5 V 5 A converter based on the TPS40057. Small, inexpensive, and great.

One of the readers asked me to review this converter, but somehow all my previous attempts to order one had gone wrong. Finally, I managed to buy this little board and test it :)
As always, the review includes an inspection, tests, and conclusions.


First, I should mention that I already had a similar converter, only that one was more expensive but had four USB outputs — review. So most of what I said there also applies to the converter reviewed here.
The only differences are the design, size, and rated output current. The other one was rated for 8 A, while this seller specifies only 5 A.

They differ in price by about a factor of two. The one reviewed here cost 1.43 (including shipping, which was paid), and since USB ports aren't always needed, that's a significant difference.

It came in a regular antistatic bag, with a neat little board inside, slightly smaller than a matchbox.
Dimensions: 45 × 30.5 × 16 mm.


The converter is built around a TPS40057 PWM controller. Next to it are two TOSHIBA power transistors, TPCA8053-H, rated at 60 V and 15 A, with an on-resistance of 13.9 mΩ.


The TPS40057 controller can be used to build synchronous rectification step-down converters. It operates at frequencies up to 1 MHz and has built-in drivers for the MOSFETs, as well as thermal and overload protection.


1. The board has a three-pin connector for the input power and output. In my opinion, this is a really awkward design. On the previous converter, ground was brought out to separate pins, so it had a four-pin connector.
2. The board also has two electrolytic capacitors, 470 µF, 35 V. And of course, they're cheap Chang capacitors, which are doing almost nothing here. Besides, a 35 V capacitor on a 5 V output looks a bit odd.
3. The inductor, on the other hand, is good: it's wound with a copper strip of a decent cross-section. The diode to the right provides reverse-polarity protection.
4. The underside is bare, so if needed, you could probably even attach the board with double-sided tape, though it also has three mounting holes.


I didn't redraw the schematic, and there's not much point anyway, since it's very similar to the typical circuit in the datasheet.


For testing, I used a setup consisting of a power supply based on an RD6006 converter and an EBD-A10H electronic load. I didn't account for the voltage drop in the input wires because they were short, and I used a four-wire connection for the load.
When power is applied, a red LED connected to the converter's output lights up.


The board also has undervoltage protection. It starts up at an input voltage of 8.2 V and shuts down at around 8.0 V, so there's a small amount of hysteresis.
When the board shuts down, the input current drops almost to zero, which means this feature can protect batteries from over-discharge. For example, with a 3S lithium battery pack and a shutdown threshold of 8 V, that's 2.66 V per cell, which is perfectly acceptable for most batteries.

In operation, its own current consumption is about 16–17 mA at 9 V and 21–22 mA at 24 V.


The output voltage is slightly high, at 5.05–5.08 V, but the board handles loads very well, up to 9.5 A, with the voltage dropping to 5.02 V.
Load graph at an input voltage of 12 V


At 24 V input, the results are almost identical.


I expected the board's efficiency to be a little higher than what I got in the test. I measured it at two input voltages, 12 and 24 V, and across load currents from 1 to 9 A. The horizontal axis shows the load current in amps.


Next came the thermal test. It was run at an input voltage of 12 V: first for 20 minutes at 50% load, then for 20 minutes at 100%, and another 10 minutes at 150% (7.5 A).


The converter performed reasonably well, although the input diode got very hot. In fact, it was the hottest component on the board.


The previous converter had much better reverse-polarity protection, using a MOSFET. Here, the manufacturer cut costs by using a diode, albeit a Schottky one. I think it would be more appropriate in this case to put a diode across the power input, with a fuse installed ahead of it, but for the second test I simply shorted it with a jumper.


The results were immediate: efficiency went up, especially with an input voltage of 12 V. At an output current of 9 A and an input voltage of 12 V, the diode dissipated about 1.5 W.


Interestingly, although the graph now looks more like the one in the datasheet, it still differs noticeably. But the datasheet graph was measured at 24 V input, 3.3 V output, and a different current range, and the components are different too.


I liked the temperature results better, though. Even at 7.5 A, the hottest component was one of the transistors, at about 100 degrees. And the rated current was only 5 A.


I captured the waveforms in four operating modes: at zero output current, and at 2.5, 5.0, and 7.5 A. The ripple amplitude barely changed, staying around 100 mV; only the amplitude of the brief spikes varied. The previous converter showed roughly the same pattern, but its ripple amplitude was about twice as high.

Below are two more interesting waveforms, where you won't see the short spikes. These were captured at the converter's output capacitor with load currents of 2.5 and 7.5 A.


In this photo, the output capacitor is at the bottom.
At the very beginning of the review, I complained about the use of a three-position terminal block. But it looks like the PCB layout isn't quite right in other ways too. For example, I don't like the long trace running from the capacitor. There are also narrow sections in the ground trace, which could be causing the effects shown above.


The manufacturer's PCB layout in the datasheet is noticeably different from the one shown above, but at the very least, that board is about one and a half times larger.


And at most, the datasheet board is four-layer, while the “Chinese” one has only two layers. That could also affect the output ripple, especially the short spikes.


Even so, I really liked the board and can definitely recommend buying it. I would, however, recommend removing (shorting out) the input diode, or putting it across the input and adding a fuse. Without it, efficiency is higher and, consequently, the board runs cooler.
If you need USB outputs, the board from the previous review is a good choice, and I recommend buying that one too. It even has USB port “cheat” circuits that tell devices to charge at the highest current they support.

That's all from me. I hope you found this useful. I'd also like to wish everyone a happy New Year and all the best—and, of course, may your devices always work reliably and never let you down :)