Review, teardown, and testing of ERPF-400-24 Mean Well power supply

General Description

Brief Specification: ERPF-400-24 is a power supply unit with a 24-volt DC output and a rated current of up to 16.7 amps. According to the specification, the unit has an extended operating range for AC input voltage from 90 to 264 V. However, with an input voltage between 90 and 200 V, the unit can only deliver 50% of its rated power. 
It can also operate from a DC input within the range of 127 to 370 V.

The unit has dimensions of around 9x5x2 inches (220x130x48 mm exactly), is built on a printed circuit board housed in a stamped metal case, and is designed to operate with no forced cooling. The board is installed into the case, like in a tray, and is covered with a compound (most likely thermally conductive), making it non-removable for repair. The unit is covered with a perforated lid on top.

It features an LED indicator for output voltage and allows adjustment of the output voltage within a range of -5% to +10%. The unit includes an active PFC (Power Factor Correction) circuit and has a high power factor of up to 0.98 at 115 VAC. It also has thermal protection, with temperature monitoring performed using a thermistor placed on the output rectifier diodes.
Traditional safe measures, such as overload and overvoltage protection, are also provided.

Circuit and construction description: Unfortunately, the compound filling makes detailed examination of the circuit difficult, so there is a high probability of interpretation errors. However, it appears that the design shares many similarities with the RSP-320-24 (https://teardownit.com/posts/review-teardown-and-testing-of-rsp-320-24-mean-well-power-supply) power supply unit, in both circuit and layout. This suggests that the ERPF-400 may be controlled by the FAN4800 controller, just like the RSP-320.

The input and output terminals are mounted on a common screw terminal block (1). The terminal block's connections from top to bottom are as follows: three terminals for phase, neutral, and ground (input), and three parallel terminals for the output, ground, and +24V.
The input voltage from the terminal block (1) passes through an EMI filter (3) and then to the diode bridge (5). A varistor (2) is installed at the filter's input to suppress dangerous voltage spikes. The rectified voltage from the bridge (5) is then routed through an NTC inrush current limiter (4) to the active PFC circuit. The PFC's power stage consists of a transistor (6), inductor (11), diode (7), and output capacitor (8).
The rectified and filtered voltage from capacitor (8) is fed into a forward converter, which comprises transistors (9) and transformer (10). The voltage from the transformer's output winding (10) is rectified by diodes (12) and filtered by an output LC filter (13) (14).

The filter’s output capacitance is provided by four 560µF, 35V capacitors rated for operation at up to 220°F or 105°C (13).

The transformer's core (10) is tied in a metal clamp, which is bolted to the case. This design likely serves to dissipate heat from the transformer.

Due to the presence of the compound, it is difficult to assess the overall build quality. The compound itself has been applied somewhat carelessly, with visible splatters and drips. Furthermore, there are gaps in the compound application, significantly reducing its protective effectiveness.

The LED output voltage indicator and output voltage adjustment resistor (16) are located near the terminal block, allowing access without removing the top cover.

Test Conditions

Most tests were conducted using Test Circuit 1 (see appendix) at 80°F (27°C), 70% humidity, and 29.8 inHg pressure.
Unless otherwise specified, measurements were taken without pre-warming the power supply, and the operating mode was momentary load. Input voltage was set to 115V AC, and the current of 8.4 A was considered 100%.
The following values were used to determine load levels:

Output voltage with static load

The unit demonstrates excellent output voltage stability.

Startup Characteristics

Startup at 100% load

Before the test, the power supply was turned off for at least 5 minutes with the load connected at 100%.
The startup waveform at 100% load is shown below (Channel 1: output voltage, Channel 2: input current):

The startup process can be divided into three phases:
1. Input current pulses charging the input capacitors upon connection to the grid, with a peak amplitude of about 3 A, consisting of two portions of one period each.
2. Waiting for the control circuit to start, about 231 ms.
3. (Output Voltage Rise Time) Output voltage rise, 57 ms.
4. (Turn On Delay Time) The total time to reach operational mode from power-on is 288 ms.

(Output Voltage Overshoot) The startup process is aperiodic with no overshoot.

Startup at 0% load

Before the test, the power supply was turned off for at least 5 minutes with the load connected at 100%, then the load was disconnected, and the unit was turned on.
The startup waveform at 0% load is shown below:

The startup process consists of three phases:
1. Input capacitors charging upon connection to the grid, with a peak amplitude of about 2.6 A, consisting of two portions of one period each.
2. Waiting for the control circuit to start, about 241 ms.
3. (Output Voltage Rise Time) Converter startup, output voltage rise, and transition to operational mode, 56 ms.
(Turn On Delay Time) The total time to reach operational mode from power-on is 297 ms.

(Output Voltage Overshoot) The startup process is aperiodic with no overshoot.

Shutdown Characteristics

The shutdown process was tested at 100% load with nominal input voltage at the moment of shutdown. The shutdown waveform is shown below:

The shutdown process can be divided into two phases:
1. (Shut Down Hold Up Time) The unit continues operating, powered by the charge stored in the input capacitors, until their voltage drops to a critical level at which maintaining the output voltage is no longer possible. This phase lasts for 26 ms.
2. (Output Voltage Fall Time) Output voltage decline, converter stop, and acceleration of the voltage drop. This phase lasts for 25 ms.

(Output Voltage Undershoot) The shutdown process is aperiodic, with no undershoot.

Current amplitude at 100% load prior to shutdown was 2.8 A.

Output voltage ripple

At 100% load:

low-frequency ripple of about 10-12 mVp-p at twice the grid frequency, and around 4 kHz with a ripple of 13-15 mVp-p.

At the converter frequency, ripple is approximately 25 mVp-p, with noise at 120 mVp-p.

At 75% load:

low-frequency ripple of about 10 mVp-p at twice the grid frequency and around 12 mVp-p at approximately 4 kHz.

Converter frequency ripple is about 20 mVp-p, with noise at 100 mVp-p.

At 50% load:

low-frequency ripple of about 10 mVp-p at twice the grid frequency and around 15 mVp-p at approximately 4 kHz.

Converter frequency ripple is about 20 mVp-p, with noise at 100 mVp-p.

At 10% load:

low-frequency ripple of about 3-5 mVp-p.

Converter frequency ripple is about 20 mVp-p, with noise at 100 mVp-p.

At 0% load:

The input current was measured with a multimeter at 60 mA.
(Power Consumption) Input current in this mode is primarily reactive in nature; thus, the power consumption value measured with handy instruments can't be correct. The input filter has two capacitors, according to the diagram.

At 0% load, low-frequency ripple is hardly distinguishable from the noise, about 3 mVp-p.

Converter frequency ripple is masked by noise around 100 mVp-p.

Dynamic characteristics

Dynamic characteristics were assessed in a mode that switches between 50% and 100% load. The oscillogram below illustrates the process:

It is evident that the unit's response to step-load changes is aperiodic, with the magnitude of the response to load changes being approximately 100 mV p-p.

Overload protection

The manufacturer specifies overload protection with a constant current limiting type, which was confirmed during testing. When overloaded or the output terminals are shorted, the unit enters current limiting mode and automatically recovers after the fault is removed.

The output current at which the limit is triggered is 22 A.

Input safety assessment

(Input Discharge) The input circuit discharge time constant was measured upon disconnection from the grid, with a value of 0.245 s. This means that when operating on a 120V grid, the time required for the input circuits to discharge to safe levels (<42 V) is 0.39 s:

Important: This result applies only to the tested unit and was obtained exclusively for research purposes. It cannot, under any circumstances, be considered a guarantee of safety.

The ground leakage current was measured at 73µA.

Thermal behavior

No significant heating of components was observed during no-load operation.
Thermograms were taken at three power levels—80%, 90%, and 100%—with and without the top cover. The thermograms show that the most thermally loaded components of the unit are the NTC inrush current limiters (4), which stand out from the other components. At 80% load, the NTC temperature is 175.6°F (80°C, with a 53°C rise above ambient temperature); at 90% load, it reaches 192.1°F (89°C, with a 62°C rise); and at 100% load, it reaches 203.9°F (95.5°C, with a 68.5°C rise).

The second hottest component, after the NTC, is a resistor, 25 degrees cooler.

80% load

90% load

100% load

Conclusions

The ERPF-400-24 generally exhibits low noise and ripple, and it maintains good accuracy in sustaining the output voltage. The unit has decent dynamic characteristics, responding to pulsating loads without overshoot.

On the startup waveform (see oscillograms), there is an N-shaped section where, after the initial rise in output voltage, a partial drop occurs, followed by another rise and stabilization at the nominal level. This behavior could potentially cause issues when powering up digital devices that require initialization.

The build quality is good. However, the compound is applied somewhat poorly, with notable imperfections and several gaps. Unfortunately, these gaps affect components with increased hygroscopicity, particularly the wire-wound inductances of the output and input noise filters. Despite the manufacturer's claim of “protecting the internal electronic components from rain splash and dust,” this unit should not be used without additional water ingress protection.

According to the specifications, the unit is designed for operation under conditions of "Cooling by free air convection" and "-22F ~ +140°C (Refer to output load derating curve)." The tested unit indeed remains safe up to 100% load during continuous operation. The presence of a temperature sensor on the rectifier diodes adds extra assurance of safety.

Important: The results and conclusions presented apply only to the tested unit and were obtained solely for research purposes. Under no circumstances should they be used to assess all devices of this type.