LED-Streetlight by 5mm leds

[Juctuc]

I like to know your opinion to that kind of streetlight.One of the lights is 48w,and there is 764 leds in it.Factory is telling that there is over 3000lm output...No ballast,no transformer,no resistor.Just some "thing" which they have patented.But im i wrong,if i say:

5mm led is not long life product
764 leds need large area where to put them,so,3000lm is wrong way to compare it for some other lights like HQL.For example: I get 100lm from ledbar which is 1 meter long,but i cant say that is same as the 10w halogen.That 100lm is all the leds together in 1 meter long bar....

So...if i REALLY say that i get 3000lm from that streetlight,i need to put those 764 leds together in same place where is just one led.

Those leds are 20deg each,the pole is 8 meters..But you cant get much light to the street from it.Even if its really 3000lm

Im sorry about my english.I hope you understand what mean.But anyway,what you are thinking about this kind of light?

[jtr1962]

That comes to about 4 lumens per LED, so 3000 lumens is possible. I've tested dozens of 5mm LEDs. At this point 4 lumens at 20 mA seems to be average for the low-cost commodity LEDs. At $0.07 each in large quantities, that's probably about $55 for the LEDs. Add in perhaps another $20 for the PC board, connectors, etc., and it still comes out way cheaper than a discharge lamp and ballast. As for light output, a 20° beam angle 4 lumen LED will give roughly 15 lux at 1 meter head-on (much less off-axis). 764 of them will give about 11,500 lux at 1 meter, or only about 180 lux at 8 meters. I never bothered measuring the lux under any sodium vapor lights but I suspect it is more.

Is this a good solution? I don't think so. They need to drive the LEDs at 20 mA or close to it to get 3000 lumens. This means a lifetime of at most 10,000 hours to 50% of initial brightness. It may be much less depending upon the LEDs used. 5mm LEDs are great for small lighting projects where the cumulative run time is likely to never reach more than a few thousand hours. You can get much longer lifetimes than this by running the LEDs at a few mA and providing fat PC board traces. The downside is that you need several times the number of LEDs.

A better solution for the same output would be to use maybe about 30 to 50 Q4 or Q5 Cree XR-Es, or a similar number of Lumileds Rebel 100s. Driving them at 350 mA with good heatsinking will give in excess of 50,000 hours before noticeable dimming occurs. This is roughly 12 years (assuming 12 hours per night use). Power consumption for 3000 lumens would be roughly 35 watts. Cost would be more, but still probably in the same ballpark as a discharge lamp and ballast.

Finally, it needs to be said that 3000 lumens is probably not enough for a streetlight unless you decrease the spacing between lamps. Most sodium vapor streetlights are in the 20,000 to 40,000 lumen area. Granted, probably about half of this is lost in the reflectors and stray light, but this still means a more inherently directional LED solution must provide 10,000 to 20,000 lumens. Of course, since the whiter LED light appears much brighter than the 2000K sodium vapor you may be able to get by with only half that, but I'd say 5000 lumens is an absolute minimum, with 10,000 being better.

Regardless of whether this is a good solution, my analysis does show the promise of LED streetlighting. At 200 lm/W combined with better directionality you can replace a 400W (without ballast losses) sodium vapor lamp with perhaps 50 to 75 watts worth of LEDs. Even using today's less efficient LEDs you can still cut the power consumption in half while giving the same or more light, and a better quality light at that (sodium vapor lamps cause tunnel vision).

[Juctuc]

Many thanks Jtr 1962!! There is a street here in Finland where that light is on.So i was there just hour ago.I get about 15lx right under the pole.(about 7 meters from the light)Its about the same as HQL but...There is dark area between the poles.So i think its not possible to make good streetlight without any optic.

Anyway...i dont know how they drive those leds..is it 20mA or what is it.There is 64led connected together,so they get voltage of about 230V,so the light is so called direct connect.This patented "thing" have to be very smart,because of the voltage difference between the leds.I really dont understand how its even possible to drive those led right without any resistor or something.

There is interesting product in the CREE-website.80pcs XR-E.8lines of leds,10 leds in every line,and those lines have different kinds of optic.

[jtr1962]

I measured 16 lux right under one of the local streetlights. This drops to 8 lux at 10 meters. Assuming then that it decays to 0 lux at 20 meters, and drops linearly, that's ~3750 lumens in the 10 meter circle, another 3750 lumens in the circle from 10 meters to 20 meters, total approximately 7500 lumens. Considering that the 250 watt sodium vapor lamp gives on the order of 25000 lumens, that's a pretty bad fixture efficiency. Only 30% of the light falls in a useful region. An LED fixture which loses only 15% in the optics would only need about 9000 lumens to light at the same levels. Running some Q5 Crees at 350 mA would require only about 95 watts to equal the 250 watt sodium lamp. Ballast losses would be less as well-about 5% to 10% for a decent LED ballast versus 15% to 25% for typical discharge lamp ballasts. Overall I think a well-designed LED fixture could cut power usage by two-thirds.

[Juctuc]

Yes,you are right what you are saying.Its possible to make good product for streetlight.I have not find that yet.I did find something with powerled.

Lamps flux: 4200lm (Tj60,Ta25)
System consumption: 75w
Illumination:15lx (8 meter pole)
Illuminated area: 26x10meter(8meter pole)

Problem with that ,is the dark area between the poles.If the poles are in every 35meters there is about 9meters of dark area.

I did measured some 125w HQL.There was about 3lx in the middle of the 2 poles.The lengt between the poles is 35meter.

[JohnB]

I have some friends that are involved in LED street lights and they have fixed the two biggest problems with street lights. Beam Spread as you mentioned and the heat management to keep the chips from melting. It seems that there are many companies that are very close to solving these issues. Once the technology has matured I am sure the larger players will move in and buy them out. LED streetlights seem to be still in a pioneer type stage. I have a prototype spotlight that they built it puts out 50 watts and it is very bright. The chip array is 19 across and 17 down.

John