200 Lumen/Watt LED announced

I thought the statement towards the bottom:

"Cree® XLamp MK-R LED samples are available now and production quantities are available with standard lead times. To learn more, please visit

means that they ARE commercially available. Though probably not yet retail.

You're right - the scientific development was a couple of years ago, and they are now announcing commercial availability.

What I meant to say was when will they be in the sheds.

I reckon the big issue will be how long it takes for the existing stocks to be flogged... :-)

hopefully spelling the end of the ballast-weight million candle power torches, with lead acid batteries that last a year at most and cost more than the whole torch to replace. A 120 lumen Cree torch is just as bright and last ages on three alkaline D cells.

rusty

No one has managed to design a lamp which can run LED junctions at less than 100-125C. That's the problem - LED efficiency drops significantly as the junction temperature rises, and the manufacturers always quote efficiency at the unachieveable 25C temperature.

LPS get more efficient as they get hotter. Their problem is that as they get hotter, they produce less waste heat to make them hotter! That's why they operature in a thermos flask with an infra-red reflective coating on the inside. In the lab, they have got up to

So far, most power LED technology has only become viable in the retail sector as patents have expired and there is competition, because it's very price sensitive. They become viable in the in commercial sectors much faster, which is not so price conscious, or at least will weigh acquisition cost against total cost of ownership.

Indeed, many councils leapt at things like LED streetlighting far too soon and without quality assessments, and have found the supposedly long life maintenance-free LED lanterns they bought lasted less time than the lamps did in their previous lights. Some have now backed away and there seems to be quite a move now to CFL.

I got this for Xmas:

Do they think all that corporate-speak is going to impress or piss off readers of their press offerings? It makes me wanns puke. Why reinvent English in such a way that nobody really knows what you mean. after a quick trip to Wikipedia, I think it ought to say, Hey guysm, we been very quiet about this here new more efficient led, so we have decided to hype up a few advanced prototypes to show you what will be coming out. However don't look at the price and reliability just yet, we need to get those sorted out before we sell them.

brian

You're right. The press-release quotes "up to 200 lumens-per-watt (at

Later on, they say "The MK-R LED features a 7 mm x 7 mm footprint with a 6 mm optical source and delivers up to 1600 lumens at 15 W, 85°C". So that's a real-world figure I guess. That's just over 100 lm/W, which is still pretty good but nowhere near 200 lm/W.

And is going to need a chunky heat sink and some pretty low thermal resistances from that tiny 6mm source to dump 15W of heat...

Is active cooling for LEDS the way forwards? Peltier or a proper heat pump?

Not quite 15W of heat, because the efficiency will be high enough that a decent proportion of that 15W will be in the form of light. But I guess most of it will still be heat.

Will the increasing efficiency of LEDs help solve the heat problem? As efficiency rises less energy is needed for the same amount of light, and a lower proportion of that energy ends up as heat.

They can dissipate far more than that on CPUs, even with passive cooling. It just means bolting the LED to a thickish ali housing with fins on it. You could even fit a fan but I doubt its needed.

My server CPU is passively cooled it's a lowly 800MHz Pentium III, the lump of finned ali to do this is about 6 x 2 x 1.5" inside the forced ventilation (power supply fan) case of course...

The fan would take a significant amount of power, screws up the efficiency something rotten... Not to mention the noise and of course yo= u have to squeeze all this into the space and shape a standard GU10 or maybe an R80 spot light takes. And its still a fairly unidirectional source of light, not the near omnidirectional light you get from a CFL o= r GLS bulb.

My heavens, an echo, echo, echo ...

not really, Until you get up well above 50% efficiency the heat to be dissipated doesn't change much.

However 15W from a 6mm die is no big deal.

Yes. A 100% efficient white light source is around 650 lumens/watt, (but it depends heavily how you work it out, and there are many different ways).

So when LEDs start reaching 300-400 lumens/watt, the decreased heat output will start making a difference.

It is if you're trying to keep the junction temperature at the 25C needed for the 200 lumens/watt :-)

I have a Maglite clone from 20 years ago - I think it's a Streamlite, but not sure. The head looks to be same size as a genuine one (52mm), so how do these drop-in upgrades work? Do you have to entirely gut the old head and re-use the glass? Is it reversible?

Partially gut and it is reversible.

You take out the reflector and bulb and screw this module onto the bulb retaining ring thread. Chainge the o ring between the glass and the end housing ring (for slightly increased spacing and screw together. Takes 2 mins. If you keep the reflector bulb and old o ring, it is reversible.