Castle Creations UBEC Test

[BrianG]

I got one of these in the mail the other day so I thought I'd put it through its paces on the bench.

First, I took off the heatshrink. The UBEC is sorta sandwiched between two PCBs, but the top one looks to be more of a noise shield and comes off fairly easily. There is some metal on it to act like a heatsink, but nowhere for the heat to go really, so I think it's mainly to minimize noise from radiating from the coil.

A peek at the circuit kinda surprised me. No regulator-on-a-chip designs common to todays BECs, but actually made up of what looks to be discrete components. A part number search on the ICs didn't turn up any direct matches, but it looks to me like the BEC circuit is based off a surface mount LM317 linear regulator set in switching mode somehow, with an FET handling the switching and high current flow - acting like how a pass transistor would function in a linear regulator. I could be wrong here, but that's what it looks like to me.

I could read the part numbers of all the ICs, except for one. It was just too darn small to see, and even if I could, the text is distorted somehow (probably intentional to keep people from doing what I am trying to do). I'm sure some of the complexity of the circuit can be attributed to the fact that this can be programmed, so there is some logic circuits in there as well.

Anyway, on to the testing and pictures.

For the input voltage, I used my regulated switching PS set to 14.8v. I figure this would closely emulate the common 4s setup. I then set various meters to measure the input voltage, input current, BEC DC output voltage, BEC AC output voltage ripple (switching noise), and BEC output noise frequency. I then dug around my parts bin and found various power resistors to load the BEC. Before testing, I set the output for 6v on the button via the Castle Link software. Below are the results of the test:



Efficiency seems up to par with most switching designs, and was pleased to see close to 90% even at higher currents.

There did seem to be a decent amount of switching noise going on, but that's somewhat expected especially at higher currents. A 1,000uF capacitor on the output helped a lot, but kept it off for the test.

Considering its small size, the temperature wasn't too bad. A little hot for my liking, but don't forget I was running for a minute straight at a constant load. This wouldn't happen in the real world, so expect lower temps. However, I think a little heatsink would help cool things down a little at the expense of increasing the size a little. Worthwhile IMO, but that's up to you. By far the hottest component is the coil (the component marked 5R6 in the blurry pic below). Can't simply slap a heatsink to the top though since the coil terminals are there, unless you machine out grooves to clear them.

All in all, this is a really nice little unit. Provides ample power in a really small package and the large ouput voltage adjustment range is a VERY nice feature. At 9v max, you could use one of these to power several cheap and easily attainable 12v fans (at lower rpm) to cool your motor/ESC.

The short-circuit protection works well. If I short the output, the current spikes to about 7A for a VERY brief time (talking milli-seconds) before shutting down. The surge may be higher, but that's the highest value I've been able to get my meter to display (stupid sampling rate issues). Input current drops just as quickly. Remove the short and it comes back up fine.

Below are some naked pictures if you like that kind of thing. One of the pics has the part numbers of the ICs where I could read them. Sorry about the blurry image on one of them - by the time I realized the picture was crappy, I had already re-heatshrinked the BEC and didn't feel like ripping it off again.

[MTBikerTim]

Impressive testing. Great work.

[glassdoctor]

That's it, you are FIRED!!!! WE can't have fuzzy horrible pics on here!! LOL

Nice info. We need to put that in with the other BEC test results... I've been curious to compare the CC with them. Cool....

[BrianG]

Thanks. :P

I'll probably end up ripping off the heatshrink again to add a heatsink, so if I remember, I'll snap a clearer picture.

For packs up to 40v and ~3A draw, I like the Koolflight UBEC best. The voltage is stable under load, no noise even under heavy load, efficiency is high, and cooling is better.

For the other UBEC tests:

Generic UBEC

KoolFlight UBEC

ParkBEC

Is the Koolflight is a switching BEC too?

[BrianG]

Yep. If a regulator is that small, has little or no heatsink, and is good for that high voltage and current, you can be certain it has to be a switching type of regulator. A linear regulator would need a gargantuan heatsink to stay cool for those specs.

[Serum]

Great info Brian.

Can you program it to the highest voltage and have it loaded at the highest current?

[jhautz]

Originally Posted by Serum

Can you program it to the highest voltage and have it loaded at the highest current?

You tryin to break Brians CCBEC.

[BrianG]

Thanks.

Don't know, didn't try it. Even if I did, I'm kinda limited by the power rating of my test resistors (between 1w and 25w). At 9v and 5-7A, output power is quite high. I'd have to invest in a set of higher power resistors if I plan to do this type of thing more often.

Even then, I think BEC temps would get out of control. I only loaded it to 4.69A max and it got quite toasty after just 1 minute - and that's with only 3.5w being dissipated on the BEC (88%). Assuming efficiency stays the same with 9v @ 5A output, that would be 5w of power on the BEC. Ouch.

[BrianG]

Originally Posted by jhautz

You tryin to break Brians CCBEC.

lol, that's ok, it's cheap. And besides, I'm using it "within specifications", so I should be covered. Seriously, this type of continuous current draw is a torture test and I wouldn't expect Castle to cover it under warranty. Even so, it's only $22. All in the name of science!

[Serum]

Amen, all in the name of science.. The neat part is that you can program it, and it's relative cheap.

[BrianG]

If the mood grabs me, I might stop by radioShack and see what they have for a power resistor selection...

I wonder if this BEC is like DimensionEngineering ones where you can hook them up in parallel for ~2x the current output? Of course, they would have to be programmed for the same output voltage.

[johnrobholmes]

I talked to Pat about paralleling them for a while, and he said it would work but not effectively enough for 2x the power. Some bit about matching components and how one BEC may put out .01v more and take more load than the other. He did say it would work well as a back up system though, and that it would probably give about 125% of one BECs rated power.

Good test there. Got a way to watch the voltage under fast changing input voltage and changing load?

[BrianG]

Update the original results table. Forgot to key in the results from the drop-out test...

JRH: I knew that you wouldn't get exactly double, but thought you'd get more than 125% of one's output. Poo.

I did try to fluctuate the input, but I'm at the mercy of the sampling rate of my meters. One is a Fluke 87 (true RMS meter) that I use to measure DC output V, AC output V, and output AC noise frequency. The other two are cheapie meters that I use to measure DC input voltage and current. I did calibrate the cheapie ones against the Fluke, so they should be accurate to about 1%. I really need to get a multichannel digital storage o-scope!

[jhautz]

Originally Posted by BrianG

I really need to get a multichannel digital storage o-scope!

Geez... who doesnt have one of those?