Another power supply: 12 V, 30 A, and 360 W

In fact, this review is just an intermediate step toward testing more powerful power supplies that are already on their way to me. But I thought this model deserved a look too, so I ordered one for review.


A few words about the packaging.
A plain white box, with nothing to identify it except the part number.


Comparing it with the power supply from my previous review, I found that this one is just a little longer. That's because it has active cooling, so with almost the same case volume, it delivers one and a half times the power.
The case measures 214 × 112 × 50 mm.


All the connections are brought out to a single terminal block. The terminal assignments are stamped into the power supply's case. This is a little more durable than a sticker, but harder to read.
The cover takes a noticeable amount of force to close and locks firmly in place. When opened, it gives full access to the terminals. Some power supplies have covers that don't open all the way, so I always check this now.


1. A label on the power supply case lists the basic specifications: power, voltage, and current.
2. It also has a 115/230 V input voltage selector, which is unnecessary and not always safe to have in our power grids.
3. The power supply was manufactured almost a year ago.
4. There is an operation indicator LED and a trimmer for adjusting the output voltage next to the terminal block.


The fan is mounted on top. As I mentioned in my previous review, 240–300 W is about the upper limit for passively cooled power supplies. Of course, fanless power supplies with higher ratings do exist, but they're much less common and quite expensive. Active cooling is used here to save money and make the power supply cheaper.


The cover is secured with six small screws, but it also fits snugly on its own. The case is aluminum and, as with other power supplies, also serves as a heatsink.


For comparison, here's a photo next to a 240 W power supply. As you can see, they're mostly identical. Essentially, the 360 W model differs from its smaller sibling only by the addition of a fan and a few minor changes to accommodate the higher output power.


For example, the power transformer is the same size in both, while the output inductor in the unit under review is noticeably larger.
Both power supplies share one feature: a very spacious layout. That's justified in the passively cooled model, but with active cooling, the case could easily have been made smaller.


Before taking it apart any further, let's check that it works.
The output voltage is initially a little higher than the specified 12 V, though that doesn't really matter. I'm more interested in the adjustment range, which is 10–14.6 V.
Finally, I set it to 12 V and move on with the inspection.


Surprisingly, the input capacitor values match what's printed on them :)
Each capacitor is 470 µF, for a total of about 230–235 µF. That's well below the recommended 350–400 µF for a 360 W power supply. Ideally, each capacitor should be at least 680 µF.

The output capacitors have a combined capacitance of 10,140 µF, which is also not much for the specified 30 A. But branded power supplies often have capacitors with similar capacitance.


The transistors and output diodes are pressed against the case through a heat spreader. The only insulation is a piece of thermally conductive rubber.
More expensive power supplies usually use a thicker rubber cap that completely covers the component. It's not particularly necessary for the output diodes, but it would certainly help with the high-voltage transistors. That's why I recommend grounding the power supply case for safety.
The heat spreaders are pressed against the aluminum case, but there's no thermal paste between them and the case.

After an incident with one of my power supplies, I now always check that the power components are clamped properly. There's no problem here. Double components rarely cause trouble anyway; problems are more common when there's a single high-power component held down by an L-shaped bracket.


The fan is a standard sleeve-bearing model, but for some reason it's rated for 14 V.
It's 60 mm in size.


Let's keep taking it apart.
The board is held in place by three screws and the mounting points for the power components. The bottom of the case has a protective insulating film.


The filter is fairly standard for power supplies of this type. The input bridge rectifier is marked KBU808 and rated for up to 8 A and 800 V.
There is no heatsink, although one would be advisable at this power level.


1. A 15 mm diameter, 5 Ω thermistor is installed at the input.
2. An X2-class interference suppression capacitor is connected across the mains.
3. The interference suppression capacitors directly connected to the mains are Y2 class.
4. A standard high-voltage capacitor is installed between the output common and the power supply case. That is sufficient here, since without grounding it is connected in series with the Y2-class capacitors shown above.


The PWM controller is a KA7500, equivalent to the classic TL494. The circuit is more than standard; manufacturers simply churn out identical power supplies that differ only in the ratings of some components and the specifications of the transformer and output inductor.
The inverter output transistors are also a classic choice for inexpensive power supplies: MJE13009.


1. As I mentioned above, the input capacitors are rated at 470 µF. Interestingly, when the capacitors have an unfamiliar name, the stated capacitance is more often accurate; when they are counterfeit, such as Rubicong, it is more often understated. Just an observation. :)
2. The output transformer core measures 40×45×13 mm. The winding is varnish-impregnated, though rather superficially.
3. There is a connector next to the transformer for a fan. Descriptions of power supplies like this usually claim automatic speed control, but there isn't any here. The fan does change speed slightly depending on output power, but that's probably just a side effect. When switched on, the fan runs very quietly; it reaches full speed at around 2.5 A, which is less than 10% of the maximum.
4. The output has a pair of MBR30100 diode assemblies, each rated at 30 A and 100 V.


1. The output inductor is noticeably larger than the one in the 240 W version. It is wound with three wires on two 35/20/11 cores.
2. As expected from the preliminary check, the output capacitors are rated at 3300 µF. Since they are new, their combined measured capacitance was 10140 µF, not 9900 µF. The voltage rating is 25 V. The manufacturer is the well-known noname.
3. Current-sense shunts for the short-circuit and overload protection circuit. Usually one such “wire” is used for 10 A, so this 30 A power supply has three of them. But there are 7 positions, so I assume there is a similar version rated for 60 A and a lower voltage.
4. Here's a small difference: the components responsible for shutdown when the output voltage is too low have been moved closer to the output, although their reference designators have been kept the same as in the schematic. In other words, R31 in the 36 V power supply schematic corresponds to R31 in the 12 V power supply schematic, even though they are in different places on the board.


At a quick glance, I'd give the soldering a solid B: everything is clean and neat.


The soldering is quite good, and protective slots have been cut into the narrow sections of the board.


But I did find a “fly in the ointment.” Some components have cold solder joints. The location isn't particularly important; the fact that they're there is.
In this case, poor soldering was found on one lead of the fuse and on a capacitor in the output undervoltage protection circuit.
It takes just a few minutes to fix, but as they say, “the spoons have been found, but the bad feeling remains.”


Since I had already drawn the schematic for a similar power supply, I simply updated the existing schematic in this case.
I also highlighted the modified components in color.
1. Red: components that change when the output voltage and current are changed.
2. Blue: I don't understand why these component values change when the output power stays the same. The input capacitors are partly understandable: they were specified as 680 µF but actually measured 470 µF. But why increase the capacitance of C10 by half?
There is an error in the schematic: C10 is 3.3 µF, not 330 nF.


That's it for the inspection; on to the tests. For these, I used my usual “test bench,” with a wattmeter added this time.
1. Electronic load
2. Multimeter
3. Oscilloscope
4. Thermal imager
5. Thermometer
6. Wattmeter, no review available.
7. A pen and paper.


At no load, ripple is practically nonexistent.


A small clarification about the test. The electronic load's display will show current readings noticeably lower than the values I'll report. The load can handle high currents in hardware, but its software limits it to 16 A. So I had to use a workaround: I calibrated the load for twice the current, meaning 5 A on the display corresponds to 10 A in reality.

At load currents of 7.5 and 15 A, the power supply behaved identically; peak-to-peak ripple was about 50 mV in both cases.


At load currents of 22.5 and 30 A, ripple increased noticeably, but remained at the same level in both cases. The ripple started to rise at around 20 A.
Peak-to-peak ripple came to 80 mV.
The output voltage regulation is excellent: as the load current changed from zero to 100%, the voltage changed by just 50 mV. What's more, the voltage rises as the load increases rather than dropping, which can be useful. The voltage remained unchanged as the unit warmed up, which is another plus.


I put the test results into a single table showing the temperatures of individual components.
Each test stage lasted 20 minutes. The full-load test was run twice to let the unit heat up.
I put the cover with the fan back in place but didn't screw it down. To measure the temperature, I removed it without disconnecting the power supply or the load.


I also took a few thermal images.
1. The wires to the electronic load heating up at maximum current; you can also see heat radiating from the internal components through the slots in the case.
2. The diode assemblies get the hottest. I think adding a heatsink, as in the 240 W version, would have significantly reduced their temperature.
3. Another major problem was getting the heat out of the whole setup, since its total power dissipation exceeded 400 W.

Speaking of heat dissipation, when I was preparing the test, I was more worried that the load would struggle to handle this much power. I've tested at this power level before, but 360–400 W is the maximum my electronic load can dissipate continuously. For short periods, though, it can handle 500 W without any trouble.
But the problem turned up somewhere else. The heatsinks for the power components have thermal switches rated for 90 degrees. One contact on each is soldered, but I couldn't solder the other, so I used terminal blocks.
At 15 A through each switch, these contacts started getting quite hot, causing the switches to trip prematurely. I had to actively cool this assembly too. I also had to partially take the load off the electronic load by connecting several high-power resistors to the power supply.

The switches are only rated for 10 A, though, so I didn't expect them to work properly at 1.5 times their maximum current. Now I'm thinking about how to modify them; it looks like I'll have to build an electronic protection circuit controlled by these thermal switches.

And now I have another task. At the request of some readers, I ordered 480 and 600 W power supplies for review. Now I'm wondering what to use as a load, since my electronic load definitely won't handle that much power (let alone currents of up to 60 A).


As last time, I measured the power supply's efficiency. I plan to include this test in future reviews as well. The measurements were taken at 0/33/66 and 100% load.

Input - Output - Efficiency.
5.2 - 0 - 0
147.1 - 120.3 - 81.7%
289 - 241 - 83.4%
437.1 - 362 - 82.8%


So, what can I say in conclusion?
The power supply passed all the tests and delivered pretty good results. It also has a noticeable thermal margin, but I wouldn't recommend loading it above 100%. I was pleased with the very stable output voltage and the fact that it doesn't depend on temperature.
What I didn't like were the unbranded input and output capacitors, some sloppy soldering, and mediocre insulation between the high-voltage transistors and the heatsink.

Otherwise, it's a pretty ordinary power supply: it works, holds its voltage, and doesn't get too hot.

That's all for now. As usual, I'm looking forward to your questions.