24 V 5 A Power Supply, Plus 1 A as a Gift :)

Since I've been working with various power supplies for a long time, I can often tell how good one is just “from a photo.” That was the case this time too: I ordered a couple that looked pretty good to test. So far, my intuition seems to have been right. As always, more after the jump :)

People often ask me whether a particular power supply is any good. I usually answer: it looks good, but I can't say for sure until I've tested it. Experience shows that appearances can be quite deceptive.
Besides appearance, price matters a lot. A cheap power supply can't be good, but that doesn't mean an expensive one can't be bad. Let's find out.
I'll try to keep this review short while still giving you as much information as possible.

First, a little about the packaging. In short, there is some :)
In fact, cardboard packaging is more important for open-frame power supplies. A power supply in a metal case is much harder to damage, but packaging is always a plus: even if it doesn't protect the unit, at least the parts won't get lost :)
No manual is included with the power supply.


As I mentioned at the beginning, the power supply has a perforated metal case. The design is quite standard: an aluminum chassis that acts as a heatsink, and a perforated cover.
Cooling is passive because the power rating is relatively low. Active cooling is used for power ratings of 240–300 W and above.


There's a label on the side with the power supply's specifications.
I'm not quite sure what the 110/220 marking means, since the power supply has no voltage selector and is rated for 220 V only. Judging by the circuit design, though, it should work across the full 85–240 V range. For now, let's assume it's a 220 V model.
Specifications:
Input voltage — 220 V ±15%
Output voltage — 24 V
Maximum load current — 5 A


For this power rating, the dimensions are fairly compact:
Length — 143 mm
Width — 58 mm
Height — 41 mm


Screw terminals are located at both ends of the power supply:
1. For connecting the power input and ground
2. For connecting the 24 V output. This end also has an LED indicating output voltage and a trimmer for adjusting the output voltage.
The manufacturer seems to have used the same type of terminal block as on the output, but wired it with two negative terminals and one positive terminal.


Removing the cover isn't very convenient. You have to undo two screws on the sides, and the clips were a bit of a struggle. I had to pry the first one open with a flathead screwdriver.


Now for some of the power supply's features, both good and not so good.
First, the input filter. It has one—and not just any filter, but an almost proper one. There's a common-mode choke (clearly rated for a decent current), two X capacitors and two Y capacitors. The only thing missing is an NTC thermistor to limit inrush current.
The bridge rectifier is a GBU6D, which left me puzzled. According to the datasheet, it's rated for 6 A and 140–200 V, yet it passed all the tests with flying colors, even though the mains voltage was 240–245 V (we often get that kind of voltage at night). It also survived about 30 power cycles (I tested this deliberately). It feels like the bridge is actually rated for 600–800 V, and they simply printed a different marking on it—like processors with different clock speeds but the same die. What's more, the more common 600–800 V bridges often cost even less.


Sometimes I think Chinese engineers read our forums :)
The input has two 82 µF capacitors, for a total of 164 µF. That's more than enough for the rated power of 120 W.


The transistor and output diode are pressed against the case through an aluminum plate, using a small metal clamp as usual.
One small question here: if there are Y1 capacitors on the input, why use an ordinary capacitor between the windings? I see this quite often, even though that capacitor costs next to nothing.


The output capacitors are also quite large: two 2200 µF, 35 V capacitors.
What's odd is that the capacitors are different sizes despite having the same capacitance and voltage rating, and they're made by different manufacturers too.


They didn't forget the fly in the ointment: instead of an output noise-suppression choke, there's a “specially trained” jumper. This cost-cutting seems a little odd given the proper input filter.


On first power-up, the output voltage was 23.78 V.
Checking the output voltage adjustment range:
Minimum: 20.71 V; maximum: 29.79 V.
In the end, I set it to the rated 24 V.


Let's take a closer look :)
Unscrew the screw holding the transistor and diode in place. It's under the specifications sticker.
The transistor and diode are pressed against the heatsink through thermally conductive rubber pads, with thermal paste applied as well.


The board is marked WJXPS-P1210B6, with a design date of September 25, 2013.
The first part of the marking suggests the board was originally designed for a 12 V 10 A power supply—that's still 120 W—and that the 24 V 5 A version was released later using the same board, with a different transformer and a few component changes.
A 24 V power supply usually has slightly higher efficiency. If the 12 V version was properly rated, that would explain the “bonus” 1 A we got.


A little more detail about the components.
1. Y1 noise-suppression capacitors
2. The PWM controller is an OB2269CP from the well-known manufacturer LITEON.
3. An Infineon high-voltage transistor, 20N60C3, in a TO-247 package.
4. The output diode assembly is a BYQ28E-200, a pair of 10 A, 200 V UltraFast diodes.
Schottky diodes are usually used at the output, but in this case, using an assembly of regular fast diodes is perfectly justified: at high voltages, Schottky diodes lose some of their key advantages—their low forward voltage drop. 200 V is more than enough here.


I had no complaints about the circuit board. It has slots in the necessary places to prevent arcing across the PCB. The traces carrying high current are tinned.


The soldering is average quality. Everything is soldered, but the component leads haven't been trimmed very neatly.


Measuring the capacitor capacitance showed that it matches the stated value, which is always nice :) But when I measured the ESR of the output capacitors, my meter gave up and displayed an impossible 0 Ω. I don't know why; I've never seen that happen before.


I drew a schematic based on the circuit board. I tried to follow the component reference designators, but I'm not sure they're all correct, since they weren't visible everywhere on the board.


All right, that's it for the teardown. Let's move on to the mandatory testing.
The test setup is the usual one: an electronic load, oscilloscope, multimeter, non-contact thermometer, pen, and paper.
The procedure is the same as in my previous reviews:
Power on, apply a 1 A load, let it warm up for 20 minutes, measure the temperatures of the main components, then increase the current by one step.


1. No load: voltage 23.98 V, ripple 100 mV
2. Load current 1 A: voltage 23.93 V, ripple 0.3 V


1. Load current 2 A: voltage 23.77 V, ripple 0.6 V. That's quite a lot.
2. Load current 3 A: voltage 23.62 V, ripple 0.6 V.


1. Load current 4 A: voltage 23.5 V, total ripple 0.8 V
2. Load current 5 A: the voltage dropped to 23.43 V, but is still within the normal range. Surprisingly, ripple is lower at 0.6 V. However, only the positive-going ripple is visible in the shot.


Now about the “bonus” amp :)
Since the power supply performed excellently in terms of heating, I decided to continue the test.
1. Load current: 6 A, voltage: 23.5 V, ripple: 0.8 V.
2. I decided to try modifying the power supply a little by installing the missing inductor and three 0.22 µF ceramic capacitors: one before the inductor and two after it.
As they say, the difference is visible to the naked (well, almost) eye: the ripple dropped by half and became much less frequent.


For the modification, I used an inductor that wasn't quite suitable: it has low inductance and is rated for high current. An inductor with about five times the inductance would easily fit in the power supply and reduce the ripple even further.
I also figured out why the voltage sags under load. At first I thought the power supply “couldn't handle it,” although that's a little odd for a PSU with output feedback. By cooling the components in the voltage-sensing circuit one by one, I found that the problem was a shift in the value of resistor R2. Heating lowers the output voltage. Replacing R2 with a precision resistor will completely solve the problem. The transformer heats the resistor; you could simply move it farther away, but replacing it is better.


And here are the temperatures I measured.
By the end of the experiment, the case had reached 70 degrees near the power components.
It's worth noting that cooling during the test wasn't very good: it was summer (I didn't turn on the air conditioner), and the unit was sitting on a table, which is a poor heat sink and makes it harder to cool the bottom of the PSU.
The final readings were taken after two hours of stepwise warm-up.


Summary.
Pros
Comes in a box
Component ratings have plenty of headroom (except for the question of the input bridge rectifier)
The thermal performance means this PSU can be used with a load current of up to 6 A
A decent mains input filter.
Neat, high-quality construction.
Output voltage can be adjusted over a wide range

Cons.
High ripple (at least without modification)
Poor thermal stability of one of the feedback divider resistors.
Incorrect type of interwinding capacitor
Unclear situation with the input bridge rectifier.

My opinion: I have very mixed feelings about this power supply. On the one hand, it's a good unit with plenty of current headroom and capacitors with sensible ratings, but it still needs some modification. The fix is cheap and simple, but it's advisable; it makes the PSU perform much better. Why skimp on something like that?
The other unclear issue is the input bridge rectifier, but I'm still inclined to think it's a proper bridge that's been mislabeled. As I wrote above, diodes and bridges rated for 600–800 V are often even cheaper than their low-voltage counterparts. Besides, the PSU passed all the tests with flying colors, even with the mains voltage above spec.
I'll try to test my theory later today. If it burns out, so be it.

Would I recommend it? Hard to say. If you know your way around a soldering iron, this is a way to get a good power supply. Or if ripple and voltage drop don't matter to you (you can set it to 24.25–24.3 V from the start, which will work better). For everyone else, I don't know—it's up to you. I've tried to provide all the information you need.

I hope this review was useful and helps you make the right choice.
For some reason, the main link redirects to the store's home page. Maybe this one will work.

The product was provided by the store for review.