Voltage converter with an output current of up to 25 A and 600 W

Recently, one of my regular readers gave me a rather interesting converter, and I’m very grateful to him for it, of course. It was a nice surprise :) And, as usual, I decided to review it.

A reminder: to keep up with new posts on the blog, I recommend following me on Instagram, where I’ll post notifications about every new article and maybe write about something interesting too - link to my account

You can also subscribe to updates and new articles on the Telegram channel - https://t.me/KirichBlog

It came in an envelope and an antistatic bag.


Inside was the converter, and honestly, my first thought was: wow, it’s tiny. Well, judging by the specs, it wasn’t supposed to be big, but in the store photo it looks somewhat larger—which is pretty typical of AliExpress.


The build is neat, and considering the price, I think it’s excellent. It has active cooling, arranged so that the fan cools both the heatsink and the inductor. The fan is very quiet.
The input terminals are on one side of the board, and the output terminals are on the other. Next to the output terminals is a trimmer for setting the output voltage.
By the way, the listing title says “DC-DC CC CV Buck Converter.” The CC mode (output current limiting) isn’t adjustable here, unfortunately, so there’s only one trimmer.



There’s something like an instruction manual on the product page, but everything here is pretty intuitive anyway.


1. The input has three 100 µF, 100 V capacitors.
2, 3. It uses a synchronous rectifier, so there’s a transistor instead of a diode at the output. Both transistors are identical and are clearly used parts - MDP1991, 100 V, 120 A, 5.9 mΩ.
There are also three output capacitors, rated at 330 µF, 63 V. Unfortunately, there’s no output LC filter, although there was nothing stopping them from adding one.
4. The fan is rated for 12 V, but starts spinning at around 3.5-4 V. It measures 30x30x10.


Unfortunately, the markings on all the ICs have been erased. I’d guess the one at the bottom left is an auxiliary DC-DC converter based on some Chinese XLxxx chip, and the PWM controller might be a TL494, but I can’t say for sure. The one at the bottom right may be the transistor driver, since the current feedback goes straight to the PWM controller and both power transistors are N-channel. The only thing I don’t understand is why there are transistors next to the driver, apparently connected as a turn-off speed-up circuit for the MOSFETs (a transistor, a diode, and a couple of resistors). Since there are two identical sets on the board, I’m almost certain that’s what they are.


The underside is empty.


The dimensions are listed on the product page, so I’ll just repeat them here.


First power-up: the converter was initially set to an output voltage of 18 V.


With 61 V at the input, the full adjustment range is 2.5-52 V. The listing, however, specifies an adjustment range of 5-48 V with an input range of 12-75 V.


1. As I mentioned, the converter starts at around 4 V. The fan starts spinning at that voltage, and the output is about 0.5 V.
2, 3. With the output set to 12 V and 12 V at the input, I got about 10.3 V at the output with no load.
4, 5. Raise the input to 14 V and the output reaches the set 12 V, but you’ll need more headroom under load, of course.
6-9. Power consumption at different input voltages.


At this point I got curious: what’s the smallest possible difference between the input and output voltages?
It turned out that if you turn the trimmer all the way to the right, the difference is 9 V. In the photo, you can see that lowering the input voltage also lowers the output voltage in step with it.
Personally, I think that difference is a bit too large.


Then I remembered that the specified maximum output is 48 V, so I set it to that and repeated the test. With no load, the output reaches 48 V only if the input is at least 56 V—that’s an 8 V difference. However, as the output voltage is reduced, the difference between input and output also gets smaller. For example, if you set the output to 24 V, it stays there with an input of at least 28 V; lower the input to 27 V and the output drops to 23.2 V.

I think the dead time is set too high here, which significantly reduces the input headroom. It’s not exactly a critical issue, but it could be a problem in some situations. And this is with no load; under load, the difference should be increased by about 2–3 V.
The specifications also show that there is a decent margin: 12V-75V to 5-48V.



Next came various tests—not too many, but I tried to check what matters.
First, a load test: I connected the converter to an electronic load and powered it from an RD6024 power supply.


The converter goes into protection at around 26 A. I didn't test CC mode, but even here you can see that the voltage doesn't cut out immediately at the end; it drops instead, which indicates that the current-limiting function is working.


I also measured efficiency in different modes. The last mode, with a 13.8 V output, was chosen for a reason: it's for using the converter as a charger. And this is where the first problem showed up: voltage adjustment is rather coarse, and I could only set it to within about ±15 mV.

Measurements were taken over a load range of 2.5–25 A in 2.5 A increments. The upper output-voltage limit was set to 20 V to get around 500 W, although I actually loaded it to 600 W as well. The load has a hard time operating in that mode, so I tested only up to 20 V.

You can see a significant drop in efficiency at an output voltage of 5 V; the other modes are noticeably better, with efficiency around 93–95%. With active cooling, overheating is probably only possible at power levels close to the maximum.
The highest efficiency was achieved with 36 V in and 20 V out. In reality, efficiency is slightly higher, since I didn't account for the voltage drop in the input wires, but that difference can be ignored.



Next was a thermal test. I didn't run it for very long; it had three stages—10, 20, and 25 A—with 10 minutes per stage.



At 10 and 20 A, the converter performed quite well. The photo on the left is at 10 A; the two on the right are at 20 A. In the third photo, I covered the shunts so they wouldn't overexpose the image and so the other components would be easier to see. As you can see in that photo, the output inductor got the hottest, reaching about 80°C.


At 25 A, things got noticeably worse. First, the shunt heated up significantly, and the inductor reached 106°C. Otherwise, things were relatively good, and the shunts were probably the biggest problem. I'll get to them a little later.


I also checked how the output voltage varied with both load and temperature.
1. Initially, with no load, I set it to about 12 V—12.005 V, to be exact.
2. 10 A, cold.
3. 10 A after warming up.
4. 20 A, at the start.
5. 20 A after warming up.
6. 25 A, at the start.
7. 25 A after warming up.
8. No-load voltage when warm.
9. And the no-load voltage after cooling down.

You can see that, first, the voltage drops noticeably under load, and it also drops further as the unit warms up. I think the second issue is caused by the trimmer potentiometer and its poor placement. The thing is, these potentiometers already have a fairly poor temperature coefficient, and this one is installed in a spot where the temperature can easily reach 150–200°C.


Now, about the shunts—this is a whole separate topic.
The thing is, at high current they don't just heat up to a crazy 200+°C; they can also easily come loose from the board, because they get so hot that the solder melts.
Below is a photo of the resistors after warming up at 25 A. You can see traces of something transparent on them. When I took the thermal image at 20 A, I covered them with a small piece of plastic bag. The plastic simply melted through in that spot and left a mark on the resistors. Then, during the 25 A test, I just touched one of the resistors with tweezers to see what had stuck to it, and the resistor shifted and stuck to the tweezers. I had to quickly press it down with something else so I wouldn't accidentally lift it off the board.

Yes, at 25 A the shunt gets so hot that you can actually knock the resistors off the board—the solder is completely molten...


Ripple was checked with an input voltage of 36 V and output voltages of 5, 12, and 20 V at currents of 12.5 and 25 A, i.e. at 50% and 100% load.
But first, here's the ripple amplitude at a 20 V output with no load: about 100 mV. At an output voltage of 5 V, the ripple amplitude was about the same.


The ripple waveform is typical of converters with synchronous rectification, i.e. it's close to triangular.
1, 2. 36 V input, 5 V output, 12.5 and 25 A
3, 4. 36 V input, 12 V output, 12.5 and 25 A
5, 6. input 36 V, output 20 V at 12.5 and 25 A.

Note that the ripple amplitude depends mainly on the load current: at 12.5 A, it is 200–300 mV, and at 25 A, 300–400 mV.
Yes, in my opinion the ripple is high. The problem is made worse by the fact that these aren't just peak spikes, but the main ripple waveform itself, which is harder to filter out with just an inductor.
Still, adding an LC filter may help a little, but replacing the output capacitors with higher-capacitance, better-quality ones will probably help more than the oversized, low-quality parts installed here.


Conclusions.
As is often the case, the converter has a number of drawbacks, most of which are offset by its low price and generally decent build quality. Personally, I really miss having output current adjustment, though I think it could be added.
The issues I noted include relatively high ripple, a noticeable dependence of the output voltage on current and temperature, and excessive heating of the current-sense shunts.
Its advantages include neat assembly, fairly high efficiency, and no overheating at load currents up to 20 A (apart from the shunts).

Overall, I don't see any major problem with loading it to 25 A, but the current-sense shunts would have to be replaced, since at that current they dissipate about 3 W and overheat. The converter could also be improved by replacing the capacitors and moving the trimmer to a cooler spot—or, better yet, replacing it with a fixed resistor.

All in all, it's very good for the money, and it's also relatively easy to modify. I plan to use it to charge a battery pack, but first I'll need to:
1. replace the trimmer with a fixed resistor
2. replace the shunts (or the feedback circuit) so that current limiting starts at 15 A, for example, rather than 26 A.
3. possibly add a simple thermal switch to control the fan.

That's all for today. And thanks again for the converter—it was unexpected and much appreciated :)