A Not-So-Great USB Charger (Power Supply)

I've been meaning to write about this charger for a long time, but never got around to it, even though there's plenty to look at.
I got it from a fairly well-known store, which took it off sale after reading my review. In my opinion, they did the right thing. That's why I'm not linking to the product. You may come across it in other stores, so I still think this review will be useful.

I received this charger (though power supply would be more accurate) in a regular resealable bag, with no blister pack or box.


I wouldn't call it particularly small. I've seen much larger models with a rated current that's not much lower.


The rated output current is 3000 mA, which is pretty good for most uses. For example, you could charge a tablet and a smartphone at the same time.
The charger has two output ports, labeled iPad and Galaxy—or, if you prefer, Apple and Samsung devices.
There's an indicator LED on top. It stays lit regardless of the operating mode.


But since there's usually nothing interesting to me on the outside, I decided to open it up, of course. It's fairly easy: pry open the small gap between the two halves of the case, then use a screwdriver to separate them. The power supply is glued shut, but it came apart without much trouble.


At first glance, it looks fairly neat and didn't raise any red flags, at least.



The circuit board is neatly made. The LED does sit right on top of the USB connectors, but they've covered them with insulating film for protection.


The board is soldered reasonably well too. There are a few minor flaws, but overall I'd give it a solid B. I took off one point for some rough soldering and the lack of safety slots in the PCB.


One thing that surprised me enough to take a separate photo was that the wires to the board have silicone insulation and can easily withstand the heat of a soldering iron tip. They're also very flexible. I wish I could buy this kind of wire separately, without the power supply.


Let's take a closer look at the board.
1. There are two input capacitors connected in parallel, with a total capacitance of about 10 µF. That's not enough for 15 W. There's no input filter, but at least there's a fuse :)
2. The IC is in a DIP package. I didn't look up its datasheet, but I remember seeing it somewhere before, and it was rated for the power supply's output. The diode bridge is quite unusual, though—I haven't come across one like this before.
3. The transformer isn't very large. The rated 15 W really is its maximum; there's no headroom :(
4. Despite that, the interwinding capacitor is the right type. There's also feedback through an optocoupler, which is something manufacturers sometimes cut corners on.
5. There are two output diodes connected in parallel. The output capacitor is only 1000 µF, which is a bit low for 3 A. There's no output filter either.
6. The feedback circuit, on the other hand, isn't very good; the cost-cutting is obvious. Instead of a proper circuit with a TL431, they simply used a Zener diode.


By the way, the input capacitor was split into two smaller ones for a reason: there's a small inductor hidden between them to reduce interference.


The IC has an external shunt for current measurement, which means it has at least short-circuit protection on the output—and that protection does work.
Voltage dividers are installed near the output connectors. They let the device being charged know how much current it can draw from the charger.


Just in case, and for general reference, I drew the schematic for this power supply. There's nothing new that sets it apart from other power supplies, except that I haven't seen one using a Zener diode instead of a dedicated IC to regulate the output voltage in quite a while.


The testing was pretty much standard for my reviews. I used:
Electronic load
Oscilloscope
Multimeter
Thermometer
Paper and pen.

1. First test, with no load: the output voltage is a little high. The limit is 5.25 V, though this is fairly common.
2. Second test: 1 A load current. Ripple increased noticeably, but the output voltage is well within spec.


1. Load current: 2 A. Ripple is about 0.7 V, which is way too much. I even had to switch the oscilloscope to 0.2 V per division instead of 0.1, as in the previous test.
2. Load current: 2.5 A. The ripple level is the same as in the previous test, and the output voltage is within spec.


The next step was going to be setting the current to 3 A, but the output diode temperature exceeded 100 degrees, so I stopped the test.
I put together a table based on the test results. The interval between measurements was 20 minutes, and the whole test took 1 hour.
As you can see from the table, the output diodes and capacitor reached quite high temperatures. I don't recommend running it like this for long, so I stopped the test.


People sometimes ask what causes power supplies to fail. Below are photos of two 5 V, 2 A power supplies. They failed about half an hour apart. The one in the middle came with a Teclast tablet. It had been working fine for several months, then suddenly burned out with a few minor fireworks while the tablet was switched on and charging. But I needed the tablet, so I took another charger off the shelf. It also passed the tests without any problems and worked fine (on the right). Half an hour later, the same thing happened, and I had to charge the tablet from a bench power supply.

Power supplies often fail because of:
1. The power transformer overheating, which causes the core's magnetic permeability to drop above its critical temperature.
2. The PWM controller itself malfunctioning, especially under overload or short-circuit conditions.
3. A loss of capacitor capacitance due to aging.


This power supply has been in service for over six months, but I had to make a few modifications. I soldered a small metal plate to the PWM controller to act as a heatsink, and drilled ventilation holes in the top and bottom of the case. It works fine this way, though I think it would also work without modifications if used at currents up to 2 A.


So, what can I say about this device? It feels like they started out trying to make something good, then suddenly ran out of money and decided to make it cheap. In other words, some parts are done properly, but there are obvious signs of cost-cutting. And the claimed 3 A output current is a bit optimistic; I wouldn't risk it and would keep the load to 2 A at most.

That's all. This was a short, but rather sad review.