I have been on the lookout for some off the shelf low cost DC-DC modules for a project that requires 5V at up to 3A. I considered doing my own but the cost of the parts is about 10 times the price of the low cost boards that are available from a number of local sellers here in Indonesia and on AliExpress etc.
I chose 3 modules that have 24Vdc input and either fixed or adjustable output. Two of the modules are rated at 5A and the other one is rated for 3A. The testing was done with a GW Instek GPS-3303 bench power supply and a Rigol DL3021 digital load. I also used the Rigol Ultraload software to capture the waveform and max and min values for each power supply.
Rp 9,000 in the local market (approx US$0.64)
This was the first module to be tested. Initial tests were not good as the output was very unstable. I added two 470uF reservoir caps to the input and output of the board to see if this would improve the performance and it made quite a difference. This board is adjustable and I had it set for 5.2V offload.
Hooked up to the DL3031 and the software setup for 0.3 amps min load and 3.0 amps max load, I configured the transient setting for 4 seconds and started the test.
As can be seen from the image below, the board did far better than I expected. Max voltage was 5.1888 at 0.3 amp load and min voltage of 5.0929 at 3 amp load. Not bad at all. I should add that without the output filter cap, the voltage goes from 4.8 to 5.6 volts so this board does need an external reservoir cap. You can see the effect of the reservoir cap on the waveform where it decays slow when switched to the 0.3 amp load.
Check with the Flir TG165
Rp 61,000 in the local market (approx US$4)
Next up was the XL4005 based board with a switching IC from XL Semi. This board is rated for 5V at 5A but the test was only done at 3A. A 5A test was also good but not recorded.
Hooked up the DL3021 and the same transient settings as used for the MP1584 above, I ran the transient test. Offload the output from this board was set for 5.2V as this too is adjustable. The voltage swing was 5.1873 at 0.3 amps and 5.0753 at 3 amps. You can see that there is little in the way of hold on the reservoir cap compared to the MP1584.
Rp 26,000 in the local market (approx US$1.74)
Lastly we have the Texas Instruments based board with dual external mosfets on the switching side. This has a fixed 5V output and offload this read 5.13 volts and is rated for 5A max output. Input is 24V.
Hooked up to the DL3021 and same settings as the other boards, it shows 5.1174 volts at 0.3 amps load and 5.0070 volts at 3 amps load. This one also remains much cooler when running at the full 5A load. This board has very good transient response.
Sadly my Flir TG165 failed to record the images of the other 2 boards so I was left only with the last one, unless I could be bothered to re-run the same tests, which I decided not to do. This was taken after running the transient test for 30 seconds and no appreciable heat increase.
All 3 boards performed very well. For projects with smaller power requirements, the MP1584 works well but does need external reservoir caps. The other 2 work with their built in caps.
Of the 3 boards tested, the TPS40057 one was chosen for a new project. The XL4500 board will replace the MP1584 board in another project.