I think the point of using so many LEDs at low current is about thermal management. On incans replacement, the main problem is to achieve junction temps low enough given the limitation on heatsink surface area. By running each LED at lower current, you can run them on a hotter heatsink while keeping Tj low enough to guarantee a lifetime over 25Kh. Aditionally, it achieves higher efficiency, thus reducing the total watts required to get the target light output and more important with the limited heatsink area, the total thermal load.
Philips 75w equivalent LED lamp, 1100 lumens
- Thread starter PhotonWrangler
- Start date
Help Support Candle Power Flashlight Forum
blasterman
Flashlight Enthusiast
- Joined
- Jul 17, 2008
- Messages
- 1,802
I know that spreading lower powered LEDs round -vs- fewer higher powered ones in theory keeps the on-die temp lower, but I'm still skeptical of this. Your heatsink efficiency, if run long enough, is ultimatley a factor of energy in -vs- radiating effiency. Using a dozen lower powered LEDs doesn't allow you to cheat because the total energy being handled by the sink is the same. The heatsink will heat up to the exact same temp and present each LED with the same surface heat.
However, that assumes were talking about heatsinks with a decent base thickness. If the heatsink is too thin then heat generated by the LED's 'pools' up in proximity on the heatsink because it can't move away fast enough. In this respect using more LEDs and lower wattages might have a thermal coefficient via the limited convection of one of these bulbs to buy you a few degrees. Given these bulbs run really, really hot that might be critical.
Given the increasing linearity of high powered LED efficiency I'm not sure if the improved efficiency would amount to much, but not sure how much a difference we're talking about. Possible they are using many smaller LEDs because they are that much cheaper.
However, that assumes were talking about heatsinks with a decent base thickness. If the heatsink is too thin then heat generated by the LED's 'pools' up in proximity on the heatsink because it can't move away fast enough. In this respect using more LEDs and lower wattages might have a thermal coefficient via the limited convection of one of these bulbs to buy you a few degrees. Given these bulbs run really, really hot that might be critical.
Given the increasing linearity of high powered LED efficiency I'm not sure if the improved efficiency would amount to much, but not sure how much a difference we're talking about. Possible they are using many smaller LEDs because they are that much cheaper.
Its clear that the heatsink will heat the same at the same heat load, if its well designed. But in thermal management, there is a near fixed figure which not depends of the heatsink temperature, the thermal resistance of the LED package. From solder to junction, the temperature raise is linear with the power of the LED.
So for a given heatsink temperature, you get lower junction temps using more lower power LEDs than fewer more powerful ones. The Rebel has a thermal resistance solder to junction of ~10 ºC/W. If you use one at 700mA, it has a typical power of 2.24W (0.7A*3.2V), thus junction temp is about 22.4ºC over heatsink. But if you use 2 Rebels running at 350mA, each one runs at just 10.5ºC over heatsink temp.
In this lamp, Rebels runs yet below 350mA, so the junction temps are below 10ºC over heatsink temp, allowing heatsink to run really hot without compromising lamp lifetime (at least, considering just LEDs reliability).
It is true that lately linearity of emission at each current is way better than in the past, but it not translates directly to similar efficiency, as efficiency is a function of total power, not just current. As current-voltage curve is flatter at lower currents, the gain on efficiency is not just achieved by the better efficacy emitting light at lower currents, but by the reduced voltage. For example, in the Rebel, still if the emission would be completely linear from 350 to 700mA, you get a gain on efficiency of 7% just due the lower voltage. As emission is not completely linear, the whole gain on efficiency is very noticeable, from 10 to 20% (it depends of the thermal management, as different power affect temperature, and temp affect efficiency) depending of the exact current compared.
I guess by using Rebels below 300mA as Phillips designed this lamp, they get an efficiency gain of minimun 10% as compared to higher currents (not 700mA, but lower), allowing to use less total watts and run junction temps low enough to get very good lifetime (if driver is designed accordingly).
It would be very interesting to know the heatsink temperature of the lamp after several hours of operation. We could make some guesses about expected degradation with time knowing it, and confirm if they were in the limit of heat load when decided touse so many LEDs. Maybe they just used so many because it is cheap for them and they want this lamp works as the flagship of LED lamps into the residential market, by using as little watts as possible to get the target output and with excelent reliability.
True 75W incan replacement with 17W and true long lifetime is the way most consumers accept the technology: 40$ is not excessive to try, for example on a place were the lamp is working many time but with frequent switching on/off. I dont see that price excessive at all, although of course it should drop with time to a price competitive with CFLs.
So for a given heatsink temperature, you get lower junction temps using more lower power LEDs than fewer more powerful ones. The Rebel has a thermal resistance solder to junction of ~10 ºC/W. If you use one at 700mA, it has a typical power of 2.24W (0.7A*3.2V), thus junction temp is about 22.4ºC over heatsink. But if you use 2 Rebels running at 350mA, each one runs at just 10.5ºC over heatsink temp.
In this lamp, Rebels runs yet below 350mA, so the junction temps are below 10ºC over heatsink temp, allowing heatsink to run really hot without compromising lamp lifetime (at least, considering just LEDs reliability).
It is true that lately linearity of emission at each current is way better than in the past, but it not translates directly to similar efficiency, as efficiency is a function of total power, not just current. As current-voltage curve is flatter at lower currents, the gain on efficiency is not just achieved by the better efficacy emitting light at lower currents, but by the reduced voltage. For example, in the Rebel, still if the emission would be completely linear from 350 to 700mA, you get a gain on efficiency of 7% just due the lower voltage. As emission is not completely linear, the whole gain on efficiency is very noticeable, from 10 to 20% (it depends of the thermal management, as different power affect temperature, and temp affect efficiency) depending of the exact current compared.
I guess by using Rebels below 300mA as Phillips designed this lamp, they get an efficiency gain of minimun 10% as compared to higher currents (not 700mA, but lower), allowing to use less total watts and run junction temps low enough to get very good lifetime (if driver is designed accordingly).
It would be very interesting to know the heatsink temperature of the lamp after several hours of operation. We could make some guesses about expected degradation with time knowing it, and confirm if they were in the limit of heat load when decided touse so many LEDs. Maybe they just used so many because it is cheap for them and they want this lamp works as the flagship of LED lamps into the residential market, by using as little watts as possible to get the target output and with excelent reliability.
True 75W incan replacement with 17W and true long lifetime is the way most consumers accept the technology: 40$ is not excessive to try, for example on a place were the lamp is working many time but with frequent switching on/off. I dont see that price excessive at all, although of course it should drop with time to a price competitive with CFLs.
patrickhuang
Newly Enlightened
- Joined
- Feb 8, 2012
- Messages
- 21
Additionally, lm per w of LED chip is upgrading every season, maybe Rebel in 75w bulb(2012) is better than Rebel in 60w bulb(2010).
I would add to Kinnza's excellent explanation that more LEDs at lower currents may also be a matter of economics. Whether Lumileds is a Philips company or not, someone still needs to pay for those LEDs either in direct costs or lost sales.
It may be cheaper to run 18 lower binned parts at lower currents to achieve effiency X, then to run fewer higher binned parts at higher currents to get the same effiency X. Premium bins command a premium which may be beyond the benefit of using less.
Semiman
It may be cheaper to run 18 lower binned parts at lower currents to achieve effiency X, then to run fewer higher binned parts at higher currents to get the same effiency X. Premium bins command a premium which may be beyond the benefit of using less.
Semiman
Theatre Booth Guy
Enlightened
Yup, they are bright. I'm concerned about the content of UV with the yellow covers removed.
don.gwapo
Enlightened
+1 coz i'm using a led bulb too with the white frosted cap removed. I'm using 40w equivalent @ 510 lumens with 18 leds in it.
