The UPS Works Normally on Mains but Shuts Off the Moment Power Fails—Why Doesn’t the LiFePO4 Battery Take Over?

Introduction

Everything looks normal while utility power is available. The UPS display is on. The connected equipment works. The LiFePO4 battery shows a healthy SOC, and the UPS may even indicate that the battery is fully charged. Then grid power fails.

Instead of switching seamlessly to battery backup, the entire UPS shuts off almost instantly. When utility power returns, the UPS starts again as though nothing happened. For an end customer, this is one of the most frustrating battery problems imaginable because the entire purpose of the UPS is to prevent exactly this interruption.

The customer often concludes that the new LiFePO4 battery “has no capacity.” In reality, an instant shutdown at the moment of grid loss usually points more strongly toward battery availability, discharge current, DC voltage collapse or transfer behavior than toward total energy capacity.

A battery with low capacity might provide only ten minutes instead of one hour. A system that dies in a fraction of a second often has a different problem.

Mains Operation Can Hide a Battery Problem

While utility power is present, the UPS may supply the load primarily from AC input. The battery is on standby. Even if the battery connection is poor or the BMS discharge path is disabled, the customer may not notice because the battery is not being asked to support the full load.

The UPS screen can still display battery voltage or SOC. Only when the grid disappears does the system discover whether the battery can actually provide power. This is why an UPS should be tested on battery after installation rather than judged only from normal mains operation.

Charging Successfully Does Not Prove Discharging Successfully

A LiFePO4 BMS can sometimes control charge and discharge independently. A battery may accept charging current while its discharge path is disabled because of a protection state or configuration. The customer sees the battery charging every day and assumes it must be ready for an outage.

Then the grid fails and the UPS receives no usable battery current. This is another reason that commissioning needs both a charging test and a controlled discharge test. One direction working does not prove the other direction works.

The Battery May Be Connected Electrically but Not Capable of the UPS Load

Suppose the battery shows normal voltage and no obvious alarm. When grid power fails, the UPS instantly demands 100A from the battery. If the BMS allows only 80A continuous current, it can enter overcurrent protection immediately.

Battery output disappears and the UPS shuts down. Once utility power returns, the battery resets or no longer sees the heavy discharge current. Everything looks normal again. This pattern is commonly misdiagnosed as a mysterious intermittent battery failure.

Calculate the Instant Battery Current

The first calculation should use the real UPS load. A simplified estimate is: Battery Current ≈ UPS AC Load ÷ UPS Efficiency ÷ Battery Voltage

Suppose the AC load is: 5kW

UPS battery-mode efficiency is approximately: 90%

Battery voltage during discharge is: 50V

Approximate battery current is: 5,000 ÷ 0.90 ÷ 50 ≈ 111A

A battery with a 100A continuous BMS has little or no operating margin. If the load contains startup surges, the peak current can be substantially higher.

The Load May Jump When the Grid Fails

Some installations have equipment that behaves differently during an outage. For example, emergency fans, pumps, communication equipment or lighting may start automatically when utility power fails.

The customer calculates backup requirements from the normal load but the actual battery-mode load is higher. This can cause an immediate UPS or BMS overload. Record the UPS output power at the moment of transfer rather than relying only on normal grid-mode power.

AC Power Supplies Can Create Transfer Surges

Servers and electronic equipment contain input capacitors and power-factor-correction circuits. Their AC input behavior during UPS transfer can create short transient loads. The UPS itself may be rated to handle these surges, but the corresponding DC current also has to come from the battery.

A battery bank designed with no peak-current margin can trip during the first second of an outage even though the subsequent steady load would have been acceptable. This is especially important when one relatively small LiFePO4 battery replaces a large VRLA string.

Lead-Acid Banks Can Deliver Very High Short-Term Current

A traditional UPS battery bank may contain many lead-acid batteries connected in series and parallel. The old bank may have been heavy and inefficient, but it could still provide substantial short-duration current. An installer may replace it with a much smaller LiFePO4 bank based only on equivalent kWh.

The new battery may provide similar or greater usable energy but have an electronic BMS current limit. If the BMS rating is not matched to UPS power, the lithium system can shut down instantly even though its energy calculation looked correct. This is one of the most important lessons in UPS lithium conversion: match both energy and current.

BMS Short-Circuit Protection Can Trigger During Startup

If current rises extremely quickly, the BMS can interpret the event as a short circuit or severe overcurrent. The battery disconnects within milliseconds. The UPS may not have enough stored energy internally to keep the load running until the battery resets.

The result is a complete outage. Check BMS event logs for discharge overcurrent or short-circuit protection at the same timestamp as utility failure. That evidence is far more useful than an SOC screenshot taken later.

Low Battery Voltage Can Cause Instant UPS Shutdown

Not every instant failure is overcurrent. Suppose the battery is at low SOC. Its resting voltage still looks acceptable. When the full UPS load is applied, battery voltage sags sharply. If the UPS DC input immediately falls below its low-voltage shutdown threshold, the UPS stops.

Once load disappears, voltage rebounds. The customer measures the battery afterward and says: “The battery still has 50V.” That does not tell us what voltage the UPS saw during the failed transfer. Loaded voltage is what matters.

One Weak Cell Can Cause the Whole Pack to Drop Out

A pack can show healthy total voltage while one cell is significantly weaker. At rest, the difference may be small. The moment a heavy UPS load is applied, the weak cell voltage collapses.

The BMS detects cell undervoltage and opens the discharge path. The UPS shuts down almost instantly. After current stops, the weak cell voltage rebounds, making the battery look normal again. For this symptom, individual cell voltage during transfer is highly valuable.

Cable Resistance Can Cause an “Instant” Failure

UPS systems operating around 48–51.2V can use very high DC current. Even a relatively small cable resistance creates significant voltage drop. Suppose current jumps from almost zero on mains to 150A on battery.

A poor cable connection that was invisible during standby suddenly creates several volts of drop. The UPS sees its DC bus collapse and shuts down. The battery itself may remain fully operational. Measure voltage both at the battery terminals and UPS terminals during the transfer attempt.

Main Breaker Rating Can Be Too Small

The battery BMS may support 200A, but the external battery breaker may be rated or configured for substantially less. When grid fails, full UPS current flows through the breaker and it trips. The customer sees the battery and UPS go offline together.

This is an external protection mismatch. The entire current path must support the required battery current: internal BMS, terminals, cables, fuse, breaker and UPS input. The lowest-rated component determines system capability.

Parallel Batteries May Not All Be Online

A system may be designed for four parallel 51.2V batteries but only two are actually active. Perhaps one breaker is off and another module has a BMS alarm. On mains power, the UPS still sees normal bus voltage.

When an outage occurs, the two remaining batteries suddenly have to support the full load. Each sees twice the intended current and can trip. Always verify active module count rather than assuming every installed cabinet is electrically participating.

Master BMS Communication Can Affect Discharge Permission

Some multi-battery systems use a master BMS to coordinate all modules. If master communication is not correctly established, the system may allow limited startup behavior but refuse full discharge.

The UPS can appear normal on mains while the battery bank is not actually ready for backup operation. Check master/slave addresses, communication status and the BMS’s Discharge Current Limit where available. A displayed SOC alone does not prove discharge permission.

Transfer Time Can Matter for Certain Loads

UPS systems are designed to transfer power within a specified time depending on topology. Online double-conversion UPS equipment typically behaves differently from line-interactive or standby architectures. If the connected load is extremely sensitive, a transfer interruption can cause the load to shut down even though the battery begins supplying power shortly afterward.

In this case the UPS itself may remain operating while the connected device restarts. This should be distinguished from a true battery-side UPS shutdown. Check whether the UPS remains alive during the event.

How to Tell Whether the UPS or the Load Shut Down

Ask the customer exactly what happens when the grid is disconnected. Does the UPS display turn off completely? Does it show an alarm? Does it remain on battery while only the servers restart? Does the battery BMS alarm?

These differences matter. If the UPS remains operating and battery current is normal, the problem may concern UPS transfer performance or the connected load. If battery output disappears simultaneously, investigate the battery/DC side first.

An UPS Self-Test Can Reveal the Problem Before a Real Outage

Many UPS systems offer a controlled battery self-test. If the LiFePO4 battery cannot support even this test, there is no reason to wait for a real grid failure. However, a successful light self-test does not always prove full-load capability.

The real outage may produce much higher current. For important installations, commissioning should include a controlled battery-mode test at representative load.

Example: Energy Is Enough but BMS Current Is Not

A customer installs a 10kWh LiFePO4 bank for a 10kW UPS and expects around one hour of backup. From an energy perspective, the calculation looks reasonable. But the system operates near 50V.

At 10kW AC output, battery current can exceed 200A after inverter losses are included. If the battery bank allows only 150A continuous discharge, the BMS trips almost immediately. The customer never gets to use the 10kWh stored energy. The installation was energy-sized but not power-sized.

Example: Loose Battery Lug

Another UPS works normally on utility power. During a battery-mode test, voltage at the battery terminals remains around 52V while UPS DC input collapses below 40V. A loose lug becomes hot.

After correcting and torquing the connection, the UPS transfers normally. The battery was never defective. This is why voltage should be measured at both ends of the cable.

Example: One Cell Trips on Undervoltage

A third battery shows 80% SOC and normal pack voltage. When the UPS load is applied, one cell drops much faster than the others and reaches the BMS low-voltage threshold.

The BMS disconnects in less than a second. Afterward, the cell recovers. This requires battery-level investigation even though pack SOC appeared high. Pack voltage alone could not reveal the weak cell.

Why Replacing the Battery with a Larger Ah Model Can Sometimes Help

A larger-capacity battery can often provide more current capability, depending on its cell and BMS design. It also experiences lower C-rate for the same UPS load.

However, do not assume that every larger-Ah battery automatically has a higher BMS rating.

Check the actual continuous and peak discharge specifications. Sometimes the correct solution is multiple parallel modules rather than one larger battery. System architecture matters.

A Controlled Transfer Test Is Essential

After a LiFePO4 retrofit, perform a planned battery-mode test before releasing the UPS to the customer. Use an appropriate representative load and monitor: UPS AC output power, battery current, battery voltage, lowest cell voltage and BMS alarms.

Then disconnect mains according to the approved procedure. A successful test should show clean transfer, stable battery current and no protection alarms. This single commissioning step can identify problems that remain completely hidden during mains operation.

Dealer Diagnostic Table

Symptom at Grid FailureLikely Area
BMS overcurrent alarmBattery current capability
BMS short-circuit alarmSurge/inrush/current path
Cell undervoltage alarmLow SOC/weak cell/high sag
Battery voltage stable, UPS input collapsesCable/breaker/connection
DC breaker tripsProtection sizing/load current
UPS remains on but load restartsTransfer/load compatibility
Some parallel modules remain offlineBattery bank configuration
No BMS alarm but UPS shows DC undervoltageVoltage sag/UPS cutoff

Capturing the alarm at the exact moment of failure is critical.

For UPS Distributors: Never Validate a Lithium Retrofit Only on Mains

A battery can charge successfully for days and still fail the first real outage. Compatibility testing must include discharge.

For UPS dealers, a minimum practical validation should check battery-mode startup, realistic load power, low-SOC behavior and return-to-grid charging. A complete cycle is far more valuable than observing a green battery icon on the UPS display.

For Battery Dealers: Ask for Actual kW, Not Only UPS kVA

When customers request a battery quotation, ask for both UPS rating and actual connected load. A 10kVA UPS supporting only 2kW requires a very different battery current from the same UPS supporting 8kW.

Then ask for required backup time. The battery bank can be sized for: energy requirement + continuous current + peak current. This produces a technically defensible recommendation.

Conclusion

If an UPS works normally on mains but shuts down the instant utility power fails, the problem is usually not simply “insufficient battery capacity.”

More likely causes include BMS overcurrent protection, high startup current, low-voltage sag, a weak cell, undersized cables or breakers, inactive parallel modules or a discharge path that is not actually enabled.

The most useful diagnostic data comes from the first second of the outage.

HIZN supplies LiFePO4 batteries for UPS, telecom and stationary backup applications with different capacity and BMS-current configurations.

If your customer’s UPS shuts down immediately when the grid fails, send us the UPS model, actual AC load, battery model/quantity, BMS discharge-current rating and alarm log at the moment of transfer. We can help determine whether the issue is battery power capability, DC installation or UPS behavior.

Facebook
Twitter
LinkedIn

Leave a Reply

Your email address will not be published. Required fields are marked *