Industrial power supply PS-220/24-1.5 from Novatek-Electro
As usual: inspection, teardown, tests, and conclusions.
The link in the title goes to the Ukrainian website. I wanted to add a link to the Russian version, but unfortunately, they don’t have these power supplies there. I did find them on Rozetka for 2500 UAH and on prom.ua for 2250, as well as on some Russian website and even in Belarus.
Of course, many of you noticed the price right away. Yes, I agree, it’s not cheap, but this is clearly not a consumer-grade device. Personally, though, I was simply interested in a power supply from a domestic manufacturer.
The packaging is as simple as it gets: a regular resealable bag and a paper manual.

The manual contains all the necessary information about the entire range of power supplies, along with the warranty terms. It states that the manufacturer guarantees trouble-free operation for 3 years. Note that this is specifically a guarantee of functionality.

The full manual is available via this link; I’ll include only the most essential excerpts.
The series includes five models, differing mainly in output voltage and current: 5, 12, 24, 36, and 48 V, with corresponding currents of 7, 3, 1.5, 1, and 0.75 A. I ordered the mid-range version for this review: 24 V, 1.5 A.

The first thing that catches your eye is the rugged construction. The power supply has an all-metal enclosure—not some “foil and cardboard” mix, but perfectly decent, fairly thick metal.
There are two status LEDs on top, along with an opening for the trimmer that adjusts the output voltage.
When power is applied, the blue “mains” LED lights up immediately. A little later, the green LED comes on to indicate that the power supply has started and power is being delivered to the output. The “mains” LED is clearly connected directly across the mains, since it has a noticeable flicker at the mains frequency.

The terminal blocks are good quality, elevator-style. What I found unusual was that the input power terminal block is on the bottom rather than the top. The output terminal block has four terminals, two per pole. There’s no “power OK” output here, though I think one would have been useful.
There are screws at the bottom of the side panels, but for some reason their heads are not countersunk, so the overall width is greater than it could be. Perhaps these screws also act as spacers, preventing anything from being mounted right up against the side panels.
The metal DIN-rail clip, held under tension by a thick spring, looks just as rugged.

Including the protruding screw heads, the enclosure is 70 mm wide, equivalent to four standard units. The power supply is noticeably taller than a standard module, but that’s normal for various types of industrial equipment.

Overall dimensions of the power supply

Before testing, I took the power supply apart for inspection, and its construction surprised me a little.
The thing is, it consists of four circuit boards, and the sides of the enclosure serve as heatsinks. All the connections between the boards use connectors, which made inspection much easier.
The cover is also made of metal, about 1 mm thick.

The top board holds a fuse, varistor, thermistor, TVS diode, multiturn potentiometer, connectors, an op-amp, X and Y capacitors, and various other small components.

The layout is quite dense.

I unscrew the side screws. By the way, they were really tightened “with all their might”—a solid job.
Here’s the power supply almost completely disassembled.

The way the sections are arranged is unusual: one board carries the input bridge rectifier, filter capacitor, high-voltage transistor, and control circuit.

The input capacitor is 100 µF, even though the power supply is rated at just 36 W. Next to it is a logic IC with six buffer amplifiers and an optocoupler.
There’s also a small bridge rectifier and a BD437 transistor.
Most of the components are on the underside of the board. The side of the case is not only made of aluminum, but also milled around the perimeter.

The high-voltage transistor is in an insulated package.

From what I can see, the circuit is based on a self-oscillating converter.

A closer look. By the way, the marking on one of the components is somehow mirrored.

The transformer, output diode assembly, and output capacitors are on a second board. In my opinion, this is a rather odd design choice: in circuits like these, the conductors are usually kept as short as possible, but here they also run through an interconnect board.

I didn’t remove this board, since the underside only has the diode assembly, which is screwed to the other side of the case.
The output capacitors have a decent total capacitance: two 2200 µF Hitano capacitors before the output inductor and one 1000 µF Samwha WL after it.

Let’s put it all back together and move on to the tests.
I connected the wires and powered it up—and was surprised for the second time. The output is 24 V with no load, which is what I expected, but the power consumption of around 7–9 W was quite a surprise, especially after using modern power supplies with a “green” controller.
I suspect the high no-load power consumption is due to the operating temperature range, which starts at -30 °C. The same may be true of the higher-capacity output capacitors.
The output voltage can be adjusted from about 22 to 30 V.

There’s also a strange behavior I can’t explain. If I turn the trimmer toward a lower voltage, the voltage keeps dropping afterward anyway. For example, it starts at 22.7 V, gradually reaches 22.0 V, and then continues to fall.
If I connect a small load to the output, say a 1–1.5 kΩ resistor, the voltage drops to 20.7 V. But when I disconnect the resistor, it rises again and then starts gradually falling once more.
Anyway, after playing around with it, I set the output to 24 V for the tests that follow.

Besides their extended temperature range, these power supplies have another key feature: they can be connected in parallel with other units of the same type.
The product manual includes a wiring diagram and instructions for setting up the power supplies to work in this mode.

The reason for this feature is quite simple: these power supplies don’t have hard overload protection. As the load current rises above a certain level, they start reducing the output voltage.
The manual includes a graph of output voltage versus load current. It shows that the voltage should start dropping when the load current exceeds 100% of the maximum.
This feature also lets you use these power supplies as battery chargers, as described in the manual.

I connected the power supply to an electronic load, and it managed to surprise me again. The graph I showed above indicated that the power supply should start reducing the voltage when the load current exceeds 1.5 A. But in my case, the voltage started dropping only above 3.3 A—almost twice the stated 1.5 A.
In practice, the power supply delivered 77 W instead of 36 W.
This surprised me, so I got in touch with the manufacturer. It turned out that this behavior is normal:
the current limit of more than 1.6–1.7 A kicks in when the mains voltage is 135 V. At the usual 220–240 V, the current limit is around 3–3.5 A.
In my opinion, this is not right. They should either have provided a graph for two input voltages so users would understand that the current depends on the input voltage, or put the current-sense resistor on the power supply output rather than the input. Then the current would not depend on the input voltage.
The power supply remained stable down to an output voltage of about 0.8 V, although it started squealing below 5 V. I also noticed a slight dependence of the output voltage on temperature: after warming up, it rose by 0.02 V, or about 0.1%.

I also measured the power supply's efficiency, and there were some quirks here too. If you look closely at the graph, you can see a point where efficiency rises sharply.
Measurements were taken from 0.25 to 5.0 A in 0.25 A increments. Between 3.25 and 3.50 A, the power supply's output voltage started to drop.

The sudden rise in efficiency caught my attention, so I separately tested the 0.2–1.7 A range in 0.1 A increments. It became clear that efficiency rises sharply between 0.9 and 1.0 A.
I think this quirk is caused by the high no-load power consumption being a “feature” rather than a “bug.” In other words, at a light load the power supply deliberately heats itself, while at a heavier load the ordinary losses in the diode assembly, transistor, and so on are enough.

At idle, the power supply can draw around 9 W with no load, while the case temperature on the high-voltage transistor side reaches 57–60 degrees. And that's just sitting on the table, without being installed in a panel.

Measuring the output ripple amplitude of the power supply.
First, with a timebase of 5 µs per division, with no load and at currents of 1, 2, 3, 4, and 5 A,
For my test conditions (25°C, 230 V mains), the manufacturer specifies ripple of up to 30 mV p-p. My results were better, especially since the 30 mV limit is specified at 1.5 A. But during testing, the “wobble” was already bothering me, making it difficult for the oscilloscope to trigger.

When I switched the oscilloscope timebase to 10 ms/div, I understood what was causing the “wobble.” In this display mode, you can see that the total low-frequency ripple reaches 100 mV at a 2 A load and 130 mV at higher load currents. The two bottom screenshots, taken at 4 and 5 A (with the power supply in current-limit mode), also show that all of this is modulated at an even lower frequency.

It also turned out that the power supply has a certain load transition range where ripple increases very noticeably.
Below are oscilloscope traces at 0.3, 0.4, 0.5, and 0.6 A. It's clear that at 0.4–0.5 A, ripple easily reaches 250 mV. The width of the range where this happens depends on temperature and narrows as the unit warms up; initially, it can span 0.4–0.8 A.
Unlike the previous traces, I had to set the vertical scale to 50 mV/div here, rather than 20 mV/div as above.

The warm-up test was carried out in four modes, at load currents of 1, 2, 3, and 3.5 A, with each stage lasting 20 minutes.
During the 3 A test, the power supply slightly reduced its output voltage after 10 minutes, so these results can be considered representative of its maximum-power operating condition.

Most of the heat is dissipated through the sides of the case. The left side was a little hotter than the right, but once temperatures had stabilized at currents above 2 A, the difference was just 1–3 degrees.
Maximum temperature at load currents of 2, 3, and 3.5 A. I haven't included the temperature at 1 A; it was only a few degrees lower than at 2 A.

Conclusions.
As usual, I'll start with the good. I liked the solid approach to the design: thick metal, a good paint job, and a case that doubles as a heatsink. I'll also praise the neat assembly and quality components. The capacitors have plenty of capacitance headroom, and it's clear the unit was designed for reliable operation at low temperatures. I also liked that the power supplies can easily be connected in parallel to increase power, and can even be used to charge batteries. The unit clearly exceeds its rated 36 W output power.
There were some drawbacks too. The ripple is too high at low frequencies, which is surprising for a power supply with this much input capacitance. I also expected the unit to limit output current to 1.5–1.7 A, but it turns out it only does that when the input voltage is low. The power supply's design and circuitry are overly complicated, which has also contributed to its price. I didn’t like how much it heats up with no load. In this mode, the temperature of the side panel where the high-voltage transistor is mounted is even higher than at full load.
As for current limiting and parallel operation, I have a power supply review with a similar feature, and in my opinion it’s implemented much more elegantly there.
And now, briefly.
To be honest, I expected more, and I was a bit disappointed during the tests. No, the power supply works, and I think it will be reliable, but dissipating 7–9 W as heat with no load seems excessive to me. I also expected the current limit to kick in at 1.5–1.6 A, not 3–3.3 A—that’s what the graph in the manual showed. And the low-frequency ripple is too high.
I contacted the manufacturer and asked them a few questions about the issues listed above. They replied that the power supply is supposed to work this way, and that it has been in production for a long time with no complaints. I have no concerns about its reliability, and in the industrial sector that’s one of the main requirements.
I think the situation is quite simple: when it was designed and production began, it met the requirements just fine, but over time the requirements became stricter, so it no longer fully meets them.
That’s all from me. I hope you found this useful.
1, блок 220/24/15 расшифровка, блок питания 220/24 это, температура блока питания 220/24, блок питания промышленный, обзор блоков питания промышленный
Questions and answers
What is the current limit of the Novatek-Elektro PS-220/24-1,5?
The author expected a limit of 1.5–1.6 A, but observed limiting at about 3–3.3 A. In the author's tests, it limits to 1.5–1.7 A only when the input voltage is reduced.
How much power does the PS-220/24-1,5 use at no load?
At an output of 24 V, no-load consumption is about 7–9 W. The author also noted that the case gets noticeably warm in this condition.
What drawbacks did the PS-220/24-1,5 tests reveal?
The author found high no-load consumption and heating, as well as excessive low-frequency ripple. However, the author has no concerns about the power supply's reliability.
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!
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



