24 V 3 A Power Supply

Many of my readers know how much I enjoy writing reviews of power supplies; it’s probably my favorite topic. But today’s review will be a little unusual for me: it’ll be short, and there are good reasons for that.
Anyway, if you’re interested in switching power supplies, come on in.


I ordered this power supply quite a while ago and had even forgotten about it, assuming it was simply out of stock. As it turned out, that was pretty much the case: the product page said “Sold out.” Still, the store found one unit somewhere in a warehouse (presumably the last one) and sent it to me.
This is a used power supply, meaning it was previously used somewhere and was most likely installed in an enclosure. It’s now being sold as a bare circuit board.

I was a little surprised when the package arrived unexpectedly and I opened it, thinking it was something else. Inside was the power supply, packed in the usual Banggood style: wrapped in white foam and put in a bag.


I ordered it because of its low price and pretty decent specs, which make it suitable for use in soldering stations. They need roughly these specs, perhaps a little more current, but I was counting on the headroom usually built into branded power supplies. I thought it might be useful to someone.
Yes, I mean branded power supplies specifically, since units like this are often branded models salvaged from equipment.

The only specs provided were 24 V and 3 A, along with the board dimensions: 11.5 x 5 x 2.5 cm.

As you can see, the power supply is almost completely covered by a copper-foil shield, so there’s no way to see what’s inside from the store page.


Only the standard input connector and a few output capacitors are visible; what’s inside is a mystery.


There’s shielding on the underside too, which is a good thing in its own way: less interference getting out.


At first I thought I’d unfold the shield, as is sometimes done with laptop power supplies, but it turned out to be much simpler here. I desoldered the copper tabs in a couple of places—though one broke off instead of coming loose—and simply slid the shield aside.


Now I can finally get a proper look. My first impression was: not bad, very not bad.


That “not bad” applies especially to the input filter. This is textbook material for how to do it right.
The mains filter has all the components a proper filter should have:
1. Fuse
2. Thermistor
3. 2 X capacitors
4. 2 Y capacitors
5. 2 dual-winding chokes
6. Varistor
7. Bleeder resistors for the X capacitors

Just excellent. There’s nothing to add.

As I often say, you can judge how seriously a manufacturer takes its work by the input filter, since that’s the first thing they “cut” when trying to save money. Here, nothing was skimped on.


Of course, you could complain about the lack of a power factor correction circuit, but it’s not really essential here, nor is it needed. Besides, there wouldn’t have been room for one.

The input capacitor is rated at 120 µF, which is more than enough for the stated 72 W when powered from 220–230 V, and should in principle be sufficient even for a 100–115 V input. The capacitor’s actual capacitance was 105 µF; ESR was 0.73 Ω at 100 Hz and 0.5 Ω at 1 kHz.
The PWM controller’s supply capacitor is rated at 47 µF, 50 V. Its actual capacitance was 45 µF, with an ESR of 2.27 Ω at 100 Hz and 0.92 Ω at 1 kHz.

The measurements show that this is a used power supply, but everything is within normal limits. The measurements were taken without desoldering the components, so there may be a small margin of error.

The high-voltage transistor is mounted on a heatsink. The heatsink is also soldered to the board via two pins, and since it extends partly over the secondary side of the power supply, it’s insulated with tape.


A Y capacitor for suppressing interference is installed between the primary and secondary sides (to the left of the transformer). Feedback is routed through an optocoupler. All very familiar and conventional.
The transformer measures 27.5 x 20 x 20 mm externally, which is pretty good for the stated 70 W. There are also pieces of thermally conductive rubber on and next to the transformer. They were positioned between the transformer and the copper shield, transferring some of the heat to the copper. Again, a sensible solution, since overheating the core can have unpleasant consequences.

There’s one rectifier diode assembly on the output, but there’s room for a second one. Most likely, the board is universal, and the product line includes power supplies with lower voltage but higher current. The heatsink here is secured at one point in addition to the rectifier assembly’s pins, and is also insulated with tape.

After the rectifier assembly there’s a pair of 470 µF, 35 V capacitors of different types (you can tell by their color), as well as an inductor to reduce ripple and another capacitor with a capacitance of about 100 µF. Some components are covered with sealant. I didn't remove it because I saw no point, but even if I got a component value wrong, it can only be off by a little. The output filter has a total capacitance of about 1000 µF, and the capacitors' voltage ratings are fine too: 35 V capacitors are usually used in power supplies with a 24 V output.
Having capacitors of different types installed is perfectly normal; I've seen this before. They were installed this way from the start, and there are no signs of anyone having resoldered anything on the board.

The board also has a power indicator LED, connected through a 23 kΩ resistor, so it barely glows.


I have no complaints about the circuit board either. Everything is generally familiar and well laid out, and the small components are additionally secured with glue. There are, however, some unpopulated spots in the secondary section.
The resistors in the high-voltage circuits are connected three in series to improve reliability and safety. The thing is, standard 1206 resistors are rated for voltages up to only 200 V, and since this circuit can reach 400 V, three resistors are used in series.


This is a single-switch flyback power supply, controlled by a PWM controller with an external power transistor. The PWM controller is the NCP1200D60R2. According to the datasheet, it has a built-in thermal protection system that shuts down the power supply when the temperature exceeds 140 degrees.
The output section contains nothing unusual: a standard adjustable TL431 shunt regulator in the feedback circuit and an output filter. There's also a footprint for a TL431 in a different package.



I usually redraw the power supply schematic next, but in this case there's no point, for several reasons:
1. The circuit is about as “standard” as they come.
2. These are no longer available for sale.
3. See the end for the third reason.

Anyway, let's move on to the tests. I used:
1. 150 W electronic load
2. Multimeter
3. Oscilloscope
4. Thermal camera



The input probe was connected with the necessary filter, consisting of two capacitors: a 0.1 µF ceramic capacitor and a 1 µF electrolytic capacitor. I described this connection method in one of my reviews.
The multimeter and load leads were also kept separate to eliminate the effect of voltage drop across them.


The test results were a bit disappointing, especially after the initial visual inspection.
1. With no load, the power supply delivered 24.25 V, which is perfectly normal.
2, 3. At load currents of 1 and 2 A, the voltage dropped slightly, which is still acceptable.
4. But at a load current of 3 A, the voltage dropped noticeably, to 21.5 V. That really shouldn't happen in a power supply rated for this current.
5, 6. Tests showed that 3.35 A is the maximum current this power supply can deliver. At that point, the output is about 17.7 V, or 60 W. Increase the load current further, and at 3.39 A the power supply goes into protection and repeatedly restarts.
After the initial inspection, I estimated that this power supply would be able to deliver at least about 4 A, but reality turned out to be considerably less impressive.


I also determined experimentally that the actual maximum current for this power supply is 2.75 A. Beyond that, the output voltage drops sharply, while the output power is 66.5 W. Up to this current, the voltage remains stable.
A quick explanation: proper overload protection works a little differently. The power supply holds the voltage steady right up to the limit, then simply shuts off the output once the protection threshold is exceeded, sometimes even in a latched mode. The fact that this power supply drops the voltage from 24 V to 17 V before shutting down is more typical of cheap power supplies.

You can see in the photo that I soldered an extra capacitor across the power supply output. I tried increasing the output capacitance a little, but it made no difference, so I simply left it in place.


Ripple waveforms.
1. No load
2. 1 A
3. 2 A
4. 3 A.

Interestingly, the ripple amplitude barely depends on the load current: it's about 40 mV peak-to-peak at 1 A and 50 mV at 3 A.


I also measured the efficiency. The results were:
1. No load: power consumption of about 1.2–1.4 W
2. 84.6% at 1 A
3. 86.4% at 2 A
4. 85.6% at 3 A.

That's an average result; it can be better or worse.


I also ran a temperature test, and its results likewise show that 66 W is the maximum power for this power supply. First, I ran the power supply for about 20 minutes, then for another 15 minutes or so. The hottest component was the transformer, which reached 80–85 °C. The high-voltage transistor reached a similar temperature. And the PSU wasn’t just outside its case—it wasn’t even inside its cover. In a closed, compact case, the temperature would most likely have exceeded 100 °C.


Just for comparison, here are the sizes of the three power supplies:
1. MeanWell EPS-65S-24 - 65 W continuous, 90 W peak.
2. MeanWell RPS-120-27 - 120 W continuous, 135 W peak
3. The unit under review.


That’s pretty much the whole review; there’s nothing else to show, so now I’ll try to draw a few conclusions.
1. The power supply is most likely genuine.
2. The input filter is excellent.
3. In terms of overall design, I’d give it a good rating, but it’s still a long way from excellent.
4. The stated specifications are clearly exaggerated.

All in all, this is a decent PSU with a real power output of around 50 W. I suspect the label may have said something like 24 V, 2.1 A. PSUs with specifications that are this far off are often bundled with branded equipment.

I won’t say I liked this PSU. Just 50 W continuous and no more than 66 W peak is a bit low for a unit of this size. But I don’t want to be too hard on it either, since it does work and has short-circuit and overload protection, as well as a decent input filter and fairly high-quality components. In any case, after this review I’d be unlikely to buy one, at least not at this price. That’s partly why I didn’t want to redraw its schematic.
By the way, this PSU would work quite well for powering a soldering station with a T12 iron, though you’d need to reduce the current-sense resistor value a little. The station doesn’t need much continuous power, and this PSU is better quality than the well-known “people’s choice” model.


That’s all for now. As always, I’ll be glad to answer any questions, and I hope this review was at least useful, even if it wasn’t interesting.