Philips Lumileds has announced a new high-brightness white LED, the "undomed" Luxeon MZ. It looks like a useful part, but the price is out of scale.
Lumileds (sponsor of this site) announced the new HBLED within a week of introducing their 200lm/W LUXEON Rebel ES Lime and LUXEON Z Lime LEDs, which we covered. This newest addition to the Lumleds portfolio, with a 4.2mm diameter emitting surface, claims to enable customers to create tighter beam angles with higher lumen density at a lower cost than was previously possible in a 600-1000 lm device, with CRI of 70, 80, 90 and a CCT range of 2700K-6500K.
Lumileds product families.
(Source: Lumileds)
The lumen levels are achieved in the 'MZ by incorporating four series-connected, 2.8V VF die in a 4.2 x 4.0, LUXEON M-compatible package. Lumileds claims the 'MZ has all of the features of LUXEON M, including an identical solder footprint, and that it allows for tighter beam control and higher punch due to a smaller apparent source size, according to Lumileds Senior Vice President Emmanuel Dieppedalle:
With the LUXEON MZ in combination with LUXEON M, designers can create two solutions with one optic and one PCB, thereby lowering overall costs. Our customers are already designing platforms with LUXEON M and LUXEON MZ to create multiple SKUs using less engineering resources.
Luxeon MZ.
(Source: Lumileds)
The ace being played several times in Philips's marcom materials is better optical control and lumen density due to a smaller source size than is found in a domed emitter: "In an MR16 the LUXEON MZ delivers greater punch, achieving 50,000 candela center beam candle power (CBCP) compared to encapsulated emitters that only achieve half the beam intensity using the same sized optics."
50,000 cd is a very impressive CBCP. Looking at the VF, the number of die, the package footprint, and the lumen output specs, they are similar to that of the encapsulated LUXEON M, which was introduced at Strategies in Light in January 2012. The dies are smaller, with a 2mm per side die in the 'M. It looks like maybe a half-area die in the 'MZ at about 1.4mm per side -- hence the "throw" and CBCP claims. What's most amazing is that Lumileds has eliminated the silicone dome of the 'M with the "domeless" 'MZ, yet the polar radiation patterns of the two products are indiscernible from each other in their respective datasheets.
Chip in Frame architecture.
(Source: Lumileds)
Thermally, the LUXEON M used Lumileds's Chip in Frame Architecture with a 2x2 array of 2mm2 thin-film flip chip die mounted on an high thermal conductivity aluminum nitride paddle. Unless there's a copy/paste error in the datasheets, the thermal resistance between the 'M and the 'MZ is identical at 1.25°C/W, suggesting that the substrate may be AlN only, as the thermal resistance with smaller die needs to be better than that if they had used alumina. The "paddle" location in both the M and MZ is metallized, providing a nice thermal path that's electrically isolated from the 11.7 V anode and cathode electrodes. This isolated pad is probably the reason this part will continue to exist, irrespective of the availability of chip-sized devices like the 2.2 mm2 Luxeon Z.
(Source: Lumileds)
Despite the attention to thermal details, and the higher current density in the half-area sized die, the MZ is expected to maintain 70% lumen levels (L70) at 50,000 hours of operation at a forward current of up to 700 mA for the LMZx-Sxxx part code. This projection is based on constant-current operation with junction temperature maintained at or below 135°C. With the parts hot-binned at 85°C, there will, of course, be a lumen depreciation, by running at 135°C, of about 10%. The latitude of running 85°C above a 50°C ambient spec with an 8W LED, makes it a walk in the park to come up with a 10°C/W cooling solution.
In summary, the LUXEON MZ looks like a nice part. Its typical efficacy of 95lm/W (3000K, 80CRI) could be better, but it's not surprising given the half-area die is being ridden hard. Looking at the product family chart above, the 'MZ is there -- that chart was shown at a conference back in September of 2013, so the part's been around for about six months. With that, a quick disty check to buy some samples to play with revealed no stock at any of them and a 14-week lead-time, which is very disappointing since the unspoken rule in semiconductor marketing is to ensure there's at least a decent supply of sample stock in disty when a part is announced. But, even that is not the knife in the back for the LUXEON MZ.
Plessey is at $1/kLumen now. Philips has been cited as predicting pricing at $1/kLumen this year. Haitz's Law says LED prices drop by a factor of 10 in 10 years, so in the two years since the LUXEON M came out, the 'MZ should be 36% lower in price than the 'M. The LMZ8-SW30 is $4.67 each in quantity 500 at Digikey, or $5.98/kLumen. The LXR8-RW30 is $5.67, or 18% lower (a "half-Haitz"). Counting optical losses and running it at a bit more efficacious power level, that says it's $9.35 for two LEDs and the rest of the light bulb's components have to be free in order to sell it at Home Depot, which itself will need to do the selling for free as well. How is that supposed to work when you're yielding twice the number of dies per wafer compared to an 'M?
Related posts
No comments:
Post a Comment