LED Lighting & Energy Efficiency

[TEG]

Both of those stories show the 'traditional' 5mm LEDs.

From my experience those don't make good home lights.
They tend to be too directional, and in many cases too blue looking.

Lots of bulbs you can find on eBay and such just use clusters of these 5mm LEDs.

I wouldn't buy anything like that except for maybe a porch light where dim bluish light could be OK.

 

[TEG]

My house is currently about 1/3 Halogen (in the kitchen mostly), 1/3 fluroescent, and 1/3 LED bulbs. The garage is all LED.

After trying different types of LEDs, I concluded that the "SMT" (surface mount technology) is the way to go. They have better color spectrum, wide angle (flood instead of spot), and tend to be brighter.

For instance, bulbs that looks like this:

 

[TEG]

Note, they tend to offer them in 'cool' white or 'warm' white. The cool tend to be in the 5000-6000K color spectrum, and warm in the 3000-4000K. Although not quite as bright, I think the warm white variants are more 'friendly' looking and have a more pleasant lighting effect.

You can actually look at the chips and if they look more yellow they are 'too cool', and if they look more orange they produce a more natural looking light.

Yellow:

vs
Orange:

 

[doug]

Both of those stories show the 'traditional' 5mm LEDs.

From my experience those don't make good home lights.
They tend to be too directional, and in many cases too blue looking.

I wouldn't buy anything like that except for maybe a porch light where you want dim bluish light could be OK.

The main point of the articles dpeilow linked is likely lost since they were written for the general public by non technical folk and they focus too much on the efficiency gains of LEDs in general. The achievement by these scientists is in the manufacture GaN, the wide bandgap semiconductor material that makes higher photon energy (shorter wavelengths such as green, blue, and UV) LEDs and lasers possible. The researchers put them in the standard 5mm packages just to demonstrate that their process works. Focusing on the package misses the point.

I haven't seen a proper journal article, but I can guess what they did. Typically these materials are grown in a vacuum chamber using MBE (molecular beam epitaxy) on a particular substrate. The article says they're now able to use Si substrates instead of more expensive sapphire. You normally choose a substrate that will not have too much of a lattice mismatch (has a similar lattice constant) with your crystal growth. Either they're putting down a buffer layer, or they've found that the stress/strain at the crystal interface is not a problem for them.

 

[TEG]

Another thing about LEDs... It is somewhat of an imperfect science to manufacture them, so they end up with "batches" and "bins" of various quality. Basically you make a bunch then you test them to sort the good from the bad and then sell them by how they turned out. A particular factory is likely to produce a variety of quality levels and rejects.

The same thing tends to be true with a lot of industries, from making LCD panels, to growing fruit.

So the "you get what you pay for" rule definitely applies. If you buy a very cheap LED light it is likely to use older technology and/or low grade LEDs.

 

[doug]

The color of "white" LEDs tends to have more to do with the phosphors in the package than the LED itself. Of course LEDs by their nature are monochromatic since the photons they emit are from electrons making a discrete energy transition across the crystal band gap (nu = Eg / h).

Low cost single emitter white LEDs use a blue LED with yellow phosphors in the package. Some of the blue photons are absorbed by the phosphors and are re-emitted as broadband yellow (in a process called florescence). The blue and the yellow mixes in a way the eye perceives as white. This mixing is often imperfect due to poor spacial overlap and/or inadequate phosphor absorption, and if you shine the LED on a piece of paper you can see a blue hot spot in the center with a yellow hallow. Also these LEDs are inappropriate for home lighting since they are not actually white (white meaning a continuum of wavelengths throughout the visible spectrum) and can make room objects look strange (since the light that gets to your eye is now convolved with the reflection functions of various materials, revealing that the source has spectral components).

A better way is to use a wider bandgap UV LED with white phosphors. In which case the LED is functioning in much the same way a standard florescent light does.

Much of the R&D on LEDs (and CFLs for that matter) intended for home and office lighting situations focuses on the the proper choice and mixture of phosphors to produce illuminations that consumers are comfortable with. This might involve the combination of several types of LEDs into a single fixture.

 

[TEG]

Thanks, Doug. I know a lot of that but you do a much better job describing it than I could. The "litmus" test for how good home LEDs work is how skin and food looks. We seem to be very observant when it comes to that. If your housemates suddenly look sickly when they walk into an LED lit room you know those bulbs are not good enough. On the other hand if it is just a reading light, or accent light on something inorganic then it isn't so critical. There is an industry measurement scale called "CRI" ("Color Rendering Index") that can be used to determine how good a light is in this regard. Most manufacturers don't even bother to publish a CRI number, but if they do and it is above 85 you are probably getting a good light. Another term to look for is "full spectrum".


LED Technology Glossary :: LED Technical Reference

CRI or Color Rendering Index - The calculated rendered color of an object. The higher the CRI (based upon a 0-100 scale), the more natural the colors appear. Natural outdoor light has a CRI of 100. Common lighting sources have a large range of CRI.

Characterizing Color Properties of LED-Based Reading Lights | Solid State Lighting | Programs | LRC
High cri led lamps utilizing single phosphor invention
LED lamp - Wikipedia, the free encyclopedia
High CRI phosphor blends for near-UV LED lamps
LED lighting CRI and the Lighting Revolution | HolidayLEDs.com

 

[TEG]

One other thing - people get used to certain kinds of light. So sometimes you change a bulb and someone goes "that looks wrong", but given a little time they get used to it and in some cases start to like it better.

 

[doug]

A figure of merit like CRI is a step in the right direction. I never liked color temperature as a metric for LEDs since clearly this is not black-body radiation.

 

[TEG]

Without buying and comparing everything, you have to make do with what published specs they use now.


Example of what I would consider a "bad" LED bulb:

• 80 lumens
• color temp 5500K 'cool white'
• CRI 72
• 60 degree spot beam angle

compared to a good one:

• 150+ lumens
• color temp 3000K 'warm white'
• CRI 90+
• 120+ degree flood angle

 

[TEG]

Related to the story dpeilow posted:

Cheap, super-efficient LED lights on the horizon - tech - 29 January 2009 - New Scientist

 

[doug]

Cheap, super-efficient LED lights on the horizon - tech - 29 January 2009 - New Scientist

Thanks, so it looks like they're using an AlGaN buffer layer.

Cracking up

The cost of production has kept the LEDs far from homes and offices, however. Gallium nitride cannot be grown on silicon like other solid-state electronic components because it shrinks at twice the rate of silicon as it cools. Crystals of GaN must be grown at 1000°C, so by the time a new LED made on silicon has cooled, it has already cracked, rendering the devices unusable.

One solution is to grow the LEDs on sapphire, which shrinks and cools at much the same rate as GaN. But the expense is too great to be commercially competitive.

Now Colin Humphreys's team at the University of Cambridge has discovered a simple solution to the shrinkage problem.

They included layers of aluminium gallium nitride in their LED design. These layers shrink at a much slower rate during cooling and help to counteract the fast-shrinkage of pure gallium nitride. These LEDs can be grown on silicon as so many other electronics components are. "They still work well as LEDs even with those extra layers inside," says Humphreys.

 

[Tdave]

Without buying and comparing everything, you have to make do with what published specs they use now.


Example of what I would consider a "bad" LED bulb:

• 80 lumens
• color temp 5500K 'cool white'
• CRI 72
• 60 degree spot beam angle

compared to a good one:

• 150+ lumens
• color temp 3000K 'warm white'
• CRI 90+
• 120+ degree flood angle

Does anyone know of an online source to buy good ones?

 

[TEG]

Searching for "SMT LED" on eBay will find some.

PAR20 White LED Flood Light 9 SMT SMD 110 240 V Wide - eBay (item 320305332935 end time Feb-24-09 20:39:20 PST)

Brightest 120 Lumen G4 Base 10 SMD LED Bulb 12V AC DC - eBay (item 220318354160 end time Feb-22-09 16:43:28 PST)

Through Amazon too:

Amazon.com: PAR38 White LED Flood Light 30 SMT 110 240 Volt - Bright LED Flood Light Color: Warm White: Home & Garden

Amazon.com: PAR30 White LED Flood Light 20 SMT 110 240 Volt - Bright LED Flood Light Color: Warm White: Health & Personal Care

Amazon.com: PAR 20 9 SMT LED (White/ warm white): Home & Garden

Amazon.com: Mr16 12 Volt 9 SMT Warm White LED Wide Angle Light Bulb: Home Improvement

 

[SByer]

LEDwaves.com - I have a 2-gen-old Par-38, really happy with it. Current Par-38 bulb hits 1000 lumens in warm white.
Lighting Science Group Corporation

http://www.theledlight.com

 

[TEG]

Another indicator to track are watts used. For low power under counter accent lights 1watt can be enough, but for wide area lighting you want 3+ watts. With that said many bulbs in the 2-3 watt range vary greatly in their output. Different LEDs are more or less efficient, and bulb optics can play into that as well.

If you aren't careful you can end up with a bulb that is too dim and/or too blue and/or or too directional.

 

[TEG]

By the way, this guy's effort to convert his house to 100% LED is interesting:

12V LED Lighting

He did this a few years ago when off the shelf options were more limited. He improvised a lot of custom fixtures.

 

[vfx]

There are lot's of these but the price is new.

Inhabitat » Ikea Unveils Solar Powered Lights for Summer!

Should this be under solar?

 

[TEG]

A science fair project we did a while back:

 

[vfx]

This company came in as an email directly to me. Anyone know them?
http://www.yjbled.com/screwinbulbs.php