MP2359DJ is a pretty good little PWM controller

I came across a mention of this PWM controller on one of the forums. Since I sometimes build various electronic devices, I decided to order some and give it a try.
And since I’d ordered them—and even received them (which doesn’t always happen these days)—it seemed like a good idea to test them before confirming receipt.
Description, testing, results—all below the fold.

I actually received these chips a long time ago, but only got around to testing them now.
I ordered them around New Year’s, and they arrived on February 3.
The seller pleasantly surprised me several times, so I’ll include a link to this review in my feedback.
First, he shipped the chips quickly, then extended the order protection period himself, without me having to remind him.
I told him I’d received the chips, but would confirm receipt after testing them.

The chips arrived in a plain paper bubble mailer, although I recently received an LED in a plastic bag with no protection at all.


The second surprise was that the seller included 11 chips instead of 10. It’s a small thing—20 cents—but a nice touch. You could say one chip is for experimenting :)


The chips come in SOT-23-6 packages, with markings on both sides.
I found no obvious issues with their appearance, although the tape wrapped around the strip made me a little wary at first.


This chip is a step-down PWM controller with an integrated power transistor.

First, the specifications (translated from the datasheet). The full datasheet in English is available here.
Peak output current: 1.2 A
Internal MOSFET on-resistance: 0.35 Ω
Stable operation with a low-ESR ceramic output capacitor
Efficiency up to 92%
Standby current: 0.1 µA.
Fixed operating frequency: 1.4 MHz
Overtemperature protection
Cycle-by-cycle peak current limiting.
Input voltage range: 4.5–24 V
Output voltage: 0.81–15 V

The typical application circuit uses only a few external components.
Of course, there are chips that need even fewer components, but this seems perfectly reasonable to me.


The chip’s internal structure.
The only real downside (and even that’s debatable) is that it uses an N-channel transistor as the power switch. This makes things more complicated: an external capacitor is required, and the chip can’t provide a 100% duty cycle because the driver’s external supply capacitor needs time to recharge.
But this approach has an upside too: N-channel transistors usually have better characteristics than P-channel ones.

Another major advantage is the low reference voltage of just 0.81 V. I’ll explain why later.


There’s also a more elaborate application circuit for this chip.
At input voltages below 5 V, it’s a good idea to add diode D3.
At output voltages below 5 V, it’s a good idea to add diode D2.
No additional components are needed in other cases.


I mentioned above that the chip has a low reference voltage.
That makes it possible to build a simple LED driver around it.
The higher this voltage is, the greater the losses in the current-sense resistor. You can get most chips to work this way, but the lower the voltage, the higher the driver’s efficiency.
And it’s always better to have some headroom on the low side, since most of the simple PWM controllers I know of have reference voltages of 1.23–2.5 V.


Since I needed to test what I’d received, I had to put together a small test board.
I decided to use the simpler circuit, with a few changes because I was building it from whatever parts I had on hand.
The changes involved the capacitors.
The manufacturer recommends 10 and 22 µF for the input and output capacitors. I used 2 × 2.8 and 2 × 5.6 µF, respectively—about half the recommended values.
The driver supply capacitor is also supposed to be around 10 nF, or up to 1 µF with additional diodes. I used 0.1 µF without any diodes.
I used a diode I had on hand too: a basic SS34.
The recommended inductor value is 4.7 µH; mine was 10 µH.

So I deliberately compromised the converter’s performance. I also wanted to see how the IC behaved with component values that differed from the datasheet recommendations.


I first laid out my own PCB, but I wasn’t happy with it, so I decided to follow the manufacturer’s recommendations instead.
PCB layout for this kind of circuit is quite important. It’s not enough to simply connect the pins according to the schematic; you also need to get the board layout right, since it can affect a lot of things.
For example, a poor PCB layout can increase output voltage ripple or even make the device completely inoperable.
This is the manufacturer’s PCB layout.


And this is the board layout I came up with.


The rest was pretty straightforward. I made the board using the toner transfer method, which I described here.
I somehow forgot that the board was etching and left it in too long, so the result wasn’t as good.
I used 1 mm-thick PCB laminate. By the way, it’s excellent stuff: once etched, the board is translucent. I’m looking for more of this laminate now, preferably a standard-size sheet.


I selected the components.
You can easily calculate the feedback divider resistor values by visiting this page. I think it’s clear what’s what without any explanation :)
The input values were an output voltage of 5 V and a voltage of 0.81 V at the divider output.
I chose 10 kΩ for the upper resistor, and the program gave me 2 kΩ for the lower one.


The capacitors were salvaged from a monitor board; the inductor and diode are new.


In the end, I got a small, reasonably neat board.


The board measures about 23 × 12.5 mm.


First, I tested the board using a “test setup” consisting of:
Power supply
Electronic load
Multimeter
Oscilloscope
Non-contact thermometer

I tested the temperature rise and how it performed under load.


But since I was interested in efficiency as well as heating, I had to use another multimeter.
The power supply's ammeter is less accurate than the multimeter, and I wanted to get more precise measurement results.
Since the load was unchanged, I grouped together the photos of the current measurements and the oscilloscope traces from the previous test.

Input voltage: 10 V
1. Load current: 0.6 A, output voltage: 4.84 V
2. Load current: 1.2 A, output voltage: 4.80 V
In both cases, the ripple was right at the limit of sensitivity, even though the probe was set to 1:1.


Input voltage: 15 V
1. Load current: 0.6 A, output voltage: 4.83 V
2. Load current: 1.2 A, output voltage: 4.81 V
The ripple level was similar to the first test.


Input voltage: 20 V
1. Load current: 0.6 A, output voltage: 4.83 V
2. Load current: 1.2 A, output voltage: 4.81 V
And again, the ripple level was right at the limit of sensitivity.


After that, I tested the board in a few more operating modes, but didn't take any photos.
1. The converter's quiescent current is 1.3 mA at 10 V and 1.4 mA at 20 V. Of that 1.3–1.4 mA, about 0.3 mA is drawn by the feedback divider. So its quiescent current (outside standby mode) is excellent.
2. I tested the converter's operation under a short circuit. The input current is about 0.1 A across an input voltage range of 10–20 V. The chip starts heating up in this mode.
3. Since the chip heated up under a short circuit, I also tested the thermal protection.
When the chip's case temperature reached about 100°C (it's difficult to measure the temperature of such a tiny component), the chip switched to a start-stop mode at about 0.5 Hz. The current draw dropped to 50 mA during the pauses.
Once the overload was removed, the chip immediately returned to normal operation.

I found a table in the datasheet showing efficiency at different operating points.
I tested at slightly different input voltages, but I don't think that makes a critical difference.
As you can see from the graph, the chip reaches its maximum efficiency at an output current of about 0.6 A and an input voltage of 12 V.


My calculations showed that the converter actually reaches almost 92% efficiency, but at an input voltage of about 15 V.
That said, I should point out that I used components that were more likely to reduce efficiency than improve it. Even so, efficiency never fell below 87.7%.


Summary.
Pros
Price
Output overcurrent and short-circuit protection work correctly
The overtemperature protection works just as reliably
Good efficiency
Simple circuit; no need for large electrolytic capacitors
Very low ripple
1.4 MHz operating frequency
Low reference voltage of 0.81 V
Excellent seller

Cons
Probably the inability to reach a 100% duty cycle, since the driver supply capacitor needs time to charge.

My take: I liked this chip. It's inexpensive, simple, and works great—and the seller gets a thumbs-up too.
Of course, there are better chips with synchronous rectification and higher current ratings, but I didn't need more. The compact size, simplicity, and price outweighed those advantages.
Overall, I recommend it.

As additional materials, here's an archive with the datasheet, schematic, and PCB layout — link.