Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 61204: Low-Voltage DC Power Supplies — Performance Characteristics and Test Methods
✅ Standard at a Glance
IEC 61204 is the international standard governing the performance characteristics, test methods, and qualification requirements for low-voltage DC power supplies. Developed by IEC Technical Committee 22 (Power electronic systems and equipment), this standard applies to AC-DC rectifiers and DC-DC converters with DC output voltages up to 1500 V DC, including power supplies for industrial, commercial, telecommunications, and railway applications. It is the key reference for specifying, testing, and comparing DC power supply performance across different topologies and manufacturers.
IEC 61204 is the international standard governing the performance characteristics, test methods, and qualification requirements for low-voltage DC power supplies. Developed by IEC Technical Committee 22 (Power electronic systems and equipment), this standard applies to AC-DC rectifiers and DC-DC converters with DC output voltages up to 1500 V DC, including power supplies for industrial, commercial, telecommunications, and railway applications. It is the key reference for specifying, testing, and comparing DC power supply performance across different topologies and manufacturers.
🔌 1. Scope and Performance Parameter Framework
1.1 Applicability and Key Definitions
IEC 61204 covers power supplies that convert AC input (single-phase or polyphase) to DC output, or DC input to DC output, with rated DC output voltages up to 1500 V. The standard does not cover specialist power supplies already covered by other IEC standards (e.g., IEC 60146 for semiconductor converters, IEC 62040 for Uninterruptible Power Supplies, or IEC 61558 for safety isolating transformers).
The standard defines several fundamental performance parameters that collectively characterize the “quality” of a DC power supply:
| Parameter | Definition | Typical Specification | Measurement Method |
|---|---|---|---|
| Voltage accuracy (set-point accuracy) | Deviation of the actual output voltage from the nominal set value under specified conditions | ±0.5% to ±2% of rated voltage | Digital voltmeter at output terminals; average over 1 minute |
| Load regulation (steady-state) | Change in output voltage when the load current changes from 0% to 100% of rated value | ≤ 1% of rated voltage | Measure Vout at no-load and full-load; calculate ΔV/Vnom |
| Line regulation (steady-state) | Change in output voltage when the input voltage varies over the specified range | ≤ 0.5% of rated voltage | Vary input from Vmin to Vmax; record ΔVout |
| Ripple and noise (peak-to-peak) | AC component superimposed on the DC output, including switching ripple and high-frequency noise | ≤ 1% of Vnom (p-p) for general-purpose supplies | Oscilloscope measurement with 20 MHz bandwidth; specified load impedance |
| Transient recovery time | Time for output voltage to return to within specified tolerance band after a step load change | ≤ 1 ms for typical SMPS; ≤ 100 µs for high-performance | Step load from 50% to 100%; measure settling time within ±1% band |
| Efficiency | Ratio of output power to input power at specified load conditions | ≥ 85% at full load (typical); ≥ 96% (high-efficiency) | Input and output power measurement using calibrated wattmeters |
| Hold-up time | Duration the output remains within regulation after loss of AC input power | ≥ 10 ms at full load (typical); ≥ 20 ms (industrial) | Measure from AC loss to Vout dropping below regulation band |
| Start-up time | Time from application of input power to output reaching regulation band | ≤ 2 s (typical) | Measure from input application to Vout reaching 90% of nominal |
💡 Engineering Insight
Transient response is often the most undervalued parameter in DC power supply specifications. A power supply may have excellent steady-state load regulation (e.g., 0.1%) but exhibit a 5% voltage dip during a 50-100% load step, with a 2 ms recovery time. This transient dip can cause downstream equipment (especially digital logic, FPGAs, or sensitive analog circuits) to malfunction or reset. IEC 61204 specifies a standardised transient test: 50% to 100% to 50% load step with a current slew rate of 1 A/µs. When selecting a power supply for a load with high di/dt (e.g., motor drives, pulsed loads), always verify the transient response specification — not just the steady-state regulation.
Transient response is often the most undervalued parameter in DC power supply specifications. A power supply may have excellent steady-state load regulation (e.g., 0.1%) but exhibit a 5% voltage dip during a 50-100% load step, with a 2 ms recovery time. This transient dip can cause downstream equipment (especially digital logic, FPGAs, or sensitive analog circuits) to malfunction or reset. IEC 61204 specifies a standardised transient test: 50% to 100% to 50% load step with a current slew rate of 1 A/µs. When selecting a power supply for a load with high di/dt (e.g., motor drives, pulsed loads), always verify the transient response specification — not just the steady-state regulation.
1.2 Classification of Power Supply Types
IEC 61204 classifies DC power supplies according to their topology and output characteristics:
| Type | Description | Typical Efficiency | Typical Ripple | Application |
|---|---|---|---|---|
| Linear regulated | Series pass transistor dissipates excess voltage; low noise but inefficient | 30-55% | < 1 mV p-p | Lab supplies, audio equipment, precision analog |
| Switched-mode (SMPS) — single transistor forward/flyback | Single switch; simple design; limited power range | 70-82% | 10-50 mV p-p | ≤ 250 W, consumer, auxiliary supplies |
| SMPS — half-bridge/full-bridge | Two or four switches; higher power; better transformer utilization | 82-90% | 10-50 mV p-p | 250 W – 5 kW, industrial, telecom |
| SMPS — resonant (LLC / phase-shifted full-bridge) | Soft-switching topology; reduced switching losses; lower EMI | 90-96% | 20-80 mV p-p | 5 kW – 50 kW, data center, EV charging |
| DC-DC converter — isolated | Galvanic isolation via transformer; input-to-output isolation ≥ 1500 V DC | 78-92% | 10-100 mV p-p | Telecom, board-level POL (point-of-load) |
| DC-DC converter — non-isolated (buck/boost/buck-boost) | No galvanic isolation; high efficiency; compact | 90-97% | 5-50 mV p-p | Intermediate bus, POL regulators |
💡 2. Detailed Test Methods and Performance Characterization
2.1 Measurement Setup and Precautions
IEC 61204 places great emphasis on the test setup because DC power supply measurements are highly sensitive to lead impedance, grounding, and probe placement. Key requirements:
- Output voltage is measured directly at the output terminals of the power supply, not at the load. For remote sensing configurations, the sense lines must be connected at the load and the measurement point specified accordingly.
- Ripple and noise measurements must use a 20 MHz bandwidth-limited oscilloscope with a 1:1 probe (not 10:1, as the 10:1 probe attenuates the signal and reduces the signal-to-noise ratio of the measurement). A 10 µF tantalum capacitor in parallel with a 0.1 µF ceramic capacitor is connected across the output terminals to simulate the typical decoupling network of the load.
- The standard specifies that the oscilloscope’s ground lead must be as short as possible — the standard ground clip (3-4 inches) forms a loop antenna that picks up radiated EMI, adding 10-20 mV of phantom noise. A spring-loaded ground tip adapter is recommended.
- Efficiency measurements require a calibrated power analyzer capable of measuring true RMS power. For AC input supplies, the measurement must account for power factor and crest factor errors. For three-phase supplies, all three phases must be measured simultaneously.
🚨 Measurement Pitfall: Ripple and Noise Using the Wrong Bandwidth
A common measurement error is using the full oscilloscope bandwidth (e.g., 500 MHz) for ripple and noise measurements. At full bandwidth, high-frequency switching noise components (typically 100-500 MHz from fast-switching GaN FETs) are included in the measurement, yielding a peak-to-peak value 2-5 times higher than the true specified value. IEC 61204 mandates a 20 MHz bandwidth limit on the oscilloscope for ripple measurements, consistent with industrial noise susceptibility standards. The 20 MHz bandwidth captures all significant switching ripple (typically 20 kHz to 2 MHz) and broadband noise components that affect most loads, while excluding VHF components that are typically radiated rather than conducted. If the load is sensitive to higher-frequency noise (e.g., RF circuits), the specification should separately state the wideband noise requirement.
A common measurement error is using the full oscilloscope bandwidth (e.g., 500 MHz) for ripple and noise measurements. At full bandwidth, high-frequency switching noise components (typically 100-500 MHz from fast-switching GaN FETs) are included in the measurement, yielding a peak-to-peak value 2-5 times higher than the true specified value. IEC 61204 mandates a 20 MHz bandwidth limit on the oscilloscope for ripple measurements, consistent with industrial noise susceptibility standards. The 20 MHz bandwidth captures all significant switching ripple (typically 20 kHz to 2 MHz) and broadband noise components that affect most loads, while excluding VHF components that are typically radiated rather than conducted. If the load is sensitive to higher-frequency noise (e.g., RF circuits), the specification should separately state the wideband noise requirement.
2.2 Dynamic Performance Testing
IEC 61204 specifies several dynamic tests that are particularly relevant for power supplies feeding pulsed or varying loads:
Step load transient response: The output voltage response to a step change in load from 50% to 100% of rated current (slew rate: 1 A/µs) is recorded. The standard defines the recovery time as the interval from the step initiation until the voltage returns to and remains within the specified regulation band. For most industrial power supplies, the band is ±2% of the nominal voltage.
Start-up overshoot: The output voltage must not exceed 110% of the nominal voltage during start-up under any load condition. This is critical for loads that are sensitive to overvoltage (e.g., ASICs, FPGAs, memory modules). IEC 61204 requires that the start-up overshoot be measured at minimum load (typically 10% of rated current) as this condition produces the highest overshoot.
Turn-on and turn-off sequencing: For multi-output power supplies, the turn-on and turn-off sequence of each output relative to the others must be documented. Some loads (e.g., dual-voltage FPGA cores) require a specific power-up sequence (core voltage before I/O voltage or vice versa) to avoid latch-up or damage.
💡 Design Guidance for Power Supply Selection
When selecting a DC power supply for a specific load, use this four-step approach: (1) Identify the steady-state voltage and current requirements, including any derating for temperature. (2) Characterize the load’s dynamic profile: maximum di/dt, pulse frequency, minimum and maximum load currents. (3) Determine the permissible voltage deviation at the load: apply the “3% rule” — the power supply’s voltage accuracy, regulation, ripple, and transient deviation should sum to no more than 3% of the nominal voltage. (4) Verify that hold-up time meets the load’s ride-through requirement: if the load has an internal energy storage (bulk capacitance) that can sustain operation for 10 ms, a power supply with 12 ms hold-up provides 2 ms of margin.
When selecting a DC power supply for a specific load, use this four-step approach: (1) Identify the steady-state voltage and current requirements, including any derating for temperature. (2) Characterize the load’s dynamic profile: maximum di/dt, pulse frequency, minimum and maximum load currents. (3) Determine the permissible voltage deviation at the load: apply the “3% rule” — the power supply’s voltage accuracy, regulation, ripple, and transient deviation should sum to no more than 3% of the nominal voltage. (4) Verify that hold-up time meets the load’s ride-through requirement: if the load has an internal energy storage (bulk capacitance) that can sustain operation for 10 ms, a power supply with 12 ms hold-up provides 2 ms of margin.
2.3 Overload and Protection Features
IEC 61204 specifies testing requirements for built-in protection features:
- Overcurrent / short-circuit protection: The power supply must survive a sustained short circuit at the output for a minimum of 1 hour without damage. Two modes are permissible: constant current limiting (the output current is clamped at Imax) or hiccup-mode (auto-restart cycling). For hiccup mode, the auto-restart interval must not exceed 10 seconds.
- Overvoltage protection (OVP): If the output voltage exceeds the OVP threshold (typically 110-130% of nominal Vout), the power supply must shut down within 100 µs and latch off (requiring a manual power cycle or remote reset to restart).
- Overtemperature protection: An internal temperature sensor must shut down the supply when the internal temperature exceeds the rated maximum (typically 85-105 °C for the baseplate/heatsink), with automatic restart after cooling to within the hysteresis band (typically 10-20 °C below the shutdown threshold).
❓ Frequently Asked Questions
Q1: What is the difference between “load regulation” and “line regulation” and why does it matter?
A: Load regulation measures how much the output voltage changes when the load current varies (e.g., from no-load to full-load). Line regulation measures how much the output voltage changes when the AC input voltage varies (e.g., from 90 V to 264 V AC). Both matter, but for different reasons. Poor load regulation affects systems where the load current changes significantly during operation — a motor starting, a transmitter keying up, or a processor entering sleep mode. Poor line regulation affects systems connected to unstable AC mains — generators, long rural lines, or sites with large adjacent loads switching on/off. IEC 61204 recommends specifying both parameters separately, as they measure fundamentally different aspects of the power supply’s control loop performance.
Q2: Does IEC 61204 cover the EMC requirements for DC power supplies?
A: Partially. IEC 61204 references the IEC 61000-6-x series and IEC 61204-3 (EMC requirements for power supplies) for conducted and radiated emission limits, as well as immunity to electrostatic discharge, radiated RF fields, electrical fast transients, and surges. The standard specifies that the power supply must meet the applicable EMC limits when tested with a representative load (typically 80% of rated resistive load). For CE marking in the European Union, compliance with IEC 61204-3 is typically required in conjunction with the relevant product family EMC standard (EN 55011, EN 55032, or EN 61204-3). EMC testing must be performed on the power supply in its intended configuration, including all cables, filters, and enclosure as supplied.
Q3: How should I interpret the “ripple and noise” specification for a switching power supply?
A: The IEC 61204 ripple and noise specification is a peak-to-peak measurement with a 20 MHz bandwidth limit, measured at the output terminals with a specified load and input voltage. The measurement includes the fundamental switching ripple (e.g., 100-200 mV p-p at 100 kHz for a typical SMPS), switching frequency harmonics, and random noise. However, it is important to note that the actual ripple at the load will differ from the measurement at the power supply terminals due to the impedance of the connecting cables and the PCB traces. Long cables act as antennas and as transmission lines, potentially increasing the ripple at the load by 50-100%. To minimize this, IEC 61204 recommends placing decoupling capacitors (10 µF electrolytic + 0.1 µF ceramic) close to the load and using twisted-pair or coaxial connections for sensitive loads.
Q4: Can two or more DC power supplies be connected in parallel or in series to achieve higher current or voltage?
A: Yes, but with specific requirements. For parallel operation (higher current), IEC 61204 requires that the power supplies have active current sharing capability with ≤ 5% imbalance between units. The current sharing bus must be connected with a shielded twisted pair to avoid noise coupling. For series operation (higher voltage), each power supply must be rated for the full series voltage to ground (the output of the top unit floats at the series voltage with respect to ground). Each unit must have its own OVP set to no more than 110% of its individual nominal voltage. IEC 61204 recommends that series operation be limited to no more than 4 units unless specifically designed for higher series count. The standard does not recommend connecting power supplies with different output voltages or control topologies (e.g., voltage-mode and current-mode) in parallel without a dedicated interface module.
