24 V 6 A power supply from Geekcreit
Inspection, tests, conclusions—as usual.
It’s worth saying right away that this power supply looks a lot like the well-known people’s model; it just has a different PCB shape and, accordingly, different trace routing.
It comes in an antistatic bag.

No complaints about the appearance. It’s no Mean Well, of course, but it’s fairly neat. The board is square, 83 × 83 mm, and 29–30 mm tall. The transformer is the tallest component.

Judging by the thousandths of a millimeter, the dimensions were clearly converted from inches.

Unlike the “people’s” series, this one has no terminal blocks. You can install them, but note that the input terminals need 7 mm spacing between contacts, not the more familiar 5 mm.

The input has a mains filter consisting of a common-mode choke and an X capacitor. There’s also a fuse and a thermistor.
Along with the L and N contacts, there’s also an earth connection, and you need to be careful here. The E (Earth) contact is connected to the output common, and its trace also runs close to the positive terminal of the high-voltage capacitor. So, for improved safety, it’s advisable either to ground it or remove this trace.

There’s also an output choke to reduce ripple. In addition, a 1 kΩ resistor is connected across the output, dissipating about 0.5 W. As with the previous model, I recommend replacing it with a resistor around 1.5–2 kΩ to reduce heating.

The topology is the same as in the “people’s” model: single-ended flyback. What’s more, it even uses the same PWM controller as the old version, the CR6842S, so there’s no point in drawing the schematic. The differences between these power supplies are minimal and mostly concern component numbering.
The interwinding capacitor here is Y-rated, unlike the ordinary capacitor used in the older power supplies.

A little more about the components.
1. Input capacitor, rated at 82 µF
2. Pair of output capacitors, 1000 µF, 35 V.
3. High-voltage transistor, 10N60C
4. Output diode assembly, MBR20200CT
Note that neither the transistor nor the diode assembly has thermal paste between it and the heatsink. The diode assembly is also mounted with a mica insulator, so if you buy this power supply, I recommend applying thermal paste.

The capacitors are ordinary no-name parts. Their actual capacitance is slightly below the stated value, but not critically so. If you want to improve reliability and performance, at least replace the output capacitors with better-quality ones.

The board is double-sided, but the components are mounted on one side only, so there’s nothing on the bottom.

Let’s move on to the tests.
With 230 V applied to the input, we get the rated 24 V at the output. It’s actually a little higher, but not critically so. After a short warm-up, the voltage rises a little more, to 24.04 V.
No-load power consumption is 1.3–1.4 W, almost half of which is dissipated by the 1 kΩ resistor connected across the output. Startup has a noticeable delay of about 1–1.5 s.

The power supply easily delivers 6.25 A. I simply don’t have a load capable of going any higher and producing a graph.

The listing claimed a maximum output current of 7.5 A. The protection trips at around 7.7–7.8 A and resets automatically afterward.

The power supply’s efficiency is about 88–89% over the 1.5–7.5 A load range. Below is a graph covering 0.5–7.5 A in 0.5 A increments.

Ripple amplitude was measured with no load and at currents of 2.5, 5, and 7.5 A. A filter made of two capacitors, 1 + 0.1 µF, was connected ahead of the oscilloscope probe.
The screenshots show that with no load, the controller switches to “green” mode, with the frequency significantly reduced. As the load current increases, the power supply operates normally. The main ripple is small, at about 20 mV peak-to-peak, but there are noticeable spikes reaching 100 mV at 5 A and 200 mV at 7.5 A.

With the probe connected directly, the spike amplitude is about 2–3 times greater. Below is a screenshot of the ripple with a direct connection at currents of 5 and 7.5 A.

Ripple at a low sweep rate. I expected worse, but I would still replace the input capacitor with a decent one rated at 100 µF.

The component thermal test was carried out in 20-minute stages at load currents of 2, 4, and 5 A. It turned out that the maximum continuous current without forced cooling, with the unit open, is 5 A. The hottest component is the transformer, which reached 104 degrees. I could smell the varnish impregnating the winding heating up.
The unit works normally for short periods at currents up to 7.5 A.

Thermal images at load currents of 2, 4, and 5 A.

While I was at it, I checked how the output voltage changes as the unit warms up.
1. With no load, the output is about 24.03 V
2. At 2 A, the output voltage drops slightly
3. After an hour of warm-up, at a load current of 5 A, the output voltage is 24.045 V
4. With the power supply hot and unloaded, the voltage rises to 24.07 V; after it cools down, it drops to 24.04 V
In my opinion, the regulation here is excellent. The voltage has a slight positive temperature coefficient and decreases as the load increases; overall, the voltage variation is no more than 40 mV.

To try to reduce the ripple amplitude, I added four 0.15 µF ceramic capacitors: two in parallel with the capacitor before the inductor, one in parallel with the capacitor after the inductor, and one across the output.
You can see that the ripple decreased, but only slightly. The probe was connected directly to the output. With a filter in front of the probe, there was almost no difference, since in this setup the filter itself suppresses the spikes better than the ceramic capacitors I added.
If you want even less ripple, it’s worth replacing the electrolytic capacitors, as I did in this review.
Results at currents of 5 and 7.5 A: before on the left, after installing four 0.15 µF capacitors on the right.

And here’s a comparison of the two power supplies.

To sum up, practically everything I wrote in my reviews of the “people’s” power supplies, both the 12 V and 24 V versions, applies to this unit too, since they are essentially the same thing.
One difference is that the transformer in the power supply reviewed here runs hotter, but the transistor runs cooler even without thermal paste.
Also, the current-sense shunt here has a resistance of 0.22 Ω, compared with 0.33 Ω in the “people’s” models. This means it can deliver more current for short periods, which is just right for use in a soldering station.
In short, it’s a decent alternative to the power supplies mentioned above if, for some reason, you need a square circuit board rather than a rectangular one.
That’s all from me. I hope you found this useful.
Questions and answers
What current can the Geekcreit 24 V power supply handle?
It easily delivered 6.25 A in the test, while protection tripped at about 7.7–7.8 A. Without forced cooling, the continuous current is limited to 5 A.
How much ripple does the Geekcreit 24 V power supply have?
The main ripple was about 20 mV peak-to-peak. Spikes reached 100 mV at 5 A and 200 mV at 7.5 A.
How hot does the Geekcreit power supply get?
During continuous operation without forced cooling, the transformer reached 104 °C. The author suggests replacing the 1 kΩ output resistor with a 1.5–2 kΩ resistor to reduce heat.
This model is included in the summary table: Comparison of power supplies: ripple, efficiency, and temperature rise
The figures from this review are in the measurement database: Power supplies
Discounted products
You may be interested in
Discounted products
All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!
Only registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.
Log inRegister
