Nichia 144A and 229A series: bigger dies, more output, 90CRI included,.......but no thermal slug :-(

We had Bridgelux SM4 but it wasn’t happening…

I’m wondering if the cathode acts as the thermal conduit, therefore must be isolated from ground, unless you want dircect drive.

Which would work better depends on which contact of the LED has better thermal connection to the die. It could end up an advantage, because packages like Cree uses must have internal insulating layers.
Djozz’s sketch would be great if the copper is thick enough.

Why Nichia… Why? :cry:

A makeshift MCPCB shouldn’t be to hard for initial testing. Just use a Cree board whose anode/cathode footprints are close and grind out a little ditch where the thermal pad is with a Dremel, so there is no contact.

Vewy intewesting!

Reminds me of Elmer Fudd hunting wabbits.

If they did a good design and uses the two main contact pads are the thermal conductors then the center pad should not be really needed. ( i guess most of us are used to having the center-pad on Cree emitters to match Noctigon stars, and like to over-drive LEDS to near burn-out. as long as the die, substrate, and two thermal/contact pads are designed well, and we have a good noctigon designed for these LED then where won;t be a problem.

So, if there is a dielectric layer under the electrical traces, how will a Noctigon help? This will be just like all the old Cree boards that have lesser thermal characteristics. The heat being dissipated through the trace pads will still have to go through the dielectric layer and as such, will be slowed down. This is ok if you don’t overclock, not at all conducive to most of us modifying flashlights.

Yeah, DTP stars are pointless if you’re not using the DTP! Best bet would be to make a board like djozz showed, with thick and large copper pours, and use the fancy nano dialectric layer that supposedly has a high thermal conductivity.

We’re just gonna have to put due diligence to work, aren’t we? Build em, work em hard, see what shakes loose…

To simplify, 1D thermal resistance is a function of conductivity, layer thickness and conducting X-section area. If we’re stuck with using dielectric and copper, we can control 2 of those three variables.

Mitigate thermal bottlenecking by designing positive/negative copper pads as thick and wide as possible, instead of narrow traces. Heat flux can flow through larger cross section of dielectric layer.
Bottom layer is copper and thinner than than +/- pad layer.

For given dielectric layer thickness,
Larger X-section of dielectric -> smaller thermal resistance of dielectric layer -> smaller ∆Temperature across dielectric layer -> cooler temp @ LED die. It won’t ever be as good as true DTP but depending on PCB size it could approach DTP performance. Also, if it works well with 16mm board it will only do better at 20mm.

Beryllium Oxide (Wikipedia)

Cheers ^:)

BeO has safety risks but a similar safer ceramic might suffice. I’m guessing this is a cheaper-easier way for Nichia to reach their efficiency goals within their stated usage parameters. That would be a good business move for an emitter manufacturer, but maybe not for us flashlight people.

It is well to remember that we’re just riding along with the emitter manufacturers- we are not driving the bus. If the emitter business slows due to market saturation or whatever maybe they’ll spend some time designing things just for us. Until then bursts of brilliance followed by dim blue haze (sometimes followed by blue air) will be our baliwick and bane but it shall not stop our relentless pursuit of more and purer lumens! Onward and forward! Full speed ahead! :+1:

Phil

Might be an instance where the nano materials dielectric would be useful combined with max sized +/- traces.

Nano-what?

FastTech blows goats for half price, once again my order got delayed 12½ days later, they don't seem to be able to get a hold of these alu-nano plates. I'm on the verge of ordering a Sinkpad and be done with it.

Cheers ^:)

P.S.: edited that not so discreet smiley. :-)

Barkuti, pretty sure an animated smiley flippin the bird is in direct violation of forum rules on profanity. You might wanna check that…

Here’s where I get a little confused with electrical arrangements in LEDs.

A common 5.6mm laser diode-can, can be powered over the ground pin, which is sometimes (-) and sometimes (+). As an example, 405nm diodes use a boost driver, 18650 for power, and the body would be negative just like a flashlight, with a diode vF of 5-6V. OK, it’s common practice to solder the negative and ground pin together on the same wire, positive pin on other wire of course. Clearly the driver still has to complete the circuit for power to flow. Why can’t an LED resolve the same outcome? Where negative could be your DTP. :question:

Anyways, I would certainly love to try these two LEDs. It’s incredible what the 219C is rated at on the spec sheet (326 lumens, I think), vs what it can actually output. Just think about the possibility of the other two then, when cooled correctly. The 144AM rated at 1300-1400 lumens at highest bin; I could only imagine a single mono die LED with moderate CRI potentially unleashing 3200-4000 lumens under a smooth reflector.

In the meantime, are they seen for sale anywhere? I would not mind doing an output test on a home-made copper board, like I did for the Luxeon Z test.

A little more 144A info:

- Emitting area: 4x4mm (16mm2) - same configuration as a XHP35 (with lower forward voltage)?

- Aluminium Nitride substrate (285 W/(m·K))

  • Gold contacts

Alumininium nitrite looks like the way to go creating a board for mounting these… it should be reasonably cheap to have someone make a bunch of 16 and 20mm disks… then glue two pieces of copper sheet to the top to act as electrodes. Done.

Welt Electronics sounds like one possible source.