/       /       /    Mean Well EPS-65S-24 power supply, the third of six
Support the project on Patreon


Mean Well EPS-65S-24 power supply, the third of six

$9.97
Go to store
So, today I have the third power supply out of the six sent to me by one of my readers. This time, it’s a relatively compact, inexpensive power supply. I’d even call it a budget model, since in some ways it’s noticeably simpler than the previous ones. But simpler doesn’t always mean worse, and I’ll try to show and explain that in this review.
In any case, if you’re interested in name-brand power supplies, I think you’ll find something new here.


To be fair, this is only technically the third review, since I’ve already reviewed a power supply from this manufacturer. But these reviews cover a batch of six units, so this one is called the third of six :)
Once again, thanks to Vladimir, without whom this series wouldn’t have happened. I also realized that there are plenty of reviews of “Chinese” power supplies, but there’s a gap when it comes to name-brand models. And they’re interesting not only for their performance, but also for the circuit designs that set them apart from their Chinese “colleagues.” That’s why I think reviews of name-brand power supplies are useful too, and I’m glad I can help fill that gap.

This power supply was bought on Taobao. The previous two came from the same seller, but this one came from a different one.
The packaging is basic. It’s certainly not as good as a cardboard box, but better than bubble wrap.


The power supply looks great on the outside, with no damage. Then again, the manufacturer pours so much potting compound into these units that nothing inside can move, and breaking anything off would be quite difficult too.



Key features include high efficiency—up to 91%, depending on the output voltage—and a no-load power consumption of just 0.1 W.
It also has overload, short-circuit, and overvoltage protection.
The power supply is designed for passive cooling and operation at temperatures from -30 to +70.


The full specifications are available in the datasheet; I’ll just include a few key pages here.
According to the description, the power supply is available with seven different output voltages. I’ll be reviewing the 24 V version.
It’s also worth noting that the power supply delivers its rated power only in versions with output voltages of 12 V and above. At lower voltages, the power rating can be as much as half the stated value—for example, in the 3.3 V version.


This power supply is a variant of the EPS-65, which has a somewhat different design.


The power supply uses a 2 × 3-inch form factor, or about 50 × 75 mm, and belongs to the compact series.


For comparison, I’ve placed a couple of Mean Well power supplies rated at 100 and 120 W on the left, and a Chinese model rated at 72–96 W on the right—the so-called “people’s choice.”


Here’s a drawing with the overall dimensions of the power supplies in this series. As you can see, this unit is very compact; its total height is just 27 mm, for example.


But let’s take a closer look.
The input has an almost complete filter; the only thing missing, I think, is a protective varistor.


As expected, there’s a thermistor to limit inrush current and an X-class capacitor. There are two common-mode chokes: a small one right after the power input, and a larger one between the X capacitor and the bridge rectifier.
The input capacitor has a label with its specifications. Its capacitance is 82 µF. Given the 65 W output power and the full input voltage range, I’d say that’s not much.


I thought I’d seen all kinds of power supplies, but they still manage to surprise me. This time I came across a very unusual inductor. It’s wound not with ordinary wire, but with a thin strip of metal—a rather unusual solution.
Sure, I’ve seen transformers with windings made from metal strip or thick foil, but here the strip is laid perpendicular to the bobbin...


1. The GBP408 bridge rectifier and 13N60M2 high-voltage transistor are mounted on a small heatsink, which seems a little undersized for the power supply’s rating.
2. The output rectifier is an MBR20200, rated at 200 V and 20 A, which is quite good for an output current of 2.7 A. To the right, you can see a footprint for another rectifier, presumably fitted in the low-voltage versions of the power supply.
3. There’s also a pair of feedback optocouplers: one handles voltage regulation, the other overvoltage protection.
4. Another unusual thing I noticed in this power supply is that it has not one or even two Y capacitors, but four.
The lower pair is connected in the standard way, between the negative terminal of the input filter capacitor and the output common. The other pair, however, has one end connected between the EMI chokes, while the other end is also connected to the output common.
It’s a somewhat unusual arrangement; I haven’t come across it before.


Above, I mentioned that it uses a 13N60M2 transistor. It’s a fairly ordinary high-voltage transistor, except for the package it comes in. This one has the standard insulated TO-220 package, but it turns out there’s also a shortened version, which I hadn’t seen before either.


The designers surprised me once again here, but this time not in a good way. The power supply output uses a diode assembly, and it makes perfect sense that it’s mounted on a heatsink. Since this power supply is also available in versions with higher output current, the designers fitted a pretty substantial heatsink, which is good.
But I was very surprised to see that it not only extends into the “hot” side, but is insulated only with tape—especially given how meticulous this company is about the safety of its products.
The heatsink is secured with a special sealant and is mounted very firmly, but in my opinion this solution is more like something you’d find in a “Chinese” knockoff than in a branded product.


1. After the diode assembly there’s a pair of 390 µF, 35 V capacitors made by Rubycon, which was to be expected. There’s also an inductor to reduce output ripple.
2. Further to the left is another 100 µF capacitor, connected after the inductor, as well as an LED and a trimmer for adjusting the output voltage.


A couple more top-down photos of the power supply.



The underside is neat, though I wasn’t expecting anything else. The board is clean, with protective slots in the “narrow” areas.


1. A 46BL065 PWM controller manages the power supply. It’s a quasi-resonant PWM controller, but interestingly, despite having a different name, its pinout is identical to that of the NCP1239 PWM controller used in the RPS-120-27 power supply. The NCP1239, however, is not quasi-resonant.
2. Here’s another thing about this power supply that surprised me.
PWM controllers often monitor the input voltage to protect against operation when it’s too low. Usually, though, they measure the voltage across the input filter capacitor. Here, the AC voltage is monitored before the bridge rectifier. To do this, the board has two diodes with their cathodes connected together and routed to the PWM controller.
That would be fine, except this feature causes at least two problems:
1. If the input voltage disappears, the PWM controller immediately shuts down instead of waiting for the capacitor to discharge. As a result, this power supply handles brief interruptions in the input power less well. In practice, everything is left “up to” the output capacitors.
2. Since the PWM controller shuts down immediately and nothing discharges the capacitor, it remains charged even after the power is disconnected. In fact, even after 12 hours, it still has a residual voltage of 250–260 V!
I found this out by accident when I decided to discharge the capacitor before measuring its capacitance. The result was somewhat… unexpected.

There’s a third problem, too. The input X capacitor also has nothing to discharge it, so if you unplug the power cord and immediately touch its prongs, you might get a little “zap.”

All in all, the designers surprised me yet again. The features mentioned above don’t affect how well it works, but in my opinion they’re unpleasant.


One more small observation: the capacitor after the output inductor is connected somewhat incorrectly, since there’s a fairly long trace in series with it. Ideally, it should be closer to the output connector.
But the tests showed that everything worked fine, although this can usually increase ripple amplitude and noise levels a little.


The power supply’s manual includes a block diagram, which differs from the one for the RPS-120-27 mainly in that it lacks a thermal protection circuit.
Yes, this power supply has no overtemperature protection, which is also unusual for Mean Well products.


Since this power supply is relatively simple, I drew up its complete schematic. There may be a few errors, of course, but I took care to copy it accurately, right down to the component reference designators.
At a glance, it looks like a perfectly ordinary power supply circuit, if it weren’t for one small detail. Not only is there no output load resistor here—which, admittedly, modern power supplies don’t really need—but the resistor in series with the power indicator LED is a whopping 680 kΩ, whereas 2–10 kΩ is more typical. Apparently, it’s another attempt to reduce no-load power consumption.


Before moving on to the tests, here’s a photo to show what’s located where in this power supply.


The output voltage adjustment range is quite impressive. With the nominal voltage set to 24 V, the output can be adjusted from 21.1 to 28.4 V, with smooth adjustment throughout.


For the first test, I’ll check how the output voltage varies with load current, and measure the power supply’s efficiency and power factor.
I used one of my electronic loads, a wattmeter, and two multimeters to measure current and voltage simultaneously.
The test was run at both 225–230 V and around 108–110 V.


The results suggest that the power supply is not only highly efficient, but also has pretty good output voltage regulation. The output voltage varies by only about 30 mV across the entire load range.
The efficiency measurements at 110 V may not be very accurate, since the wattmeter doesn’t work quite correctly in this mode.


Next, I’ll measure the output ripple level.


First, here’s what the output looks like with no load. There’s a small low-frequency sawtooth because the controller is operating in “green mode.” Unlike the RPS-120-27, though, this power supply runs quietly; the noise is noticeable only at currents between 50 and 150 mA.


I’ll hide most of the oscilloscope traces under a spoiler, since they’re included just as supplementary material.


Combined view of the high- and low-frequency ripple at input voltages of 230/110 V and 100% load.
1, 2 - 230 V
3, 4 - 110 V

The ripple amplitude is low, especially compared with what the power supply's description claimed.
For the 24 V model, the specified ripple amplitude is no more than 240 mV. In reality, the result was about four times lower.


As usual, I'll finish the review with the thermal test results. The tests followed the usual routine: 20 minutes each, followed by a temperature measurement and an increase in load current. I covered the power supply with its lid to simulate an enclosed space.


A brief aside here. The manual specifies fairly strict limits on the power supply's operating conditions. For example, if the ambient temperature exceeds 50°C, the load current must be reduced. Similar limits also apply when the input voltage drops. At my test voltage of 108–110 V, for instance, the output power should be reduced to 90% of the maximum continuous rating—in this case, to about 60 W.


In fact, I didn't reduce the load power. What's more, at 230 V input I loaded it to 3 A, rather than the 2.7 A specified by the manufacturer. In my opinion, the power supply handled all of this just fine, only starting to give up at the very limits of its operating range.
At 230 V, the high-voltage transistor was the first component to overheat, though it was still far from a critical temperature.
When the voltage was reduced to 110 V, the diode bridge predictably heated up more because the current through it increased.
The other components stayed relatively cool, so I think increasing the heatsink area in the "hot" section could allow for higher output power without overheating.


Thermal image after the test at 230 V


And after the test at 110 V.


As last time, I ran a DC test to determine the maximum output current.
1, 2. With an input voltage of 230 V, the power supply shut down at 3.878 A.
3, 4. At 110 V, it shut down slightly earlier, at 3.569 A.

In both cases, the power supply shut down abruptly, and the output voltage remained at its nominal 24 V right up to shutdown, as you can see in the first and third photos.
Once the overload was removed, the power supply resumed operation automatically.


Just for fun, I made a GIF showing the short-circuit test.


Video review


Time to sum up the test results.
The power supply passed all the tests, delivering good output voltage regulation, low temperatures, low ripple, and high efficiency.
Overall, I don't have much to complain about in terms of performance, but I do have a few comments about its "features."
First, there are no bleeder resistors across the input capacitor. This doesn't affect operation, but it does affect safety.
Second, the heatsink for the output diode assembly is poorly designed, which is also a safety concern.
And third, there's no thermal protection. That's not ideal either, even though the power supply doesn't overheat during operation.

Personally, I think this power supply would be a great fit for compact soldering stations built around T12 soldering irons.

That's all from me. I'll be glad to answer any questions or read your comments. I hope you found the review useful.
$9.97
Go to store
This page was found when searching for:
meanwell тест, блок питания 3-24в с регулировкой обзор, 1239d065, 2. Блоки питания - Mean Well EPS-65S-24 (65 Ватт, 24 Вольта), 3, бп mean well 24в 6а, mean well блок питания 24v схема электрическая принципиальная, 5, rosleds блок питания 24в что внутри, для импульсного бп микросхема 46bl065 характеристики, блоки питания mean well 24v, 46bl065 купить, сколько потребляет блок питания без нагрузки mean well, что находится в блоке питания телефона, https://www.kirich.blog/obzory/bloki-pitaniya/564-blok-pitaniya-mean-well-eps-65s-24-tretiy-iz-shesti, eps65-24 схема принципиальная, блок питания 24в 65 вт, mean well ac/dc схема принципиальная, блок питания 24 mean well отзывы, plp 60 24 причина поломки, блок питания mean well в закрытом пространстве, лёгкий способ найти неисправности в электронике, ncp1239, dc-dc meanwell обзор, mean wel блок питания 24v регулировка напряжения, блок питания Mean Well, блок питания на 24 Вольта, тест блока питания, технические обзоры, схема блока питания

Questions and answers

What maximum output current did the Mean Well EPS-65S-24 reach?

It shut down at 3.878 A with 230 V input and at 3.569 A with 110 V input. It resumed operation automatically after the overload was removed.

What shortcomings did the author find in the Mean Well EPS-65S-24?

The author noted the lack of a discharge circuit across the input capacitor and the lack of thermal protection. He also questioned the insulation of the output diode assembly’s heatsink.

What does the author recommend the Mean Well EPS-65S-24 for?

The author considers it suitable for compact soldering stations using T12 soldering irons.

The figures from this review are in the measurement database: Power supplies

Calculate manuallyAll calculators

Discounted products


You may be interested in


Discounted products


Support the project

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!

Comments20

Comments are published in their original language.

  1. PavelKS
    9 May 2019 22:43#1
    Здравтвуйте.
    Спасибо за обзор.
    Ещё про силовой MOSFET можно дополнить. Если сравнивать с народным 24V блоком, в котором установлен 20N60, то 13N60M2  существенно быстрее. А если 13NM60, он ещё немного быстрее. Меньше потери на переключение, меньше нагрев.
    0
  2. PavelKS
    15 May 2019 12:50#2
    И ещё, в обзорах нигде у Вас не встречал, в мощных блоках питания во входных фильтрах ставят вместо разрядного резистора на 1МОм специальную микросхему: TEA1708T или HF81.

    ...И ещё в некоторых блоках питания сегодня уже на входе фильтра ~220VAC можно увидеть мелкую микросхему, в корпусе SO8 вместо высокоомного резистора. Это HF81, "зелёный" разрядник для X2 конденсаторов входного фильра. Похожая по функционалу микросхема TEA1708T, но у неё другая распиновка.
    0
    1. 15 May 2019 15:48#3
      Да вроде было что-то такое, в одном из Минвелов, но уже не припомню где именно. Как по мне, то вполне хватает и резистора, то уже "зеленые" не знают чем им заняться :)
      0
  3. 17 May 2019 11:47#4
    Добрый день,
    Подскажите можно ли на основе этого БП сделать зарядное устройство LIFePo4? Имеется 6-ть банок соеденены последовательно, BMS стоит с ограничением 5А. Напряжение зарядки 3,65*6=21,9В. Можно ли переделать БП, чтобы он работал в режиме стабилизации по току, а не по напряжению. У меня есть БП EPP-200-24 но там напряжение не уменьшить до 21,9 В, так же есть EPS-65-24. Или нкжно искать другой блок питания?

    P.S.
    Эти БП рассматриваю из за их размеров.
    0
    1. 17 May 2019 12:29#5
      В принципе переделать скорее всего можно, я так как-то давно переделывал один БП, это есть в моих первых обзорах. но не проще поставить DC-DC?
      0
      1. 17 May 2019 13:34#6
        DC-DC с ШИМ? Просто есть один простенький ШИМ, который хорошо ограничивает мощность, но он не успевает её ограничить в момент старта и БП EPS-65-24 уходит в защиту.
        Нашел такой PLP-60-24. Буду порбывать.
        0
        1. 17 May 2019 23:17#7
          Вообще такого не должно быть, у вас DC-DC с регулировкой тока?
          0
          1. 18 May 2019 10:34#8
            Есть ограничитель мощности https://masterkit.ru/shop/1320410 у него правда частота всего 500 Гц.
            0
            1. 18 May 2019 10:44#9
              Вам он не нужен, надо плату DC-DC которая умеет регулировать не только выходное напряжение, а и ток.
              Если Ваша плата умеет только первое, то конечно блок питания может уйти в защиту.
              0
              1. 18 May 2019 11:07#10
                Только что попробывал  https://mcustore.ru/store/moduli-pitaniya/preobrazovatel-dc-dc-xl4005e1-5
                a-s-regulirovkoj-toka/?gclid=Cj0KCQjw2v7mBRC1ARIsAAiw349MLtytcQSkdbe4odtRFhZxqK
                oYXslGF76yHffSHxj8rihYgd_IHDEaAibnEALw_wcB
                БП в защиту не уходит, ток ограничил на 4 А( на выходе с DC-DC), но греется ощутимо.

                Может всетаки приобрести PLP-60-24, а уже на выходе поставить DC-DC для подстройки конечного напряжения заряда АКБ? Или на самом PLP-60-24 найти место для подстройки конечного напряжения на 22 В.
                0
                1. 18 May 2019 11:49#11
                  А почему Вы просто не хотите взять модуль DPS, под нужное напряжение и ток ?
                  Сложные варианты делаете. 
                  0
  4. 18 May 2019 12:43#12
    Ограничен пространством корпуса где должны располагаться шесть банок LIFePo4 АКБ и зарядное устройство. Вот и ломаю голову над компактным зарядником пусть даже не совсем правильным с точки зрения процесса заряда.
    0
    1. 18 May 2019 13:12#13
      Маленько не понятно, что Вы хотите разместить в корпусе ?
      БП с зарядным устройством будет отдельно ?  В корпусе будет только 6 банок LIFePo4 АКБ с защитой ?
      Или Вы хотите всё в одном корпусе ?

      0
      1. 18 May 2019 14:18#14
        Всё в одном, что бы не таскать несколько разных коробочек. Нашёл не по далёку DC-DC  с возможностью регулировки тока и напруги, но более мощный чем есть. Завтра куплю попробую разместить.
        0
  5. Artem Puponin
    9 September 2019 12:28#15
    Спасибо за обзор. Удивило что данный БП не имеет заземления ни по входу ни по выходу.  Стоит ли его добавить  через Y конденсаторы? Питаю устройство для пром применения от этого блока питания. Блок питания на заземлен на корпус. Корпус устройства заземлен.  Если заземляю на корпус минус то получаю очень нестабильную работу устройства. Вот и думаю заземлить бп может стоит по описанию ваших видео. Минус заземлить мягко через конденсатор?

    0
  6. 11 December 2020 17:57#16
    Спасибо за обзор. Встретился мне MeanWell rs-35-24. Нагрузка 400-500 ма. 24 вольта стабильны. Фильтр 2 по 390, индуктивность, 100.
    Но пульсации огромные +- 1 вольт. С частотой 66кГц. Заменил конденсаторы в выходном фильтре поставил 3 по 470, индуктивность, 100. Пульсации уменьшились до +- 300 мВ. Подскажите пожалуйста возможную причину неисправности. Снабберная цепь?
    0
    1. 11 December 2020 20:40#17
      Вы измерения, на осциллографе, правильно делаете ? Очень похоже, что проблема с измерениями размаха пульсаций.
      Щупы осциллографа сделайте как у Kirich(а). ЗДЕСЬ есть про щупы.
      Для нагрузки возьмите просто резистор. 
      0
  7. 13 December 2020 12:51#18
    Думаю правильно. Кроме дефектного на столе лежит БП от ноута. На нем и без навесных емкостей  все хорошо. А вот у MeanWell не установлен Y конденсатор между обмотками. И снаббер вызывает сомнения 34кОм и 10 nF. Нагрузка 100 ом.
    0
  8. 14 December 2020 15:49#19
    Проверил. Да, это конденсатор подавления помех. Пульсации стали 80 mV. БП много. Какая-то партия кривая. Оборудование проработало 7 лет).
    0
    1. 20 December 2020 21:29#20
      Странно что нет Y конденсатора, не встречал такого, из-за этого и пульсации могут быть завышены. Можете сделать фото?
      0

Discuss on the forumHave a question or some experience to share? Discuss the article with the author and readers: the forum is a convenient place to share schematics and photos.
Go to the forum

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