Modiary MDA238 42 V 2 A charger for 10S lithium battery packs

I’ve reviewed lithium battery chargers several times. They were for 3S packs, but the bigger issue was that their circuitry was far too simplified. After that, I was repeatedly asked which charger to buy for better battery safety, without having to “finish” it myself—even if it cost more.
So I ordered a charger that supposedly belongs to a more properly designed type. As always, I’ll get into the details in the review.

Previous reviews
12.6 V 3 A
12.6 V 3 A
12.6 V 1 A

One of the key safety features when charging lithium batteries and battery packs is for the charger to shut off when charging is complete. This is often where manufacturers cut corners: once the charger switches to CV (Constant Voltage) mode, it stays there. This is an extremely simple solution and is essentially just a regular power supply with current limiting for CC (Constant Current) mode. In fact, you could even build a charger like this yourself “with sticks and string.”

For the most part, this charging mode is safe if the process is monitored by the user: when the light turns green, you know it’s time to disconnect the charger. You can unplug it right away or an hour later, but if you don’t, the risk of battery failure increases—especially with worn-out batteries. I think many people have noticed how quickly the terminal voltage of old batteries drops after disconnecting the charger.

I’m no chemist, but I suspect the problem is dendrites, which can increase self-discharge and even cause an almost direct short circuit. This in turn heats up the cell, with the usual special effects. And the bigger the battery, the more spectacular the effects.

I ordered this charger almost “on a whim,” since I didn’t know whether it was properly designed or not—I only had a few guesses.
I also chose it based on its voltage: 10S packs are quite common. As for the power, I figured that if the 2 A model was designed properly, there was a good chance the manufacturer’s other models would use similar circuitry.

Specifications from the product page:
Input: 100–240 V AC
Output: 2 A (85 W), CC/CV, with a final voltage of 42 V + 0.2 V
DC plug: 5.5 mm / 2.1 mm (barrel plug); in fact, it’s a 2.5 mm connector.
Cable length: about 100 cm
For: 36 V lithium-ion bicycle batteries
GS-certified safety standards in accordance with § 21 ProdSG

The packaging is very basic.


The contents are just as simple: the charger, a cable, and a manual.


Though in this case, it’s not really a user manual, but a list of things you mustn’t do.
The cable is standard: an unearthed plug at one end and the familiar figure-eight connector at the other. It looks decent.


The charger looks like a laptop power supply. It measures 146 × 65 × 34 mm. I didn’t weigh it, but it feels quite hefty.


The bottom is covered with warning labels and specifications, along with marks indicating compliance with various standards. There’s also a very useful diagram showing the connector options and their pinouts.
Mine has the most common type of plug, like the one used on many power supplies, but versions with three- and four-pin plugs are also available.


The power supply’s case is clearly sturdy. The power input connector is on one side, the output cable on the other, with a status LED on top and four feet on the bottom.
The output cable has an unusual cross-sectional area by our standards: 0.824 mm². It’s visibly thick, too. The connector is listed as 5.5 × 2.1 mm, but as far as I can tell, its actual inner diameter is 2.5 mm. At any rate, it fits sockets of that size.


The LED isn’t very bright. It glows green when charging is complete or when no load is connected, and red while charging.


For testing, I connected the charger to an electronic load in CV mode, which in a way simulates a battery as the load.
The load’s measurement error is a bit high, but it’s good enough for an evaluation.

At 41.83 V, the charging current was 1.8–1.9 A. Raising the voltage slightly, to 42 V (4.2 V per cell), brought the current down to 0.8 A. Increasing the voltage further reduced the current, and at 42.14 V it dropped to zero, with the charger switching on its green LED.


When the voltage is lowered, the charger behaves a little differently.
At 25–30 V (2.5–3.0 V per cell), the charging current is still 1.8–1.9 A, but at 20 V it drops to 0.6 A. Lower the voltage further and the charger shuts off its output, although it continues running.
That is, If, when trying to charge a battery with a voltage below 18–20 V, the voltage does not rise above 2 V per cell as a result of the charging current, the protection will trip.

The protection is latched, so to reset it you need to unplug the charger, wait for the LED to go out, and plug it back in.


For more realistic tests, I used one of the batteries from a hoverboard. It’s a 10S2P pack with a capacity of 4.4 Ah, so this charger is a great fit for it.
But the battery was badly out of balance, so the first test didn’t go quite as planned.

It started as expected: the charging current was 1.9 A, which is actually less than the stated 2 A. After a while, the current began to drop gradually, and when it reached about 1.1–1.2 A, the charger shut off.
I then checked again, using one tester to log the voltage and the other as an ammeter. The same thing happened, and the last photo shows the battery voltage starting to drop slightly after the charger shut off.

It turned out that it wasn’t the charger shutting off, but the battery’s protection board, due to the imbalance. Though it’s not entirely clear why the voltage dropped in that case. The relay in the charger clicked, and everything appeared to be working normally.


I wasn’t happy with this, so I tried a second identical battery. It behaved much the same, so I decided to connect directly to the battery’s protection board connector—that is, upstream of the protection circuit. Before testing, though, I did balance it a little.

Warning: only do this if you know exactly what you’re doing. Otherwise, you could run into problems, including the battery catching fire!!!

After that, everything went smoothly. The charger charged the battery, and when the charging current dropped to 200–220 mA, the LED turned green, but charging continued. When the current fell to around 100 mA, the relay clicked and the charger shut off. The LED stayed green, but no current was flowing.


Next, I wanted to see how it would behave if I disconnected and reconnected the power, or disconnected and reconnected the battery.
It turned out that in both cases the charger behaved exactly the same—and not very gracefully.

So, the battery is almost fully charged. I unplug the charger, wait a few seconds, and plug it back in. The charger starts charging the battery at 0.31 A (an arbitrary value), and the current begins to drop. When it reached 50 mA, I simply unplugged the charger because there was no point waiting any longer: it hadn’t “armed.”

Further tests showed that for the charger to enter normal operating mode, the charging current needs to rise to around 500–700 mA (I may pin down the exact value later). Below that, it simply works like a regular charger, switching to CV mode without automatically shutting off. This is most likely related to the auto-start function, rather than a separate RUN button that you press to start charging. I once tried to do something similar in one of my reviews, and I can tell you that implementing this function properly isn’t easy. It would be much simpler if the device had a microcontroller, though.


I’d say I’ve figured out the charging mode, so I moved on to heat-up tests.
To run the test under conditions close to the maximum, I connected the charger to an electronic load instead of a battery and set the voltage close to the maximum, simulating operation at maximum output power.

1. About 5 minutes after startup, the thermal image showed some hot spots on the left, presumably around the transformer or output diode, and a slight warm-up on the right.
2. After half an hour, the case had warmed up to 55 °C.

Below are pairs of photos taken with a thermometer and a thermal camera. There’s a reason for showing both. A thermal camera and a thermometer measure temperature differently: the thermal camera “looks for” the hottest spot, while the thermometer averages the temperature across the entire case, so its reading will always be lower. Even if you use a “laser-guided” infrared thermometer, that doesn’t change anything—the spot is simply at the center of the measurement area, nothing more. I use a compact infrared thermometer for a reason: unlike larger models, it can measure the temperature of small objects, such as transistors and diodes.



1. After one hour of operation, the case temperature was 65 °C.
2. After two and a half hours, the case had warmed up to 71 °C at its hottest point, and to an average of 56 °C. Just a reminder: this is the maximum setting. In practice, it won't happen, since the voltage will either be lower or rise to 42 V, at which point the current will start to drop.

One more thing: some plastics are partially transparent to infrared, so the thermal camera can “see” through them.


Naturally, I took the charger apart. It came apart fairly easily: all it took was running a knife along the seam while tapping it with a small hammer.


The layout is similar to monitor and laptop power supplies: two heatsinks along the long sides of the board, with everything else inside.


1. The input bridge rectifier is glued to the heatsink with sealant. The power dissipation there is low, though, so it was probably just secured in place rather than mounted to dissipate heat.
2. The board is also secured with sealant near the output cable.


After cutting away the sealant, I removed the board from the case. It was held in there quite firmly even apart from the sealant.


A few words about the components and circuit design.
1. The input mains filter gets top marks: it has a filter made up of two inductors and X-class capacitors, a fuse, a thermistor, and even a varistor.
2. The input capacitor is 120 µF, which gives plenty of headroom for our mains voltage and is even suitable for operation across the full 100–240 V mains range. The amount of sealant poured over the board is surprising.
3. I couldn't make out the full marking on the high-voltage transistor. Judging by the suffix, it's rated for 600 V. The tight layout, traces of coating on the transistor, and—unexpectedly—an insulator got in the way. I haven't come across that kind of arrangement before. I didn't like that the current-sense resistor is right next to it and almost touching the heatsink.
4. To the left of the transformer are two Y capacitors connected in series for increased safety; to the right is a trimmer for setting the output voltage.
5. The output uses an MUR1040CA diode assembly, rated at 10 A, 400 V. To its right, inside the heat-shrink tubing, is the snubber resistor. The output disconnect relay is also visible here.
6. The output also has two inductors, one standard and one with two windings, plus 470 + 220 µF, 50 V capacitors. The transformer has two windings: a power winding and an auxiliary one for powering the control circuit and monitoring the current. To the right and below, there's a footprint for a temperature sensor connector, but the connector itself is missing.



I have no major complaints about the build quality. Everything is clean and neat; the components are secured with glue, and the board itself is coated with protective varnish.


The PWM controller is an OB2202CP. Nothing unusual here—it's a controller like any other. What caught my attention is that it also receives a voltage taken from a divider across the AC input.
There are no isolation slots in the board, but the required creepage and clearance distances are maintained. The high-voltage circuits also use two resistors in series, although ideally there should be three. Two are still within spec, since each resistor is rated for up to 200 V.




The output section is more complicated. It has two chips: on the right is a basic AS358M op-amp (an LM358 equivalent); unfortunately, the marking on the one on the left has been rubbed off :(
I didn't redraw the schematic because one of the chips has no marking. It could well be some kind of microcontroller, though I wouldn't be surprised if it were just a comparator. Also, a fair amount of sealant covers the upper part of the board.
If there's interest in this charger, I may redraw at least the output section. The primary side isn't anything special and could be replaced with any similar circuit.



Since I'd opened the case, I ran a thermal test and measured the component temperatures. I connected the charger to the same load and set it to the same operating mode. The test lasted 1 hour 30 minutes. To take the measurements, I removed the power supply cover without disconnecting the load—in other words, while it was running.


The MAX field shows a temperature of around 100 degrees. It turned out to be the snubber resistor for the output diode heating up; in this case, that's not critical.
Thermal images taken 30, 60, and 90 minutes after startup.


The temperatures of the individual components all stayed within their limits. I'd even say there's about 20 degrees of headroom, despite the test conditions being harsher than real-world use.


Conclusions.
After testing and inspecting it in detail, I came away with a positive impression, for several reasons:
1. Good build quality and design
2. Although the components are “Chinese,” I didn't notice any obvious cost-cutting
3. The battery is disconnected when charging is complete.
4. There is protection that stops charging if you try to charge an over-discharged or faulty battery.
5. No overheating
6. Available with other types of plugs.
7. No modifications needed

There is one drawback: after reconnecting the power or restarting, the charger sometimes does not stop charging when the battery is full. This happens when the current after power is applied does not exceed roughly 500–600 mA. One caveat, though: this can happen with a fully charged new battery. With an old, worn-out battery, the voltage drops even after a short pause, so the charging current will most likely jump up to the cutoff threshold of the protection system. In other words, it should work properly in the situations where protection is actually needed.

In short, it seems that although the charger was manufactured (and most likely designed) in China, it was made for the European market. That’s why they didn’t skimp on filters and so on. The main thing that sets it apart from power supplies in the Mean Well class is that its capacitors (at least the output ones) are from little-known manufacturers.

Otherwise, I can say that yes, it’s expensive, especially since shipping on Nkon costs extra. If you buy one, it’s best to add some batteries to the order. But in my opinion, safety is worth more. I’d also appreciate it if anyone could share links to similar chargers with an automatic shutoff function.