Inexpensive 12.6 V 3 A Charger

I posted a couple of charger reviews just recently, but as it happened, another one landed in my hands by chance. Unfortunately, it’s also a 12.6 V charger (for a 3S lithium battery pack), but I decided the review might still be useful because of its low price. Sadly, things aren’t quite as I’d hoped, but more on that in the review.


I ordered 10 chargers. At the time, they cost $8.13 each. Maybe there was a sale, or maybe the seller has raised the price since then—I don’t know. To avoid any customs issues, I placed two separate orders.
Interestingly, the packaging was different. Looks like they used whatever boxes they had on hand, but everything was packed tightly.


In any case, everything arrived. Each charger was packed in its own cardboard box, with the cables packed separately.


The package includes the charger itself and a power cable.


One of the ten cables came with a plug that has flat pins, even though the order specified EU. Not a dealbreaker, but annoying.
The second issue is much more interesting. The listing says “Liitokala 12.6 V 3 A charger.” The 12.6 V and 3 A part is clear enough, but the Liitokala name raised a few questions. As far as I know, Liitokala doesn’t make chargers like this. Still, the chargers have Liitokala stickers on them—and oddly enough, some boxes had them and some didn’t. Though looking at the photos, you can see there’s no difference between the chargers themselves; the only difference is the sticker.


The case is the familiar black “brick” shape. There’s a power cable connector on one end and a cable for connecting to the device being charged on the other. The connector is 5.5/2.1 mm.
There’s an indicator LED at the cable end.


But I was interested in this charger not only on its own, but also in comparison with the one I reviewed earlier.
As a reminder, that charger has the same claimed specs—12.6 V, 3 A—and looks almost identical, though its case is slightly larger. Here’s a link to , so you know which one I mean.


The one under review is on the right; the one I took apart earlier is on the left. You can already see some differences here.


I bought these chargers for someone else, so before taking one apart I had to ask my friend if he was okay with me opening it up for the review, since the two halves of the case are glued together. He didn’t object, so I took it apart.


There are many more differences inside. For one thing, the previous model has a larger transformer core. It’s not very noticeable in the photo because the tape gets in the way. The heatsinks are also less well insulated—or rather, there’s a small piece of insulation only on the transistor heatsink.


And then there’s the input filter. The unit under review is on the right: a simpler bridge rectifier, no inductor, and a standard fuse.


The output side is a little better. Actually, it’s more accurate to say it’s not much different from the previous model: two capacitors, and no output inductor. And, just like the previous one, there’s a footprint for a second diode assembly.


Let’s take the board out of the case for a closer look. Even at first glance, I thought there were more differences.


1. The input diodes are 1N4007. There’s no filter, but the capacitor is rated at 82 µF. Even allowing for the fact that Chinese capacitors usually have less actual capacitance than marked, that’s still fine for a 35–40 W charger.
2. The transistor is an 8N65, which is perfectly adequate for this power level.
3. The EMI suppression capacitor is the correct type, so the main safety concerns are the lack of insulation on the heatsinks and the lack of protective slots in the board.
4. The output diode assembly is rated for 10 A and 100 V, which is fine for both voltage and current. The capacitors are 1000 µF, 25 V, and there’s not much to complain about there either, apart from their “no-name” status.


Surprisingly, the board is soldered quite neatly. It’s nowhere near the quality of a name-brand unit, of course, but overall it’s fine.
There are no protective slots, but the clearance between the “hot” and “cold” sides is pretty good.


The primary side of the power supply, in case anyone needs to repair a similar charger.


And here’s the first blunder. To be honest, I’m not even sure what to call what I found.
You can see a yellow Class X EMI suppression capacitor on top of the board, but it isn’t actually connected to the circuit. Well, sometimes they use jumpers instead of an inductor; I got used to that a long time ago. But soldering in a capacitor and then not using it?
In the photo, I marked how the thermistor and fuse are soldered in. You can see that the capacitor (on the right) isn't connected to anything. Strange choice :)


As with the previous charger, I'm much more interested in the secondary side, since the primary side usually differs so little from one design to another that I can draw it from memory by now.
Like the previous chargers, this one is based on an LM358 op-amp; there are no “smart” controllers anywhere in sight.


The entire circuit consists of a 6853K09 PWM controller, connected just like the 63D39 and 63D12 controllers, which are all very similar to [leech=https://docs.google.com/viewer?url=https%3A%2F%2Fwww.kirich.blog%2Fonline%2F290154802700162501500662321]FAN6862. There's also an LM358 op-amp—the classic choice for cheap chargers.


I redrew the schematic, although in this case it's essentially a combination of the power supply schematic and the previous 12.6 V, 1 A charger I covered earlier, with a few differences.
The component reference designators match the schematic, at least in most cases :)

The output section is very similar to the one in that charger's schematic, and in some ways this circuit is even simpler. In any case, both circuits are much simpler than the previous 3 A charger. That one had a dual power supply, and with a little work you could get its power consumption down to almost zero when it wasn't connected to the mains.


The output section is also very basic: blue is voltage regulation, red is current regulation, blue is the indicator, and green is the reference voltage.
This is one of the simplest charger designs. The only simpler options are based on an LM317 or a resistor, but the latter isn't used with lithium batteries (at least, it's very rare).


Here are the first tests, following my charger testing method.
1. The no-load output voltage is noticeably too high, by about 40 mV per cell. That means each cell will be charged to 4.24 V instead of 4.20 V. This makes it more likely that the battery pack's protection board will trip. The previous version was 20 mV over.
2. The charger's own current draw with the mains disconnected is 11 mA, compared with 7 mA for the previous version and 14 mA for the 1 A version. The current draw of the previous 3 A version can be reduced significantly; that's much harder to do with the one under review, though still possible.
3. The charging current is 3.23 A, almost 10% above the rated value. That's not really a problem; the batteries will simply charge a little faster. But in my case, the higher current came back to bite me.
4. The indicator switches from red to green at 359 mA, a little above the standard 1/10 of the initial current. Not a critical issue.
5, 6. The charging current 5 and 10 minutes after the indicator switches. As the schematic suggests, this charger can't disconnect the batteries when charging is complete, so it keeps supplying current. That's not an issue if you typically charge and then disconnect the batteries, but I wouldn't leave them connected for a week.


Next, a load test. As always, we'll check two things:
1. Heating.
2. How much the voltage drifts after warm-up.

For this test, the electronic load is connected before the shunt so the charger doesn't switch to constant-current mode (although the red indicator still lit up in the end). The load current is set to the value measured in the previous test.

After half an hour of warm-up, the voltage has drifted noticeably from its initial value. Of course, heating drops once charging is complete, but first the charger will bring the battery voltage up to 12.7 V, then, after cooling, it will drop to 12.68 V. Wait, why would it drop? With no load, the output was 12.72 V, so it will probably rise instead. Too bad there's no trimmer to adjust it.


The graph shows the voltage drifting as the charger heats up. The previous 3 A charger drifted by just 0.005 V! As they say, compare the difference.


The thermal performance isn't great either. First, here's the temperature of the case and components after half an hour of warm-up.


And now after 1 hour and 14 minutes. The highest temperature was recorded near the transformer winding: over 100 degrees.
I wouldn't say it's all that bad, since the charger usually runs for an hour or two, three at most. By then, the battery is usually charged and the temperature drops. Also, heating will be a little lower at the start, since the charger's output power is lower. For example, at 3.8 V per battery, the total is 3.8 × 3 × 3.2 = 36.5 W, and near the very end of charging (this is the mode I used for the test), it’s 4.2 × 3 × 3.2 = 40.3 W, 10 percent more.

Temperature of individual components at the end of the test:
Input bridge rectifier - 74.5
High-voltage transistor - 86.3
Transformer - 94.8
Transformer winding - 102.8
Output diode assembly - 99.9
Output capacitors - 82.4

Thermal images from two angles.


In my opinion, the overheating problem comes down to a few things. First, the transformer has too little power headroom. Second, the output current is too high—almost 10% is a lot. I think it should be reduced at least to the specified value, and ideally down to 2.8 A. It should work fine that way.
As I did last time (in my review of the 1 A charger), I recommend changing the resistor divider values. In this case, either increase R20 or decrease R22. Since decreasing it is easier than increasing it, that’s the better option—for example, by soldering an 8.2–10 kΩ resistor in parallel. The lower the resistor’s resistance, the lower the output current.


The conclusions are simple. The main advantage of this charger is its price; I haven’t found a cheaper one yet. As you can imagine, the price usually reflects the quality of the assembly and performance. In this case, the manufacturer has clearly cut corners on almost everything. Still, the charger works, though I’d recommend making a few modifications. The modification itself is simple; the hardest part is opening it carefully.
In any case, these products have about as much to do with Liitokala as I do with ballet :)

That’s all. I hope you found the review useful. As always, I look forward to your comments and questions.