Review, teardown, and testing of APV-35-24 Mean Well power supply

All information about this project is available in a special website section.
We thank you for your attention!
No items found.
Click image to zoom

General Description

Brief Specification: APV-35-24 is a power supply unit with an output voltage of 24 volts DC and a current up to 1.5 amperes. According to the specifications, the PSU has an extended operating range for input AC voltage from 90 to 264 V. When input voltage decreases from 100 V to 90 V, the supported load power decreases linearly from 100% to 80% of the nominal rating.

This power supply can also operate on a DC network within a range of 127 to 370 V.

The unit’s dimensions are approximately 3.3 x 2.2 x 1.2 inches (exactly 84 x 57 x 29.5 mm). It has a neat, tactile, light gray plastic casing, presumably made of ABS. The casing appears solid and unrepairable (or hides its secret).

The PSU lacks external indicators, output voltage, or current adjustments. It does not have power factor correction (PFC) and offers an efficiency of around 84%. While it has overload and output overvoltage protection, it lacks thermal protection.

Test Conditions

Most tests were conducted using Test Circuit 1 (see appendix) at 80°F (27°C), 70% humidity, and 29.8 inHg atmospheric pressure.

Unless specified otherwise, measurements were taken without pre-warming the power supply in a short-term load mode. Input voltage was set at 115V AC, and 1.5 A was considered 100% load.

To determine load level, the following values were used:

Static Load Output Voltage

The change in output voltage with load changes did not exceed 0.7%.

Startup Characteristics

Power-On at 100% Load

The unit was kept off for at least 5 minutes with a 100% load connected before the test.

The oscillogram below illustrates the startup process at 100% load (Channel 1: output voltage, Channel 2: input current):

The startup waveform has three phases:

  1. Input current spike to charge capacitors upon power connection, peaking at 7 A, lasting less than half a cycle of the input voltage (5 ms).
  2. Waiting for the control circuit to initialize, approximately 17 ms.
  3. (Output Voltage Rise Time) Voltage increase, taking 7 ms.

(Turn-On Delay Time) The entire startup process takes 24 ms.

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

Power-On at 0% Load

Before testing, the unit was kept off for at least 5 minutes with 100% load. Then, the load was disconnected, and power was turned on.

The oscillogram below illustrates the startup at 0% load:

The startup waveform has three phases:

  1. Input current spike to charge capacitors, peaking at 6 A, lasting less than half a cycle (5 ms).
  2. Waiting for control circuit initialization, around 23 ms.
  3. (Output Voltage Rise Time) Converter startup and voltage increase, lasting 6 ms.

(Turn-On Delay Time) The entire startup process takes 29 ms.

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

Power-Off Characteristics

Shutdown was conducted at 100% load with nominal input voltage. The oscillogram below illustrates the shutdown process:

The power-off process has two phases:

  1. (Shut Down Hold Up Time) The source continues running on capacitor charge until voltage drops below a critical level, after which nominal output can no longer be maintained. This phase lasts 78 ms.
  2. (Output Voltage Fall Time) Output voltage drops, the converter halts, and the voltage drop accelerates. This phase lasts 27 ms.

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

Peak current at 100% load before shutdown: 1 A.

Output Voltage Ripple

100% Load

A low-frequency ripple with an amplitude of approximately 60 mV is observed. The same diagram shows the waveform of the input current (Channel 2) with an amplitude of 1 A.

Both switching frequency (~62 kHz) ripple and noise at 100% load have an amplitude of about 350 mVp-p.

75% Load

Low-frequency ripple at approximately 60 mV with a sin-like waveform; amplitude of input current is 0.75 A. 

Switching frequency ripple at 75% load is around 350 mVp-p, switching noise of 450 mVp-p.

50% Load

Low-frequency ripple reduces and appears as noise, masking harmonics at grid frequencies, around 50 mVp-p. Input current peaks at 0.5 A.

Switching ripple at 50% load is 300 mVp-p, and switching noise reaches up to 700-750 mVp-p.

10% Load

Low-frequency ripple appears as noise with an amplitude of 110 mVp-p, masking grid harmonics. Input current peaks at 0.3 A.

Switching ripple is around 200 mVp-p, with noise near 800 mVp-p.

Notable is the change in the converter's operation—specifically, the interval between energy transfer cycles to the load increases from 16 microseconds to approximately 39-41 microseconds. It's unclear what causes this; one might initially assume the converter simply lowers its frequency as the load decreases. However, when running with no load, this interval returns to 16 microseconds. On the other hand, it doesn’t seem to be subharmonic oscillation due to insufficient slope compensation either, as the periods aren’t multiples of each other. It would be helpful to look "under the hood," but the case is too pretty to break open.

0% Load

Current consumption without load is 19 mA, according to the multimeter. 

(Power Consumption) Input current in this mode is primarily reactive in nature; thus, the power consumption value could be challenging to measure with simple instruments.

Low-frequency ripple at 0% load is approximately 50 mVp-p. The input current waveform resembles a sinusoid with some noise.

Switching frequency ripple with no load is masked by noise of 200 mVp-p.

Dynamic Characteristics

Dynamic testing involved periodic switching between 50% and 100% load. The oscillogram below illustrates the process:

The oscillogram shows a smooth response to load changes without overshoot, though the load change response reaches 2 V p-p, which is relatively high. Channel 2 shows input current changes in relation to load changes.

Overload Protection

Manufacturer-specified Hiccup Mode protection was verified during testing. In the event of an overload or a short circuit across the output terminals, the power supply switches to a pulsed (hiccup) mode and automatically resumes normal operation once the overload condition is removed.

Overload protection triggers at 2.1 A output current.

Input Circuit Safety Assessment

(Input Discharge) Time constant for input discharge was measured at 0.199 s, resulting in a discharge time of 0.32 s to safe voltage levels (<42 V) when used with 120 V AC.

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.

Thermal Characteristics

No noticable heating of components was observed during no-load operation.

Thermograms were taken at 80%, 90%, and 100% load. At 80% load, the hotspot reached 131.3°F (55.2°C, +28° above ambient). At 90%, it reached 142.5°F (61.4°C, +34°). At 100% load, the maximum temperature was 151.7°F (66.5°C, +40°).

80% load

90% load

100% load

Conclusions

The APV-35-24 is a simple power supply unit with considerable noise and moderate dynamic response and should be used accordingly. It does not reach unsafe temperatures but lacks thermal protection, so it should not be left unattended.

Unfortunately, internal inspection is not possible without destroying the casing.

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.

Products in this post

Sign up for my newsletter

Kevin Gibbs

Hi! I'm Kevin! I am a very curious engineer :))
I'm the website founder and author of many posts.

I invite you to follow exciting experiments, research, and challenges.
Let's go on to new knowledge and adventures!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.