Sanmim GPN30E-15: a power supply that really surprised me
To keep up with new posts on the blog, I recommend following me on Instagram, where I’ll post notifications about new articles and maybe write about something interesting now and then: link to my account
You can also subscribe to updates and new articles on the Telegram channel - https://t.me/KirichBlog
I’ll show you what I ended up using the power supply for at the end. For now, let’s take a look at the unit under test.
It’s made by a fairly well-known company that, in my view, is something of a competitor to the well-known Mean Well, whose power supplies I’ve written about several times.
This model belongs to the GPN30E series and, as you’d expect from a decent power supply, it comes with the manufacturer’s datasheet.
It’s also sold on AliExpress, but it costs more there: $11.30 + $1.07 = $12.37 for shipping, compared with 253.5 UAH ($6.76) at Kosmodrom. The savings are obvious, in my opinion, and you don’t have to wait, either.

It comes in a plain white box. I think all of this company’s switching power supplies are packaged the same way.

The model is available with six output voltages: 5, 9, 12, 15, 24, and 36 V. I’ll be testing the 15 V version.
I’ll also point out the specified temperature range, which starts at -30 degrees.

Compared with modern GaN chargers, I wouldn’t call this model compact. Even next to the familiar Mean Well units, it looks relatively bulky. On the other hand, it’s neatly built and, importantly, looks repairable.
It measures 100 × 50 mm, roughly the size of a standard matchbox, with a 2:1 aspect ratio.

The datasheet includes a dimensioned drawing, and it was clearly made by someone with an artistic streak—they even added color to make it easier to understand. It looks nice, all in all :)

The build looks neat. If you ignore the fact that some components are from lesser-known manufacturers, it’s hard to tell it apart from a Mean Well unit.
The input and output connect via terminal blocks, and I do have some reservations about them. They’ll work, of course, but in my opinion they don’t clamp the wires very well, so crimped wires are a better choice.

1. There’s a proper mains filter at the input, and you can also connect protective earth.
2. Thermistor, varistor, X capacitor, fuse—the usual setup. The only oddity is the 560 V varistor (about 400 V RMS), which is clearly only there to suppress transients.
3. The input capacitor is rated at 400 V and 47 µF, which is quite sufficient for a 30 W power supply.
4. The high-voltage transistor is a 7N65C, rated for 7 A and 650 V.

1. The feedback optocoupler is positioned between the two heatsinks. There’s a Y capacitor nearby. I have no concerns about the clearances between the primary and secondary sides, though I’d add isolation slots.
2. The output section; you can also see the inductor that reduces output ripple.
3. The output rectifier is an MBR20100F Schottky diode assembly, rated for 20 A and 100 V—more than enough for this model.
4. Aishi output capacitors, 1000 µF, 25 V

The PWM controller is marked 63H50. A search keeps pointing to NCP1251, and judging by the PCB traces, that does seem to be what it is.
There are three current-sense resistors, each 1.2 Ω, for a total resistance of 0.4 Ω. I’ll come back to them later.

There’s nothing on the underside but traces, which can actually be a good thing.

And, of course, the tests.
First, I simply plug it into the wall. The output is 15.086 V, and the wattmeter reads zero. I don’t really trust it, though—the one I did trust burned out :(

For the rest of the tests, I’ll use an EBC-A10H battery tester with a four-wire connection, so the measurements should be as accurate as possible.

Next came the usual load test, and this is where I got a surprise. Out of habit, I set the limit to 150% of the maximum current—that is, 3 A for a 2 A output current. But the load itself stopped the test; the power supply’s protection didn’t trip. After that, I raised it to 3.5 A, then 4 A. I was going to stop there, since that’s twice the rated current, but curiosity got the better of me: I set it to 4.5 A and decided to see what would happen...
The protection kicked in at around 4.2 A. That’s quite high, but what’s interesting is that the output voltage stayed rock solid: 15.074 V at 0.1 A and 15.10 V at 4.1 A. That’s a very good result. Still, I was concerned that the protection only kicked in under such a heavy overload.

While I was at it, I measured efficiency over the 5–40 W range. Each horizontal division represents 5 W. At the top end, I got around 80–82%. I wouldn’t call that particularly high, although it’s typical for power supplies like this.

I ran the thermal test at load currents of 1, 2, 2.5, 3, 3.5, and even 4 A. All tests except the last one ran for about 30 minutes; the last one ran for about 15 minutes. I stopped the test myself—the protection didn’t trip. I should also point out how stable the output voltage was: it started at 15.093 V and ended at 15.08 V, with the supply warmed up and delivering 4 A.

And the results surprised me a second time, after I’d already been ready to criticize this model for its excessively high protection threshold.
But first, here’s the graph from the datasheet, where the manufacturer specifies normal operation at 100% load, with an input voltage of 110–240 V and an ambient temperature of up to 55 degrees.

It was of course noticeably cooler in my room—just 20 degrees—but here’s what I got:
1. 1 A, 50% load
2. 2 A, 100% load
3. 2.5 A, 125% load.
No issues at all; you could say everything stayed cool. The only things that stood out were the snubbers and the output load resistor.

It gets more interesting from here
1. 3 A, 150% load
2. 3.5 A, 175% load
3. 4 A, 200% load.
Basically, nothing changed dramatically. At 200% load, the temperatures of the key components were:
1. Transformer core — 69 degrees. This reading was taken at 175% load; I forgot to measure at full load. At 200%, it was around 75–80 degrees.
2. High-voltage transistor — 85–90 degrees
3. Output diode assembly — around 100 degrees.
Those are genuinely excellent results for a twofold overload. I now understand that the protection trip current is quite reasonable: even at elevated temperatures, the supply is unlikely to burn out. Still, I’d personally recommend removing one of the current-sense resistors. That will reduce the protection trip current by a third, to around 2.8 A, making the supply even more reliable.

And, of course, the ripple. The datasheet specifies up to 100 mV peak-to-peak with the 20 MHz filter enabled. I didn’t use an additional C+C filter; I measured directly at the power supply’s terminals.
Results for these operating conditions:
1. No load
2. 1 A, 50% load
3. 2 A, 100% load
4. 3 A, 150% load.
At 100% load, I got just 60 mV, and even at 150% load the supply stayed within the specified limit. The oscilloscope showed just 80 mV.

The low-frequency ripple at 100 Hz is also well within spec. The test conditions were the same as for the high-frequency ripple measurements.

And, as promised, here’s the device this power supply was intended for. The photo was taken during testing; the dissipated power shown is 1018 W. It’s a prototype load designed for a continuous power rating of at least 1500 W. In a recent video, I said I was planning to build one. Next, I plan to give it a proper finish, then work on a 3000 W model. Essentially, this is an evolution of this project. But I’ll go into more detail another time.

Well, what can I say? At first, I was unpleasantly surprised that the overload protection only kicks in above 200% load. But after running the load tests and checking the temperatures, it turned out that the supply handles that kind of overload just fine. The situation will of course change at higher ambient temperatures, but even then, the key components shouldn’t exceed their critical temperature limits. I still increased the current-sense shunt value a little, though.
The only real drawback is the cheap input and output terminal blocks. They look out of place and don’t match the otherwise quite high standard of the device.
That’s all for now. I hope you found this information useful.
sm-gpn30e, sanmim gpn30e-24, dain/mpx1 40/100/21/c
Questions and answers
What output voltage does the Sanmim GPN30E-15 maintain under load?
The no-load reading was 15.086 V. From 0.1 A to 4.1 A, output remained around 15.07–15.10 V.
How much ripple does the Sanmim GPN30E-15 have?
With the 20 MHz filter enabled, high-frequency ripple measured 60 mV at 2 A and 80 mV at 3 A.
What happened when the Sanmim GPN30E-15 was overloaded?
Protection tripped at about 4.2 A, more than twice the rated current. The supply completed a 4 A load test, although key components heated up; the author nevertheless slightly increased the current-sense resistor value.
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!
Comments19
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




Comments are published in their original language.