ИБП в режиме ожидания или ИБП в режиме онлайн: какая топология подходит для вашего серверного помещения?

The Real-World Problem: A Server That Blinks Off

Picture this: a small server room in a suburban office. The IT manager installed a cheap standby UPS because the budget was tight. One afternoon, a forklift in the warehouse next door bumps a breaker. The lights flicker for 300 milliseconds. The server reboots. That's 45 minutes of downtime, a dozen angry users, and a support ticket that eats the rest of the day. The standby UPS did its job—it switched to battery—but the transfer took too long, and the server's power supply couldn't ride through the gap.

This is the classic standby UPS versus online UPS debate. The short answer: for anything running 24/7—servers, storage, network gear—online double conversion is usually the right call. But it's not always that simple. Let's break down how each topology works, where it shines, and where it falls short, so you can make an informed choice for your server room.

How Standby UPS Versus Online UPS Actually Work

The core difference comes down to how the load is powered under normal conditions. A standby UPS—sometimes called offline or line-preferred—feeds utility power straight to the load through a filter. The inverter stays off until the input voltage drops or disappears. When that happens, a transfer switch flips the load to battery power. That switch takes time—typically 4 to 10 milliseconds, sometimes up to 20 ms on cheaper units.

An online UPS, also known as double conversion, runs the load through the inverter at all times. The input AC is rectified to DC, then inverted back to clean AC. The battery sits on the DC bus, so when utility fails, there's zero transfer time. The load never sees a gap. That's the fundamental advantage: continuous, conditioned power with no interruption.

There's also a middle ground: line-interactive UPS. These units regulate voltage with an autotransformer and only switch to battery when the input fails. Transfer time is similar to standby, but voltage regulation is better. For some IT loads, line-interactive is acceptable, but it still has that gap.

Key Differences: Transfer Time, Waveform, and Voltage Regulation

Let's get into the numbers. Transfer time matters because modern server power supplies have hold-up times—typically 10 to 20 ms—but they're not guaranteed. If your standby UPS takes 15 ms to transfer, you're flirting with disaster. Online UPS has zero transfer time, which is why it's the standard for critical loads.

Output waveform is another differentiator. Cheap standby units often produce a stepped or simulated sine wave on battery. Many server power supplies can handle that, but some sensitive electronics—especially older equipment or certain medical devices—may misbehave or even shut down. Online UPS always outputs a true sine wave, because the inverter is always running.

Voltage regulation is where online UPS really shines. Standby units typically pass through utility voltage until it goes out of a narrow window—say ±10%—and then they switch to battery. That means your servers might be running on utility power that's sagging or spiking, which can stress power supplies over time. Line-interactive units correct moderate fluctuations with taps, but online UPS regulates output to ±1-2% regardless of input, because it's always regenerating the waveform.

Особенность	Standby UPS	Line-Interactive UPS	Online Double Conversion UPS
Transfer time	4-10 ms (up to 20 ms)	2-6 ms	0 ms
Output waveform on battery	Stepped or simulated sine wave	Stepped or pure sine wave (varies)	Pure sine wave always
Voltage regulation	None until switch	Taps ±10-20%	Regulated ±1-2%
Efficiency (typical)	95-98%	95-98%	90-95%
Cost per kVA	Низкий	Средний	Высокий
Typical use	PCs, home office	Small servers, network gear	Server rooms, data centers, critical loads

Efficiency is the trade-off. Online UPS burns more energy because it's constantly rectifying and inverting. A 10 kVA online unit at 92% efficiency wastes about 800 W as heat. Standby units are more efficient—up to 98%—but they don't condition power. That extra heat means your cooling system has to work harder, which eats into the efficiency savings. In a small server room, the difference might be a few hundred dollars a year in electricity, not nothing, but usually worth it for protection.

Why Standby UPS Is Often Inadequate for Server Loads

The problem with standby UPS for servers isn't just transfer time. It's also the lack of frequency regulation. If your utility frequency drifts—which can happen with generators or unstable grids—a standby UPS will pass that drift through. Online UPS regenerates at a stable 50 or 60 Hz, which some equipment requires.

Another issue: standby UPS units often have limited surge protection and minimal filtering. They might clip a spike, but they don't actively correct brownouts. In areas with frequent voltage sags, your servers might be constantly running on the edge. I've seen standby UPS units that switch to battery dozens of times a day because the voltage dips just below the threshold. That wears out the battery fast and increases the chance of a failed transfer.

Worth checking: the input voltage window on a standby UPS. Many are set to switch at ±10%, which means if your utility runs at 115 V in a 120 V system, the UPS might see that as normal and pass it through. But if it drops to 108 V, it switches. Some servers can handle that, but others can't. Online UPS eliminates this variability entirely.

When Standby UPS Makes Sense (and When It Doesn't)

Let's be fair. Standby UPS has its place. For a single workstation, a router, or a small network switch that can tolerate a few milliseconds of interruption, a standby unit is fine. It's cheap, efficient, and easy to replace. If you're protecting a home office or a non-critical lab, don't overspend on online double conversion.

But for a server room—even a small one with a couple of servers—the stakes are higher. A single reboot can cost hours of productivity, and if you're running databases or virtual machines, the risk of corruption is real. In that context, the extra cost of an online UPS is insurance. The rule of thumb I use: if the load is critical and runs 24/7, go online. If it's non-critical and can tolerate a blip, standby is acceptable.

There's also the generator consideration. If you have a backup generator, you need a UPS that can handle the generator's frequency and voltage swings. Online UPS units are much better at this because they decouple the load from the input. Standby units often struggle with generator power, causing nuisance transfers or even damage.

Practical Checklist Before You Choose

Before you pick a topology, run through this checklist. It'll save you from a costly mistake.

1. Identify the criticality of the load. Will a 10 ms interruption cause downtime or data corruption? If yes, online is required.
2. Check the input voltage range of your equipment. Some server PSUs can handle 100-240 V, but others are narrow. Compare with the UPS's transfer window.
3. Measure the actual load in watts and VA. Don't guess. Use a power meter or check nameplates. Size the UPS at 1.2-1.5 times the load to avoid overloading.
4. Consider the runtime you need. Calculate battery capacity based on load and desired runtime, not just the UPS rating.
5. Assess power quality in your area. If you have frequent sags or surges, online UPS will protect better.
6. Think about generator compatibility. If you have a generator, verify the UPS can handle its output.
7. Check efficiency and cooling. Online UPS generates more heat; ensure your cooling can handle it.
8. Look at the output waveform. For sensitive loads, require pure sine wave.
9. Plan for maintenance and battery replacement. Online UPS units often have hot-swappable batteries, but check.

That's a lot to consider, but it's worth the effort. The cost of a wrong choice is downtime, and downtime is always more expensive than the UPS.

Sizing and Runtime Considerations

Sizing a UPS isn't just about watts. You need to know the VA rating and the power factor. Most modern servers have a power factor around 0.9-0.99, but older ones might be lower. A 10 kVA UPS at 0.9 power factor gives you 9 kW. Don't oversize too much—running a UPS at 20-30% load reduces efficiency and can cause battery issues. Aim for 60-80% load for best performance.

Runtime is a function of battery capacity. A typical server room might need 15-30 minutes to shut down gracefully or switch to a generator. You can add external battery cabinets to extend runtime, but that adds cost and floor space. Online UPS units are more efficient at battery use because they don't have a transfer loss, but the difference is small.

Worth checking: the battery charging time. Some UPS units recharge slowly, which means if you have multiple outages in a day, you might not have full runtime. Look for units with fast recharge or consider a larger battery bank.

The Verdict: Standby UPS Versus Online UPS for Servers

For server room loads, online double conversion UPS is the safer choice. The zero transfer time, consistent output waveform, and tight voltage regulation protect your equipment from the everyday quirks of utility power. Yes, it costs more and uses more energy, but the protection is worth it. Standby UPS has its place, but that place isn't in a rack with critical servers.

That said, don't just buy the first online UPS you see. Check the specs, match the load, and think about how the UPS will interact with your generators and cooling. A well-chosen online UPS will give you years of reliable service. A poorly chosen standby UPS might give you a headache on day one.

VERHI offers a range of UPS solutions, including online double conversion models for server rooms and micro-module data centers. Our engineers can help you size and select the right topology for your specific load. Contact us to discuss your requirements.