250W DC-DC Boost Converter Adjustable 10A Step Up Constant Current Power Supply Module Led Driver For Arduino (WHL #21)
What do you do when none of your existing power supplies can provide enough power for some decent LED module? Yeah, right, you buy a bigger one!
So I did. I had some 100W/150W DC-DC converters laying around, but most of them cap at around 36V. And with the very nice Toshiba E-Core 92W LED module that wants 40V or more, that’s just not good enough (same goes for the popular 25W/50W/100W LED modules!). So the new one in my collection is this:
Some more shots:
I bought it on AliExpress back in mid-January for 4.31€ including shipping.
Tech specs are, according to the seller: 250W total output power (boost = step-up only), non-isolated output, input current 8A or 10A with additional heat sinking. As the output voltage is adjustable from 10V to 50V (and there should be a CC mode or limit as well), it’s obviously limiting input source selection if you want something near the maximum output power. No big deal, plan your circuit accordingly.
Operating frequency is said to be 150 kHz which is okay, “efficiency” is “up to 96%” (these might be Chinese percents, which are slightly smaller than European ones).
There’s not much on the board, you can see the huge inductor (my AVR transistor tester says 70µH/0.3Ω), some beefy diodes (SS56 – 60V/5A Schottky), some SMD caps (220µF/50V) and two 10-turn pots for voltage and current setup. The chip underneath is a TI TL494 PWM controller…which looks a bit dodgy to me. Not only does it have a strange stepped line at the side, but it also features no date code (I clearly need a microscope for documenting that). And, according to the TI data sheet, it is recommended for AC-DC applications and DC-DC is not even mentioned. Ah, and it’s rated for 41V absolute max.
As it explicitly says “DIY a power supply, 12V can input and output can 12-50V adjustable”, I tried it on 12V first. I didn’t go above 5A in, as I know the wiring diameter of my ATX power supply, and especially the weakness of the 4mm banana plug output terminals. Still, the supply ramps up the voltage just fine and the LED module (actually both) came to life. Great!
So for the test under full load for this LED, I needed a higher voltage for the input. With 12V to be converted to I assume 44V/2.1A, that’s pretty damn near the 8A specced input current limit. I got me a nice 8S 26650 LiFePo pack from work over the weekend, which also needed load testing (great employees like me do that totally selfless for fun in their free time…). This roughly doubles input voltage for the DC-DC to 24V, which halves current. The cells handle the current just fine, so no worries from the input side.
And it worked. For around a minute. It really didn’t get warm, so the 96% efficacy might be not that far off from reality. 4% losses at 100W can be perfectly handled by the aluminium PCB as shown above. But something had their magic smoke escaped… After that, the inputs act as a short-circuit – you can warm your input terminals as much as you like, but there is no output at all.
Tracing the component down isn’t really difficult with the Half Ohm. It’s the NCEPOWER NCE7560K TO-252 N-Channel MOSFET which is the fully gunked part below the inductor. Obvious guess: It was handling the switching to make the DC-DC circuit work. It’s rated at 75V DS, ±20V GS, 42A drain current at 100°C and a typical 6.8mΩ of RDSon…sounds beefy and isn’t easy to kill one would assume. It was also properly soldered and therefore not too easy to take it off the huge heat sink. But if your controller does strange things and the massive inductor just follows the laws of physics, overloading the thing (sans protection diodes or other unnecessary stuff) seems possible. So now the RDSon of less than 5mΩ is valid at all times…even when you have no gate voltage at all. Guess what’s hindering proper operation
…I complained and got my money back. Maybe I put some replacement MOSFET on there, but I doubt it will work for long. Scrap the board, try the next. And if you ever wondered how to deliver photo or video proof of a non-working item in the Ali dispute process – this is what I uploaded and instead of lengthy discussions, I got an instant refund.
My verdict? Bad Wan Hung Lo product. Failed too soon, customer was able to file a complaint and get a full refund. Needs improvement in the planned obsolescence part of the design!
More junk to come, I’m on a Wan Hung Lo killing spree now.