Compact voltage converter with Quick Charge

A review of another converter, this time designed for a single port, but with QC support. I’ve already posted a review of a similar board, but this one is cheaper and smaller.


As with the 5 V, 8 A board, I ordered these as an add-on to another order, simply because that worked out cheaper. Shipping was paid, and ordering a single board would have cost more than the item itself: the board costs $1.23, while shipping to my country is $2.03. But I bought 22 boards, so each one came to $1.34 including shipping, which I think is pretty good.

They were packed compactly but securely, with the boards wrapped in bubble wrap. Nothing was damaged in transit.


You might think they broke during shipping, but they’re actually not as easy to snap off as they might look.


I gave 20 boards to a friend, so from here on I’ll be looking at my two. Each board is exactly one-third the size of a standard matchbox. It has components on one side only, so it’s quite easy to stick inside an enclosure with double-sided tape.


Board dimensions from the seller’s page.


Let’s snap the panel in half and move on to the description and tests.


The circuit is very simple: a single chip with its markings removed handles everything. It controls both QC mode and the conversion itself, and has a synchronous rectifier. On the right is a TVS diode rated for 12 V, which provides some protection for the load if the converter fails. Keep in mind, though, that this protection won’t save a regular smartphone or tablet designed for 5 V only, since it clamps the voltage at a higher level.

Specifications:
Input voltage: 8–32 V, but since this is a step-down converter, you’ll need to supply around 14–15 V to get 12 V out.
Operating modes: QC2.0, QC3.0, but it supports a cut-down version of the protocol, limited to 12 V.
Output ratings:
5 V, 3.4 A
9 V, 2.5 A
12 V, 2 A.

I can say right away that the output ratings are more than enough for the vast majority of devices. Typical current draw is 3 A at 5 V and 2 A or 1.5 A at 9 V or 12 V, respectively.



The controller’s markings have been carefully removed. There’s a resettable fuse on the input.


First, I’ll connect the converter to a power supply to see which modes it supports.


So, by default the output is 5.16 V. Excellent.
It supports most protocols, and Samsung AFC was detected too. But although I tried connecting a Samsung phone, it wouldn’t work in this mode. Perhaps the phone itself doesn’t support it.
The minimum voltage in QC3.0 mode is 3.48 V, and the maximum is 12.13 V—pretty standard.


For the remaining tests, I connected the load directly to the board, bypassing the USB port, while the USB tester controlled the operating modes.


The converter delivered 5 V without any trouble, with current up to 3.4 A. At first the voltage dropped to 4.98 V, but it rose as the load increased, reaching the same 5.16 V at the maximum 3.4 A as it did with no load.


At an output voltage of 9 V, the results were much the same. I got 2.5 A at the output, but at a higher current the board went into protection mode.


At 12 V, the voltage also rose slightly with the load. The board delivered 2.1 A without any trouble, a little above the rated 2.0 A, but at 2.3 A it started reducing the output voltage.
All tests were performed with an input voltage of 24 V.


Let’s connect an oscilloscope and see what’s happening at the converter’s output. To reduce the electronic load's influence, it was powered through a transformer-based power supply.


The three traces on the left show the output at 5 V with no load, at 50% load, and at 100% load. The traces on the right show the same for 12 V.

In short, the results aren't particularly pretty. With no load, the ripple is 100 mV at 5 V and just 50 mV at 12 V, but under load it reaches 300 mV. The waveform is almost unaffected by the load current or output voltage.
300 mV is a lot at 5 V, but tolerable at 12 V. Tests with an actual phone showed that it worked just fine, though.


After that, I moved on to thermal tests. First, I set the output to 5 V and applied a 50% load (1.7 A). After about 7 minutes, the board had warmed up to 60 degrees—more precisely, that was the temperature of the PWM controller.
At 100% load, the converter ran for 7 minutes before shutting down in protection mode.


The graph looked something like this. Additional cooling is needed to get the rated 3.4 A, although later I began to suspect that the fuse had tripped. Its current rating depends on temperature, and the hot PWM controller is right next to it.


At an output voltage of 12 V and load currents of 1 A and 2 A, the converter ran 5–10 degrees cooler and worked perfectly.


During testing, I measured the input current and efficiency at input voltages of 15 V and 24 V, output voltages of 5 V, 9 V, and 12 V, and load levels of 20%, 40%, 60%, 80%, and 100%.
Green — 5 V
Red — 9 V
Blue — 12 V


I also plotted efficiency against operating mode and voltage; the colors correspond to those shown above.
While we're at it, maybe someone knows the answer to this. When I select several columns in Excel to plot, they appear in the chart, but in reverse order: 3-2-1 instead of 1-2-3. How can I control this?


Conclusions.
I don't want to be too hard on these converters, though there are things to criticize—for example, the high output ripple and their inability to deliver 5 V at the maximum load current for extended periods. But at this price, I think those shortcomings are forgivable. The main issue is the ripple, since in real life it's unlikely you'll find a device that needs 3.4 A at 5 V.
In my view, the build quality, price, size, stable operation in the other modes, the range of available modes, and the output protection TVS diode more than make up for the drawbacks.

That's about it. I hope this information was useful, and as always, I look forward to your comments.