I recently tested the output of the
Koolflight UBEC Mike has in his store, and was pleasantly surprised. Here are a couple pictures of the UBEC without the heatshrink:
This UBEC is based on the adjustable version of the
LM2596 switching IC, which is the same IC used in a cheaper UBEC I
previously tested - except the cheaper one used the 5v-specific IC and used circuitry to boost it to an advertised 6v.
While the cheap version was a little disappointing, the Koolflight actually does put out a true 6v even under load. The below table shows the test results. I had made an adjustable loading circuit, but fried the constant-current transistors when I tried set the current too high. So anyway, I used a few power resistors I had laying around instead:
The table also shows a very low drop-out voltage (0.17v max) so this can be used all the way down to 2s and still have a solid 6v out.
Using a Fluke true-RMS meter, I was not able to get any AC ripple on the output even at high current. So, it produces a nice clean DC voltage.
Efficiency is pretty good too, even under extreme loads. The peak efficiency might be a bit higher, but I didn't have enough loading options to get more amperage samples. Efficiency will also go up if used with higher input voltage, so you can consider the figures above pretty much worst-case.
Yeah, this UBEC is larger than many others, but the larger PCB really helps control the heat. Even after 20 seconds of almost 5A, the heat was manageable IMO. And the inductor is larger and uses slightly larger gauge wire for better performance.
Speaking of the inductor; in the top view image above, there is a ferrite ring around the inductor to help shield nearby electronics from noise. Also notice the underside of the PCB - it's just about a solid ground plane. Koolflight took some pains to make this pretty noise-free. Like the KF instructions say; if you have noise/servo glitching, orient the UBEC so the backside faces your electronics. Makes sense to me!