UFO … That’s what they’re called, maybe because they are around, maybe because the vendors grew weary of calling them luxeon clones. No, I’m not writing about some errant weather balloon, the planet venus, or my neighbors truck seen through beer goggles… I’m writing about some high-output LEDs
These are the first of any ‘next gen’ leds I’ve worked with … I’m stoked they are so tiny … its often hard to tell the scale when you just see pictures of these things online with nothing of known size in the shot.
the big black module is an RGB ufo, with three one watt emitters in a single package. It’s wired as a common anode device… should work out well with my led controller project. The small led in the top of the picture is a white piranha led. The smaller silver module is a blue UFO.
This is the die of a blue ufo, running at a greatly reduced capacity. I still had to stop my camera all the way down to take this shot with any amount of detail … (1/4000s f7.3)
More pictures to come. I have a red, blue and green mounted to an old 586 heatsink which provides ample cooling to the 3 watt load, I’m going to work on mounting the driver circuitry to the backside of the heatsink and should have some pictures shortly!
Wondering aloud, would it be possilbe to socket those LEDs so you could reconfigure the array later? (any thoughts, JustDIY?)
The quote came from a longer thread, where the led being discussed is the high-flux “piranha” package. Well, he must have me pegged as some sort of piranha expert 😛
So I pulled out a few leds, and some different sockets. Here is what I came up with:
The black strips are Molex C-Grid board to board connectors … the mating board would have pin headers on it, and the two shall mate. But, they also work decently well for holding other things, sorta like a breadboard does.
Depending on how you wire the pins underneath, the headers accept the LED in either orientation (piranha have two pins for each cathode and anode, which are connected inside the package)
I also tried other cheaper and easier to find sockets, like machined pin and standard spring … they just don’t line up with the piranha’s leads.
Here’s a cheap incandescent garden / landscape light I retrofitted with 4 RGB leds. It works OK, but the retrofit module is difficult to assemble without a proper pc-board and there are problems with diffusing the light properly.
This breadboard is the first time I had dealt with high / super flux leds… otherwise refered to piranha or ufo package leds.
These came from www.lsdiodes.com and have a viewing angle of 45 deg… it makes them super tremendously bright looking indicators, and probably would work great for tail lights and such.
The piranha led has the chip or die mounted on a large heatsink connected to the anode … the cathode is also a heavy duty looking assembly, but its only connection to the die is a hair thin gold wire. The piranha led is an illumination grade led; it is designed to have a very high light output and be installed in close proximity to other leds of its kind. The large heatsink carries heat away from the die, but you still need to do something with that heat. The PCB board should have as much copper on it as you can fit, and the traces feeding the anode should be as beefy as possible so they can conduct heat from the pins.
I built this to test some “color washing” ideas I had, but it didn’t turn out like a hoped. In my mad rush to get it done (or maybe it was the post 1am state of mind), I wired it up backward… I was shootin for common anode, that is each led shares a common voltage supply and is controlled via their ground connection… but what I built is common cathode. Each led shares a common ground connection and is controlled by the supply.
Why is that bad? Well… If you’re going to use digital logic to control things, the easiest way is with a MOSFET transistor … the cheap and common N-Channel mosfets deal with low-side or ground connection … so with all the leds sharing the same ground, I cannot controll them independently.
These little jems are serial controlled (i2c) PWM generators, designed by Philips Semiconductor for the purpose of controlling leds. I received samples of both the 9531 and 9533, which contain two separate pwm generators, and either four or eight outputs.
Above is a 9531 mounted to a prototyping adapter so it would go into a breadboard. Original tests with the chips were very interesting. The onboard pwm generator is capable of a wide range of frequencies and a full 8-bit gamut of duty cycles (256 steps). I ended up not liking these chips for a couple of reasons…
The biggest drawback is the chips only contain 2 pwm generators … while this would be great for a dichromatic white lite (yellow + blue), it’s no good for trichromatic (red + blue + green) … so the development would require two chips, and additional logic in the program to figure out which chip controls which color. The programming isn’t that big of deal… but I’m lazy. The second drawback is I couldn’t find anywhere to buy these. I don’t have any venture capital, so I can’t buy them in lots of 1000 from a regional distributor, and the supply houses I normally buy from didn’t have them in stock at the time I was working with them… so they’re on the back burner for now.
This is my protopod. I think I built four or five of these, some came out better than others … point to point soldering of jumper wires isn’t the best approach, especially when they’re all overlapping and what-not
each of those leds is rated at 200mW, but I’m only driving them around 100mW because they need a fairly heavy pcb with good thick traces to sink the heat.
you can see that I’m not the worlds cleanest prototype builder … but heck, I don’t care as long as it works
Here is junior, in a mould of wax. I chose wax because its all I could think of that was:
A) easy to shape
So I quickly made a shallow square hole for my pc board to rest in… I was able to use some regular screwdrivers to carve the wax. After a lot of searching, I discovered blocks of wax at the grocery store in the canning section.
The epoxy is too thick to really ‘pour’ … you sort of ooze it into something. I scooped up a big glob with a stir stick and drizzled it into the mould. Then I placed the light in, and drizzled more epoxy on top.
Here’s Junior … I call it junior, its the smallest led light I’ve succesfully designed so far. roughly the same size as a quarter and about three thick. The black stuff is MG Chemicals 832-TC thermally conductive epoxy. It has harnded into an impregnable shell around my light, and helps keep the leds cool at the same time. The epoxy is fully curied now, so its time to abuse it some. I think it will spend a few nights in my freezer, and then a few days on the dashboard of my car, baking in the sun.
Junior is a light for a larger project I am developing and hope to sell someday… I’ll be sure to share more details here as I get things finalized.
This led light contains three high flux leds and combined with its controller, can project any color of the rainbow.
Check out this video of junior in action, before encapsulation