The core of high frequency online UPS dynamic response test (mainly refers to load step response and mode switching response) is to verify the fluctuation amplitude and recovery speed of output voltage and frequency when load suddenly changes or working mode switches. The test should strictly follow GB/T 7260.3 and IEC 62040-3 standards, and the precautions are mainly divided into three categories: pre-test preparation, process control, safety and evaluation.
First, preparation before the test (to ensure that the data is valid)
UPS preheating and steady state
Before testing, UPS should be preheated for at least 30~60 minutes under full load condition to reach thermal stability.
Confirm that the input voltage and frequency are within the nominal value (such as 220V/50Hz), and the steady-state accuracy of the output voltage and frequency is up to standard (usually within 1%).
Instrument accuracy and calibration
Must use 0.5 and above power analyzer, high bandwidth oscilloscope (≥100MHz), high sampling rate data collector.
During the effective calibration period of the instrument, the voltage/current probe bandwidth covers the UPS switching frequency (usually 20kHz).
Load and wiring specification
Using programmable AC electronic load simulation, typical steps such as 0% 100%, 25% 75% and 50% 100% are preferentially covered.
The impedance of the test cable should be low and the length should be as short as possible to avoid the influence of line voltage drop on transient measurement.
Distinguish between linear load and nonlinear load (rectifier+capacitor) and test them separately.
environmental conditions
The ambient temperature is 25 5℃, the humidity is 30%~70%RH, and there is no strong electromagnetic interference.
Second, the test process control (core points)
1. Load step test (mains mode)
Step amplitude and rate
The step change time should be less than 1ms (simulating "sudden change").
Common combinations: 0%→100%, 100%→0%, 25%→75%, 50%→100% and vice versa.
Waveform capture and criterion
Voltage transient amplitude: generally, 8% ~+10% is allowed (the standards of different grades are slightly different).
Recovery time: return to the allowable zone of steady state within 20ms (such as 2% or 3%).
Oscillation times: ringing < 2~3 times, without continuous oscillation.
Oscilloscope triggers synchronously, and records the complete waveform of 5 cycles before step+20~30 cycles after step.
Key indicators:
Repetition and consistency
Repeat the same conditions for 3~5 times, take the average value and the worst value, and eliminate accidental errors.
2. Mode switching response (mains battery, inverter bypass)
Switch without breakpoint
On-line double conversion UPS requires zero interruption (0ms breakpoint).
If there is a breakpoint, > 20ms is unqualified.
Switching transient
When switching, the voltage overshoot/drop is ≤ 5% ~ 10%, and it will recover within one cycle (20ms).
Frequency transient: the frequency jump is < 3%~5% and it is fast and stable when the mains tracks the local oscillation of the battery.
3. Special attention to high frequency machine
Control loop and current limiting
High frequency machine (without output transformer) has fast dynamic response and easy overcurrent.
During the test, observe whether the inverter current limit is normal, whether there is any protection misoperation, and whether there is overvoltage in the output capacitor.
Harmonics and EMI
THD will increase when the load suddenly changes, and it is necessary to confirm that THD is ≤ 5% (nonlinear load) after recovery.
Transient process avoids electromagnetic interference from affecting control panel and measurement.
III. Security and boundary conditions
Overload and protection
The step does not exceed the rated capacity of UPS; 125%~150% short-term overload should be able to maintain or turn to bypass normally without blowing up the machine.
Battery mode test
The battery should be fully charged and in good condition; During dynamic discharge, the battery voltage drops normally, and there is no false alarm of low voltage.
Parallel system
When paralleling, do single machine dynamics first, and then paralleling test; Make sure that the load is evenly distributed, the circulation is small and the switching is consistent.
Fourth, common problems and pit avoidance
The bandwidth of the instrument is not enough → the real peak and oscillation cannot be measured, and the result is optimistic.
Insufficient preheating → the dynamic difference between low temperature and high temperature is great.
The response is worse when only linear load → actually nonlinear load such as server/PC is used.
Single test → accidental fluctuation leads to misjudgment.
Cable is too long/thin → line voltage drop masks the real dynamic performance of UPS.
Simple summary: preheat first, use the instrument well, short thick line, multi-step, grasp the waveform, see the amplitude/recovery time, repeat verification, distinguish linearity/nonlinearity, and ensure safety.
