12 V 0.5 (1) A Power Supplies

Many readers know how much I enjoy writing reviews of power supplies. As it happens, I recently got my hands on quite a few of these devices. A well-known store started selling various power supplies pulled from equipment not long ago, and today I’ll tell you about one of them.


Last year, I wrote in the comments that reviews of various power supplies were coming soon, and these are exactly the ones I had in mind. I ordered several types: three small used ones and one new, fairly powerful model. I’ll cover them in order of size, starting with the smallest.
Since I use power supplies often, I ordered a lot of three, though lots of 1, 5, and 10 are also available. This power supply won’t be an exception: I’ll use it in one of the reviews I’m planning to put together fairly soon.

The power supplies come in separate large bags, not all three in one bag as I initially thought. In other words, the warehouse simply marks how many items to put in the basket.
I have no complaints about the packaging; everything was thoroughly wrapped in polyethylene foam.


In the title, I wrote 0.5 (1) A for the current. I’ll explain what that means as we go through the review.
The product page listed 12 V, 1 A, which is clear enough. It also said the power supplies were disassembled, meaning they aren’t new but have been pulled out of some equipment. In my experience, these used power supplies often have better build quality and circuitry than new ones.


The power supplies are quite compact, with actual dimensions of about 57 × 35 × 19 mm.


The board layout is quite dense and partly covered with silicone, which I had to cut away in a few places.
Since the board is used, you can see where the wires have been cut.


The boards have different colors of phenolic laminate and were made at different times, but all three date from 2007–2008.


The boards also have the model number marked on them: 3A-064WU12. I used it to find their actual specifications.
12 V, 0.5 A, 6 W, 74% efficiency at 115 V. The manufacturer’s name is also listed there: Eng Electric Co., LTD. So these are genuine branded power supplies.


The product page also mentions a current of 0.5 A, but only in passing. I think the idea was that 0.5 A is the rated current and 1.0 A is for short-term use. In any case, these specifications should be listed in the specifications section, not in the product title.


All right, back to our power supplies.
1. The input has a 1 A fuse. It’s a slow-blow fuse (T stands for Träge, German for slow), which is needed because of the pulsed current when the power supply is switched on.
2. There’s also a 7 mm varistor at the input, rated for a peak voltage of 470 V. Next to it is an X-class interference suppression capacitor with a capacitance of 0.1 µF.
3. Next come the common-mode choke and the bridge rectifier.
4. The primary and secondary sides are connected through a 2.2 nF Y-class capacitor.
Overall, I could give the filter five stars, if not for two shortcomings:
1. There’s no thermistor, though perhaps one isn’t really needed here, since the input capacitors don’t have a particularly high capacitance.
2. There’s no bleeder resistor across the X-class capacitor. Without one, the power supply can give you a little shock if you unplug it and immediately touch the plug contacts.

Still, credit to the manufacturer for including an interference suppression filter and a varistor.


1. Two 6.8 µF capacitors are installed at the power supply input, for a total capacitance of 13.6 µF, which is perfectly reasonable for the rated power of 6 W.
2. But the capacitors aren’t simply connected in parallel; there’s also an inductor between them. The color bands aren’t visible in the photo: brown-black-red-gold.
3. The power supply is controlled by the well-known VIPer-12A PWM controller.
4. Next to the controller is its supply filter capacitor. These capacitors can often fail without any obvious signs and cause a lot of trouble, since they still look fine from the outside. If you have a used power supply, I recommend replacing them first.


The silicone covering the board has a slight yellow tint. At first I thought it was caused by heat from the components, but it’s the same color even near components that don’t get hot.


As I mentioned above, it uses a VIPer-series PWM controller. This is a family of integrated PWM controllers. The chip package contains not only the PWM controller itself, but also a high-voltage transistor and protection circuits for overload, overheating, and overvoltage.
I usually use similar controllers from another, equally well-known company, Power Integrations. I like them better. But by and large, they are very similar in many ways.
The specified power rating for the DIP-8 package is 13 W over a narrow input-voltage range (230 V) and 8 W over a wide range (115–230 V). Since this power supply is rated for 115–230 V, its actual power is therefore up to 8 W.


The block diagram shows the output transistor and the protection circuits. I could explain all of this in more detail, but in my opinion it would be better suited to a separate article.


The secondary side of the power supply contains:
1. A 2 A Schottky output diode, which again indicates that the maximum output current is no more than 650–700 mA. One of the diode leads has a ferrite bead.
2. There are two output capacitors, 470 and 220 µF, both from Samxon, as are the input capacitors. I wouldn't call them high-end capacitors; they're more mid-range. They were originally OEM parts from Matsushita, sold under their own brand. What disappointed me personally is that they're rated for 16 V rather than 25 V, as they should be for this voltage.
3. There is a space between the capacitors for an inductor to reduce ripple, but a jumper is installed instead.
4. The feedback circuit is standard: an AZ431 adjustable shunt regulator (equivalent to the TL431) and an EL817 optocoupler (equivalent to the PC817).

There are two things I don't like about the output circuit:
1. No output inductor.
2. 16 V capacitors instead of 25 V ones.

Otherwise, it's put together quite well.


The soldering quality is perfectly acceptable. The remaining components are on the underside, along with a pair of Zener diodes that I'll discuss below.
The clearance between the high-voltage and low-voltage sides is quite adequate. There are no isolation slots, but since the power supply was originally designed to be installed in an enclosed case, this is acceptable.


The power supply circuit is fairly standard and is essentially based on the PWM controller's datasheet. One useful safety feature for protecting the load is a pair of Zener diodes.
ZD1 is rated at 14 V and connected across the output. Its purpose is to prevent the output voltage from rising above 14–14.5 V.
ZD2 is rated at 16 V and connected across the optocoupler transistor. Its purpose is to limit the output voltage if the feedback circuit is interrupted or fails.



I've been told several times in the comments that my approach to measuring ripple levels isn't quite right. I've taken that on board and will try to do it more correctly this time and in future tests.

The thing is, when taking measurements I usually connect the probe using the “wrong” method because it's more convenient. In this setup, the probe's ground lead acts partly like an antenna, picking up interference and distorting the waveform. This method is fine for a general assessment, but it is indeed technically incorrect.
The image below is from the description of the power supply testing procedure.


To capture waveforms correctly, the probe should be connected directly to the power supply output, without long leads.


As you can see in the photo, in addition to the ground lead with an alligator clip, the oscilloscope probe can also be connected right next to the probe tip.
Using “sticks and string,” I made a makeshift version of a dedicated probe for testing power supplies. The most awkward part was connecting to the center contact, since it's conical.
Two capacitors are connected in parallel across the input: a 1 µF, 63 V electrolytic and a 0.1 µF ceramic capacitor.


Of course, what I showed above could be called a hack, but even fairly well-known companies (Power Integrations, for example) aren't above doing something similar. They use a connector for it, though, and I didn't have one :(.
Photo from the application notes for Power Integrations' TOP-series PWM controllers; the component values are taken from there as well.


The oscilloscope probe was connected directly to the power supply's output terminals, and the load was connected to an extra wire I had soldered on.
While preparing the test, I compared the waveform with and without a load connected, under no-load conditions. There was no difference.


The first thing that surprised me when I switched it on was the output voltage: 12 V, accurate to at least two decimal places. Strictly speaking, it doesn't matter; even if the voltage had been in the 11.5–12.5 V range, I would have considered that normal. Still, it's nice to see.
1. No load.
2. 0.25 A
3. 0.5 A
4. 0.75 A
5. 1 A
6. 1.2 A.

As you can see, the output voltage only started to drop when the load current exceeded 0.75 A, which is one and a half times the rated value. Before that, the voltage remained very stable, dropping by about 0.001 V for every 0.25 A of load.


I wouldn't call the ripple level low: at the rated current of 0.5 A, it was 100 mV, but even under overload it stayed below 140 mV.


The tests showed that the maximum current at which the power supply can maintain a stable output voltage is 0.9 A. That's with an older PSU and at almost twice its rated output current.


I've also been told that it's wrong to test power supplies using an electronic load. In this case, I disagree with that conclusion, since in linear mode the load's MOSFETs are essentially just resistors with feedback.
In any case, as an experiment, I compared the power supply's behavior under load from a regular 10 Ω resistor (it was what I had on hand). You can see in the photo that the load's positive lead isn't connected.
The voltage dropped, of course, since the current was clearly above the calculated value.


On the left is the waveform under a 1 A load from the electronic load; on the right, 1.08 A with a resistor as the load.
I wouldn't say there's any major difference.


Next up: a temperature test. I covered the power supply with an improvised "case" and applied loads in sequence, from 0.25 A to 0.9 A. I chose 0.9 A because the PSU can still maintain its output voltage reasonably well at that current. Each test lasted 20 minutes, for a total test time of 1 hour 20 minutes.


I put all the data into a table and added a new column showing the voltage at the start of the test (V1) and at the end (V2). This makes it possible to track voltage drift as the unit heats up.
At first, the voltage may look less stable than in the test above, but there I connected directly to the PSU terminals, whereas here I used a piece of wire, which accounts for the difference. Still, I can say that the output voltage is practically unaffected by temperature.
What I did find was that at a load current of 0.9 A, the PSU reduced its output voltage after about 5–7 minutes.


After the test, the components reached a maximum temperature of about 100 degrees for the transformer and 118 for the PWM controller. At currents up to 0.75 A (1.5 times the rated value), there is no overheating.


This is what the output power limiting looked like. I repeated the test with the PSU already warmed up to make the effect clearer.
At startup, then after 6 minutes the voltage gradually began to drop. At the 20-minute mark, I removed the cover and the voltage slowly started to rise. About 15 minutes later, I had to blow on the board several times, and the voltage quickly returned to normal.


Above, I complained about the lack of an output inductor, so I decided to compare the two setups and see how the results changed.
I used a small homemade inductor, made from whatever I had on hand. It's small and wound with 0.68 mm wire.


The results speak for themselves.
1, 2. Current: 0.5 A; without the inductor on the left, with it on the right.
3, 4. Current: 1.0 A.

A word of warning: the inductor shouldn't have too high an inductance. Increasing the inductance will cause the ripple on the first filter capacitor to rise sharply, which is bad for both the capacitor itself and the zener diode connected in parallel with it. You'll need to replace the capacitor with a similar one rated for 25 V, and move the zener diode to the PSU output.


That's it. In short, although these power supplies have some of the shortcomings mentioned in the review, they're pretty good overall and can be used in various DIY projects that don't require much power (6–8 W). They're genuine branded units and reasonably well made.
Buying them individually costs more, so if you're going to buy some, get a lot of 3 or 5.

I hope you found the review useful. As always, I'll be happy to answer questions in the comments.