a teaching diagram.

It should not become a universal rule.

A facility-level automatic transfer device can make sense when the system needs to select between incoming utility, generator or alternate facility sources.

A static transfer device can make sense downstream where two conditioned sources feed a load that requires fast source selection.

But real projects can also include dual utility feeds, multiple generators, distributed redundant UPS architectures, maintenance bypass paths, dual-corded IT equipment and rack-level transfer devices.

The correct topology depends on the continuity strategy of the whole facility.

A transfer switch does not create redundancy by itself.

It only makes use of the sources and paths that actually exist around it.

Source Quality Matters During the Transfer

Transfer time gets most of the attention.

Source condition deserves just as much.

Before switching, the system needs to know whether the alternate source is acceptable.

Voltage matters.

Frequency matters.

Phase relationship may matter.

The behavior of the load during transfer matters.

Protection and fault conditions matter.

Some transfer devices include logic for deciding when the alternate source should be accepted and when a transfer should occur.

That is why the datasheet line labelled transfer time cannot be read alone.

A fast transfer to an unsuitable source is not a successful transfer.

Maintenance Can Change the Answer

One thing is often missing from STS-versus-ATS comparisons:

What happens when the transfer equipment itself needs maintenance?

If a critical load depends on one transfer point, that device has become part of the critical path.

The project therefore needs to understand how the equipment can be isolated, bypassed, inspected or replaced.

The same applies to the upstream sources.

If Source A is taken out of service, can the load remain on Source B?

Can the transfer device itself be serviced while the load remains supported?

Are bypass arrangements part of the selected equipment?

Does the maintenance procedure introduce a temporary single point of failure?

These questions are less interesting than a simple millisecond comparison.

They are often more important after the facility is live.

A Faster Switch Does Not Repair a Weak Power Architecture

Suppose a project installs an extremely fast transfer device.

Source A and Source B are both available.

Everything looks redundant.

Then someone traces the single-line diagram and discovers that both sources depend on the same upstream distribution section.

The switch is fast.

The architecture is still not redundant.

Or perhaps the two sources are independent, but both downstream paths eventually pass through the same feeder.

Again, the transfer device cannot solve that problem.

Redundancy at one stage does not automatically mean the entire downstream system is redundant.

The same principle applies here.

STS and ATS are switching tools.

They are not substitutes for a power-path review.

So When Does an STS Make Sense?

An STS becomes especially relevant when two suitable AC sources are available and a sensitive downstream load requires rapid automatic source selection.

That may occur within a UPS-backed distribution architecture, around precision distribution equipment or at another deliberately selected transfer point.

The actual suitability still depends on the selected product, source conditions, load characteristics, fault requirements, protection coordination, bypass arrangement and project standard.

CONLUXS currently includes the JSTS Series Static Transfer Switch within its data-center power distribution portfolio.

The product should be selected against defined project requirements rather than treated as a universal replacement for other transfer architectures.

That distinction is important.

An STS is not simply “better ATS.”

It solves a different transfer problem.

The Five Questions I Would Put on the Single-Line Diagram

Before selecting either device, I would write these questions next to the proposed transfer point:

What exactly are Source A and Source B?

Utility, generator, UPS outputs or independent distribution paths are not equivalent.

What load sits downstream?

Single-corded IT equipment, dual-corded servers, mechanical loads and general facility loads can have very different continuity requirements.

What interruption or disturbance can that load tolerate?

Use the actual equipment or project requirement, not a generic hold-up-time assumption.

What happens when one source — or the transfer device itself — is under maintenance?

The normal diagram and the maintenance diagram should both work.

Does the complete path remain independent upstream and downstream of the transfer point?

If not, a fast transfer device may simply hide an existing single point of failure.

Once those answers are clear, the equipment choice becomes much easier.

The Real Difference Between STS and ATS

The useful distinction is not:

STS means no interruption. ATS means interruption.

And it is not:

STS belongs beside every server. ATS belongs before every UPS.

A better way to think about them is this:

An STS is a static transfer system intended for controlled transfer between AC sources where continuity and transfer performance are central parts of the application.

An ATSE automatically moves a load between power sources under its defined transfer logic and operating conditions.

The device that belongs in a particular data center depends on where that transfer is happening and what the downstream load expects from it.

Sometimes that points clearly toward an STS.

Sometimes an automatic transfer device at facility or rack level is entirely appropriate.

Sometimes the best answer is to maintain two independent paths and avoid introducing another common transfer point at all.

The acronym comes later.

The architecture comes first.

Planning a Critical-Power Transfer Point?

For a useful transfer review, start with the single-line diagram rather than the equipment catalogue.

Define the two available sources, UPS topology, downstream load type, single- or dual-cord arrangement, required continuity behavior, voltage and frequency, protection arrangement, maintenance strategy and the intended location of the transfer point.

CONLUXS can review defined STS and low-voltage distribution interfaces with the project team and discuss an appropriate equipment configuration.