Introduction
A customer complains that a LiFePO4 battery does not provide enough UPS backup time. The distributor tests the battery at the warehouse and finds that it delivers close to its rated Ah or kWh capacity.
The battery appears normal. The customer installs it again, and the UPS still shuts down much earlier than expected.
This situation can create a difficult argument. The end user says the battery capacity is insufficient, while the battery supplier says the laboratory test proves that the product meets specification. Both sides may be looking at valid data, because battery capacity testing and real UPS runtime are not necessarily the same test condition.
A controlled capacity test may discharge the battery at a relatively moderate current until a defined battery cutoff. A UPS can demand much higher current and may stop operating at a higher DC voltage before the battery’s full laboratory-rated energy has been used. Understanding this difference is essential for UPS dealers, battery distributors and system integrators.
Ah Capacity Does Not Directly Equal UPS Runtime
Battery capacity is often expressed in amp-hours: Ah
UPS load is normally expressed in: W, kW, VA or kVA
To estimate backup time, the system needs to be considered in terms of energy.
A 51.2V 100Ah LiFePO4 battery has nominal energy of approximately: 51.2V × 100Ah = 5.12kWh
That does not mean a 5kW UPS load will necessarily operate for exactly one hour. Some energy is lost in conversion, the battery voltage changes during discharge and the UPS may stop before every theoretical watt-hour is extracted.
UPS Efficiency Reduces Useful AC Energy
The battery provides DC energy. The UPS converts that DC into AC power for the connected equipment. No inverter is 100% efficient. If the UPS is operating at 92% efficiency, supplying 3kW of AC load requires more than 3kW from the battery.
A simplified estimate is: Battery Power ≈ AC Load ÷ UPS Efficiency
For a 3kW load: 3kW ÷ 0.92 ≈ 3.26kW
The battery therefore discharges faster than a simple 5.12kWh ÷ 3kW calculation suggests.
The UPS Itself Also Uses Power
In addition to conversion losses, a UPS has internal consumption from control electronics, fans, relays and other components. For a large UPS, auxiliary power can become noticeable during long backup periods.
The load meter may show only the customer equipment while the battery is supplying additional internal power to the UPS. For precise runtime analysis, use actual DC battery power or total input/output efficiency rather than only the external AC load.
High Load Means High DC Current
Low-voltage UPS systems can draw very high current from the battery. Suppose the battery voltage is around 51V and DC power demand is 5kW.
Current is approximately: 5,000W ÷ 51V ≈ 98A
At 10kW, current approaches 200A. This has consequences for battery voltage, BMS limits, cables and terminals. A capacity test performed at 20A is not electrically equivalent to a UPS discharge at 150A.
Battery Terminal Voltage Drops More at High Current
When current increases, voltage drop across internal resistance also increases.
A simplified relationship is: Voltage Drop = Current × Resistance
This resistance includes the battery cells, internal busbars, BMS components and external DC connections. At light load, voltage drop may be small. At high UPS current, the same system can experience a much larger drop. The UPS may therefore reach its low-DC shutdown threshold even though meaningful energy remains inside the battery.
Example: Capacity Test vs UPS Test
Suppose a 51.2V 100Ah battery is discharged in a controlled test at 20A. It delivers close to 100Ah before reaching the battery test endpoint. The same battery is then connected to a UPS drawing approximately 100A.
Under the higher current, battery terminal voltage is lower. The UPS has its own low-voltage cutoff and stops before the battery reaches the same endpoint used in the laboratory test. The battery can therefore pass its capacity test while providing less usable UPS energy. Neither result is necessarily incorrect. The tests are measuring performance under different conditions.
UPS Cutoff Voltage Can Leave Energy Unused
A UPS originally designed for lead-acid batteries may have a fixed low-battery shutdown voltage. LiFePO4 has a different discharge curve. The UPS threshold may not allow the lithium battery to use its intended SOC window.
For example, the UPS may shut down at a voltage where the LiFePO4 battery still has significant usable capacity. This creates the impression of an undersized battery. The solution is not automatically to lower the UPS cutoff as far as possible. The correct limit must remain above the battery’s safe discharge endpoint and within UPS specifications.
BMS Cutoff Can Do the Opposite
In another system, the UPS cutoff may be too low. The UPS continues drawing current until the battery BMS reaches cell undervoltage or discharge overcurrent protection. The BMS then disconnects abruptly.
From the customer’s perspective, the UPS goes from working normally to completely off. This is not ideal daily behavior. The UPS should normally stop or reduce load before the BMS’s emergency protection is required.
One Weak Cell Can Limit the Entire Pack
A battery pack contains multiple series-connected cells. During discharge, the same current passes through all series cells. If one cell has lower usable capacity or significantly higher resistance, its voltage can fall faster than the others. The BMS protects the entire battery when that cell reaches its undervoltage threshold.
The remaining cells may still contain energy, but it cannot be used safely because the weakest cell has become the limiting factor. A pack-level Ah result under light test current may not reveal the issue as clearly as high-power UPS operation. Cell-voltage logs under load can provide important diagnostic information.
BMS Current Rating Can Limit Runtime Under High Load
A 51.2V 100Ah battery with a 100A BMS is not automatically suitable for every UPS up to 5kW or more. At low battery voltage and after conversion losses, the required current may exceed 100A. The BMS can then trip on overcurrent even though the battery SOC is still high.
The customer reports: “The battery still showed 70%, but the UPS stopped.” That can happen if the problem is power, not energy. For UPS applications, battery selection must include both kWh and maximum DC current.
Continuous Current and Peak Current Are Different
A BMS specification may include:
- Continuous discharge current
- Peak discharge current
- Peak duration
A UPS can present both steady loads and short transients. Servers, motors, transformers and power supplies can have startup or dynamic current behavior. If the UPS load briefly drives battery current above the allowed peak or keeps it above the continuous limit too long, protection can occur. Do not select a battery based only on the highest number printed in the BMS specification.
UPS Load Percentage Can Be Misleading
A UPS may show: 50% load
but that percentage refers to the UPS’s own rated output. For a 20kVA UPS, 50% can still represent a very large load. Battery distributors should ask for actual watts or kilowatts rather than only “UPS load = 50%.” Power factor can also mean that kVA and kW are not identical. For runtime calculations, actual real load in kW is much more useful.
Example: 10kVA UPS at 40% Load
Suppose a 10kVA UPS reports 40% load. A customer may assume this means only 4kW, but actual real power depends on the load and UPS reporting method.
The safest approach is to read the UPS’s real output power where available. If the actual load is 3.5kW and conversion efficiency is 92%, battery power is already approximately: 3.5 ÷ 0.92 ≈ 3.8kW
A 5.12kWh battery will therefore have substantially less than 1.46 hours of practical runtime once usable SOC limits and system losses are included.
Rated kWh Is Not Always Fully Usable kWh
Battery nameplate energy is usually calculated from nominal voltage and rated Ah. The system may intentionally use less than 100% of that nominal energy to protect the battery or maintain reserve.
For example, the inverter may operate only between a defined upper and lower SOC window. If the customer has configured minimum SOC at 20%, then approximately 20% of nominal energy is intentionally reserved. A runtime calculation based on 100% of nameplate energy will therefore be optimistic.
Starting SOC Must Be Verified
One surprisingly common cause of short runtime is that the test did not actually begin with a fully charged battery. The UPS may display 100% using its own lead-acid-based estimation, while the lithium BMS shows 80%.
The customer then performs a runtime test expecting full battery energy. The result appears 20% shorter than expected. Before evaluating runtime, confirm SOC from the battery BMS and verify that the charging process reached the intended starting condition.
Actual Load Often Changes During the Test
Customers may calculate runtime using an assumed 2kW load, then unknowingly add other equipment. During an outage, cooling fans, networking devices, emergency lighting or other systems may start automatically.
The UPS battery therefore sees a larger load than the customer expects. A good runtime investigation should record actual UPS output power throughout the discharge rather than relying on an equipment list created months earlier.
Battery Temperature Changes Available Performance
LiFePO4 battery performance is affected by temperature. Low temperature can reduce usable capacity and increase internal resistance. A battery that passes a warehouse capacity test at 25°C may provide less usable energy at a cold remote site.
Higher internal resistance at low temperature can also increase voltage sag during high UPS load, causing earlier shutdown. Always compare test and site temperatures when investigating a runtime difference.
Cable Voltage Drop Is Part of the Runtime System
Imagine the battery itself remains at 48.5V during a heavy discharge, but the UPS sees only 46.5V because of cable and connector losses. The UPS reaches its shutdown threshold even though the battery has not.
From the customer’s perspective, backup time is too short. A capacity test performed directly on the battery does not include the site’s external cable losses. Measure voltage simultaneously at the battery terminals and UPS DC input during high load. The difference can reveal how much runtime is being lost in the installation.
Breakers and Fuse Holders Can Add Resistance
The DC path may contain a main breaker, individual battery fuses, busbars and disconnect switches. Every connection contributes some resistance. A poor-quality or damaged breaker can create significant voltage drop at high current while appearing normal at rest.
Thermal imaging or voltage-drop measurements under load can identify these problems. Do not focus only on the battery cells. The complete DC path determines UPS performance.
Parallel Batteries Must Share Current Properly
Suppose four batteries are expected to share 200A total discharge, approximately 50A each. If cable resistance causes one battery to carry 90A while another carries only 20A, the heavily loaded battery may reach its BMS limit early.
Once it disconnects, current through the remaining modules increases further. This can create a cascade of battery dropouts and unexpectedly short UPS runtime. A light-load capacity test of each individual battery may show that all four are healthy. The problem exists in the parallel installation.
Capacity in Ah Can Be Misleading at Different Voltages
For UPS distributors comparing lead-acid and LiFePO4, it is better to compare energy in Wh or kWh rather than Ah alone. For example, 100Ah at 12V and 100Ah at 51.2V represent completely different stored energy.
Even within a nominal 48V class, actual voltage and usable operating range matter. Using kWh makes it easier to compare battery energy with UPS load in kW. This is especially useful when explaining runtime to end customers.
A Better Runtime Formula
A simplified first estimate is: Runtime ≈ Usable Battery Energy × Conversion Efficiency ÷ AC Load
For example: Nominal battery energy = 10.24kWh
Usable fraction = 80%
UPS efficiency = 92%
AC load = 4kW
Approximate usable AC energy: 10.24 × 0.8 × 0.92 ≈ 7.54kWh
Estimated runtime: 7.54 ÷ 4 ≈ 1.89 hours
This remains an estimate because voltage cutoff, temperature, current and BMS behavior can change the result. Still, it is much more useful than dividing Ah by load current without considering system efficiency.
Why Laboratory Capacity Tests Should Record Conditions
When a battery supplier reports a capacity result, the test should ideally identify discharge current or C-rate, starting condition, endpoint and temperature.
A result of: 101Ah means much more when the test conditions are known.
UPS dealers should not compare a low-rate factory capacity test directly with a high-power UPS runtime test and assume the two should produce identical usable energy. The difference in operating conditions must be considered.
When Should You Suspect a Genuine Battery Capacity Problem?
A real battery problem becomes more likely when the battery also delivers low capacity under an appropriate controlled test, when one cell reaches its discharge endpoint much earlier than the others or when the battery’s measured usable energy has declined significantly compared with earlier tests under similar conditions.
By contrast, if controlled capacity is normal but runtime is poor only on one UPS installation, the complete system deserves investigation. This distinction can save both the dealer and manufacturer from unnecessary warranty replacement.
Dealer Troubleshooting Checklist
When an end customer says “This battery is rated 10kWh but my UPS runs for only one hour,” collect:
| Parameter | What to Check |
|---|---|
| Battery nominal kWh | Actual stored energy |
| Starting BMS SOC | Was it really full? |
| UPS AC load in kW | Actual energy demand |
| UPS efficiency | Conversion losses |
| Battery DC current | Current capability |
| BMS alarm | Protection event |
| Lowest cell voltage | Weak-cell limitation |
| Battery/UPS terminal voltage | Cable voltage drop |
| Temperature | Site vs test conditions |
| UPS low-DC cutoff | Usable SOC window |
This converts a vague runtime complaint into an engineering problem that can be solved.
For Battery Distributors: Quote Runtime Carefully
Do not tell a customer that a 10kWh battery “will run your 5kW UPS for two hours” without qualifications. That calculation assumes 100% usable energy and 100% conversion efficiency. A more professional quotation states that actual backup time depends on load, UPS efficiency, discharge limit, battery current capability and site conditions.
Where runtime is critical, obtain the exact UPS model and load before recommending battery capacity. This reduces after-sales disputes.
For UPS Dealers: Size the Battery for Both Runtime and Current
Traditional UPS battery sizing often focuses heavily on runtime. With LiFePO4, BMS current capability adds another clear requirement.
A battery bank must satisfy two questions: Does it contain enough energy? and Can it deliver the required power?
A 20kWh battery that can only safely deliver 5kW is not suitable for a 15kW UPS simply because its energy capacity is large. The electrical design must satisfy both dimensions.
Conclusion
A LiFePO4 battery can pass a controlled capacity test and still provide shorter-than-expected UPS runtime without being defective. The difference can come from UPS efficiency, high discharge current, voltage sag, BMS current limits, UPS cutoff voltage, cable losses, temperature, starting SOC and real load power.
Capacity testing answers how much energy the battery can deliver under defined conditions. UPS runtime measures the performance of the complete battery-plus-UPS-plus-installation system.
HIZN supplies LiFePO4 batteries for UPS, telecom and stationary backup applications and can provide battery-side energy, current and BMS specifications for system matching.
If an end customer’s UPS runtime is shorter than expected, send us the UPS model, actual load in kW, battery model/quantity, starting SOC and shutdown/BMS data. We can help determine whether the problem is battery capacity, power capability or system configuration.