100 W LED array.

It all started when I spotted a lot for a 100 W LED on sale at a big discount in a group-buy. I love gadgets like this, so I couldn't resist.

To be honest, I hadn't planned to post this review, but after a bit of a flame war in my review of defective 10 W arrays, where someone suggested that when looking through dark glass I simply couldn't see that the other dies were working too, I decided to show here that if an LED is working, everything should be clearly visible.

So, here's the review. The LED module arrived in packaging designed in the style of late minimalism. The envelope is the color of cheap cardboard. But that's not where the story began. The previous evening, I was walking past the post office and, knowing my order was due to arrive, decided to pick it up. They told me the parcel hadn't arrived yet.

Imagine my surprise when I checked the order status the next morning and saw that the parcel had been delivered to the recipient the day before, before I even got there. I got dressed, already gearing up for an argument, and on the way checked the mailbox, where I spotted the envelope with my order. I opened it at home and found this—


Of course, I appreciate the quality of service at our post office (I hope they aren't reading this, or they'll start doing a worse job), but I was surprised the order made it here intact. The inside of the envelope is lined with bubble wrap, but I doubt it would have helped if something reasonably heavy had landed on it.

I decided to photograph it next to a 10 W LED to compare their sizes.


By the way, it's 40–45 mm across (excluding the contacts); the array itself is about 25 × 25 mm.

The first surprise was the LED's weight. It's light. The LED weighs 11 grams.


For comparison, here's a 10 W LED:


The second surprise is the markings. Take a close look and try to guess which side is + and which is -.

The plus is marked at the top of the housing, the minus at the bottom, with contacts on the sides.
For the test, I had to use a C-8 filter this time, since the camera overexposed the shot with the C-4.


To my surprise, it all works. The photo shows a current of 1.6 A, a voltage of 33 V across the array, and power of about 53 W. Unfortunately, my adjustable power supply doesn't have a high enough voltage, so I had to improvise. I used a transformer rated at about 60 W with an output voltage of 38 V, followed by a bridge rectifier, a bank of capacitors with a total capacitance of 5000 µF, and a current-limiting resistor (three 10 Ω resistors in parallel). The resulting current is given above.

The array does have a slight variation in LED parameters, but it only shows up after the power is switched off, while the capacitors are discharging.


But as you could see above, everything looks fine at the operating current. By the way, you can see here that, for some reason, the LED strings are arranged crosswise rather than in the usual layout. I don't understand the point of this design choice.

Unfortunately, this isn't a light bulb, so it's hard to judge the brightness, but it seems roughly equivalent to an 80 W fluorescent lamp (with 53 W going into the array). In other words, if you put the LED and the fluorescent lamp side by side on a table, the amount of light in the room would be about the same (I sometimes get 4 × 18 W office fixtures brought in for repair or modification).

The photo was taken with a desk lamp on, fitted with a 35 W halogen bulb, and the mains voltage was 240 V.


In reality, the light looks brighter and less blue. Although the matrix is listed as cool white, it doesn't look nearly as bad in person as I expected. Sorry about the mess on the table—it's the result of searching for a way to power this matrix.

Pros:
it works :)
the entire matrix is working
the light is quite tolerable

Cons:
poor packaging
strange material used for the heat-spreading plate
Verdict: it's perfectly suitable for non-critical applications. I was concerned about the suspicious material used for the heat spreader (it was too light), but at half power it easily warmed up a CPU heatsink. After a 2–3-minute photo session and a few brief test runs with pauses beforehand, the heatsink reached about 55 degrees (I didn't bother measuring more precisely). This was with only КПТ-8 thermal paste, covering about 60% of the surface (mostly in the center), and without the heatsink being screwed down. So the housing material transfers heat quite well. Considering the low price (it was 5.39 when I bought it; now it's 8.99), it's definitely a good deal. Now I just need to build an enclosure for it (housing, heatsink, reflector, driver).