Well CREE do produce LEDS with a CRI of 90 but as you said the efficiency takes a hit. I think this is just incorporating more red emitting phosphors into the mix as explained in the article provided by Scott.
The R2 bin for the standard XP-G LEDs @ 80 CRI has only come out recently where you get 114 lumen at a CCT of 3k. By contrast the maximum flux bin for the 90 CRI LEDs is only the Q2 bin at 87.4 lumen. This is quite a sizeable difference but the R2 bin has only been around for a couple of months and until then we were looking at the Q5 = 107 lumen.
87.4 > 107 lumen is quite a lot and represents the difference between one generation and the next. Still, if they manage to push efficacy up to around 150-160 lumen for a standard 80 CRI LED, 120-130 lumen for the 90 CRI version isn't too bad.
Of course in CREEs own fixtures they produce the light using a mixture of LEDs throwing in red along with the mix to increase the CRI of the product.
Remote phosphors seem like a very intriguing idea although I can't see them being easily incorporated into standard LED packages. I do wonder how large one needs the individual lighting element to be to do this however. Most LEDs are tiny, yet require large external mounting components to get rid of heat so really one could increase the size of the LED considerably without really affecting the end product design.
The R2 bin for the standard XP-G LEDs @ 80 CRI has only come out recently where you get 114 lumen at a CCT of 3k. By contrast the maximum flux bin for the 90 CRI LEDs is only the Q2 bin at 87.4 lumen. This is quite a sizeable difference but the R2 bin has only been around for a couple of months and until then we were looking at the Q5 = 107 lumen.
87.4 > 107 lumen is quite a lot and represents the difference between one generation and the next. Still, if they manage to push efficacy up to around 150-160 lumen for a standard 80 CRI LED, 120-130 lumen for the 90 CRI version isn't too bad.
Of course in CREEs own fixtures they produce the light using a mixture of LEDs throwing in red along with the mix to increase the CRI of the product.
Remote phosphors seem like a very intriguing idea although I can't see them being easily incorporated into standard LED packages. I do wonder how large one needs the individual lighting element to be to do this however. Most LEDs are tiny, yet require large external mounting components to get rid of heat so really one could increase the size of the LED considerably without really affecting the end product design.
I bought some remote phosphors and had PCBs made to hold SMD LED arrays to play around with, just waiting for them to arrive. I could post my results on how they turn out.
I don't think it needs to be big at all.
At work, I once knocked the lens off a LED by accident listed in the BOM as an LXM3-something (which I just found out is a Luxeon Rebel) and ended up with a LED emitting blue light. Obviously the phosphors weren't attached to the die, but the lens. It's actually quite easy to knock the lens off a Rebel (probably the reason why the datasheet mentions not to handle the Rebel by its lens).
Wouldn't that mean that the Rebel, in a way, is a remote phosphor LED?
Last edited:
To answer myself: no, the Rebel is not a remote phosphor LED. I just saw the cross section drawing, and it has a phosphor plate on top of the LED-chip (like figure 5 in Mike Wood's article).
The plate must have gotten stuck to the silicone lens on the LED I damaged.
Looking at figure 7 of the same article, I can't help but think remote phosphor LEDs don't need to be very much bigger, though...
The plate must have gotten stuck to the silicone lens on the LED I damaged.
Looking at figure 7 of the same article, I can't help but think remote phosphor LEDs don't need to be very much bigger, though...
The trouble with car headlamps is there's nowhere for the heat to escape. This is why they are in expensive cars because I am guessing that you effectively replace the entire headlamp unit rather then just a bulb.
Indeed Jitter, some of the older LEDs were much larger in comparison to the tiny affairs that CREE now produce. I wouldn't mind betting that you could increase the size of the lens a little and have a remote phosphor on the inside of it, especially if you're using a tiny 3x3mm blue LED to power the thing. But then, what do I know? Sounds like a good idea from the uneducated point of view.
Indeed Jitter, some of the older LEDs were much larger in comparison to the tiny affairs that CREE now produce. I wouldn't mind betting that you could increase the size of the lens a little and have a remote phosphor on the inside of it, especially if you're using a tiny 3x3mm blue LED to power the thing. But then, what do I know? Sounds like a good idea from the uneducated point of view.
Still more expensive than the average car, but no supercar: the first series-production LED headlamps are to be found in the Lexus LS600h (2008-model), according to Wikipedia (Headlamp - Wikipedia, the free encyclopedia).
But as always, innovations are to be found in the most expensive cars first and then find their way down the line into the cheaper models.
I'm looking to replace the lighting in my fish tanks with LED's. Also just want to screw around with them for accent lighting in the house. Can anyone recommend a good supplier?
Also any recommendations on good driver manufacturers would be appreciated as I see lots of reviews of the led's but very few on the drivers.
thanks,
revb.
Also any recommendations on good driver manufacturers would be appreciated as I see lots of reviews of the led's but very few on the drivers.
thanks,
revb.
It depends how DIY you want to get with all of it. Farnell stock a large number of the raw LEDs, but you can't do much with them unless you've got the ability to make your own PCBs and in such a way as to carry the heat away from the LEDs themselves. I do this by drilling a hole through the PCB underneath the LEDs heatslug and then push a suitably sized copper rod through the hole. When soldering the LED into place using solder paste + hot air gun/oven, the heatslug solders to the copper rod (1.5mm dia jewellers wire) and the copper rod then carries the heat through to the reverse side of the board where it can be filed down/soldered to the reverse copper plain and then interfaced to a heatsink/metal enclosure.
This works well but is fiddly.
The other option is to buy LEDs that are already soldered onto aluminium backed PCBs. LEDrise is the only company I've seen that sell a variety of LEDs ready soldered to suitable PCBs.
LedRise - High Power LEDs
There will no doubt be others too.
Then there's CREEs CXA2011 LED array that comes ready made on an aluminium mounting pad. This packs a lot of light into an easily mountable and relatively small area. It is nowhere near as efficient as say the XML LEDs, but this is only when driven at high currents, at low currents the CXA starts to become a lot more energy efficient. The CXA however is a lowish current high voltage part which isn't suitable for off the shelf LED drivers I'd bet.
I build all of my LED drivers, I've used single chip solutions from maxim, linear tech and zetex and they all work well. Maxim make a pretty neat little chip called the MAX16832 which is one of my favourites. It comes in a handy 8 pin SOIC case so is easier to manage then most of the other LED drivers and it also only needs very few supporting components, it will drive up to an amp of current and accept up to 65 volts from the power supply. Naturally these drivers are DC>DC converters and I power my LED lamps from surplus SMPS.
This works well but is fiddly.
The other option is to buy LEDs that are already soldered onto aluminium backed PCBs. LEDrise is the only company I've seen that sell a variety of LEDs ready soldered to suitable PCBs.
LedRise - High Power LEDs
There will no doubt be others too.
Then there's CREEs CXA2011 LED array that comes ready made on an aluminium mounting pad. This packs a lot of light into an easily mountable and relatively small area. It is nowhere near as efficient as say the XML LEDs, but this is only when driven at high currents, at low currents the CXA starts to become a lot more energy efficient. The CXA however is a lowish current high voltage part which isn't suitable for off the shelf LED drivers I'd bet.
I build all of my LED drivers, I've used single chip solutions from maxim, linear tech and zetex and they all work well. Maxim make a pretty neat little chip called the MAX16832 which is one of my favourites. It comes in a handy 8 pin SOIC case so is easier to manage then most of the other LED drivers and it also only needs very few supporting components, it will drive up to an amp of current and accept up to 65 volts from the power supply. Naturally these drivers are DC>DC converters and I power my LED lamps from surplus SMPS.
You can buy LEDs on stars quite cheaply from eBay.
MR11 and MR16 LED lamps are also pretty cheap, about $4 and up (including shipping).
As for drivers, adding one resistor to a typical car cellphone charger will create a constant-current driver. I've opened up a half-dozen of those and they all contained some version of the MC34063 switching regulator. (As did a camcorder charger, and a $2 USB lighter jack) This isn't the most efficient driver, though, because the chip uses a 1.25V reference voltage. Google for the ON Semiconductor app note ONS_MC34063A_AN1.pdf.
A more efficient switching driver chip is the PT4115; you can get drivers with that chip from Dealextreme.
If you want to be able to dim your LEDs, it's best to use PWM control. LED colour can change if you just lower the current, so switching it on and off at a constant current makes it dimmer without altering colour. The ATTiny13 is commonly used in flashlight drivers, and various people have come up with custom firmware for it. If older PICs are your thing, there's the freeware lightbrain dimmer that was originally designed for incandescent bike lights, but should be adaptable to LEDs. And the source code is there.
PWM is also the "correct" way to regulate a Peltier cooler (or TEC: thermo-Electric Cooler); here's one that looked easy to adapt to LED control using a common chip:
W's simple high power PWM TEC controller You should be able to find a similar chip in an old PC power supply, and maybe a power MOSFET to use as well.
Using a resistor as a driver isn't an awful idea. With a correctly sized resistor, reliability should be much higher than any active driver. And with a bunch of LEDs in series efficiency may be as good as with a switching driver.
You should find some advice on fish tank forums; I know I've run across some while googling for LED information.
MR11 and MR16 LED lamps are also pretty cheap, about $4 and up (including shipping).
As for drivers, adding one resistor to a typical car cellphone charger will create a constant-current driver. I've opened up a half-dozen of those and they all contained some version of the MC34063 switching regulator. (As did a camcorder charger, and a $2 USB lighter jack) This isn't the most efficient driver, though, because the chip uses a 1.25V reference voltage. Google for the ON Semiconductor app note ONS_MC34063A_AN1.pdf.
A more efficient switching driver chip is the PT4115; you can get drivers with that chip from Dealextreme.
If you want to be able to dim your LEDs, it's best to use PWM control. LED colour can change if you just lower the current, so switching it on and off at a constant current makes it dimmer without altering colour. The ATTiny13 is commonly used in flashlight drivers, and various people have come up with custom firmware for it. If older PICs are your thing, there's the freeware lightbrain dimmer that was originally designed for incandescent bike lights, but should be adaptable to LEDs. And the source code is there.
PWM is also the "correct" way to regulate a Peltier cooler (or TEC: thermo-Electric Cooler); here's one that looked easy to adapt to LED control using a common chip:
W's simple high power PWM TEC controller You should be able to find a similar chip in an old PC power supply, and maybe a power MOSFET to use as well.
Using a resistor as a driver isn't an awful idea. With a correctly sized resistor, reliability should be much higher than any active driver. And with a bunch of LEDs in series efficiency may be as good as with a switching driver.
You should find some advice on fish tank forums; I know I've run across some while googling for LED information.
On a lark, recent trip to Home Despot (the big box home improvement store) I bought a "40 watt equivalent" 350 lumens 8 watt (draw) "Ecosmart" bright white PAR20 LED bulb.
Pretty darn good. Spot. Good color light 3000K (like a halogen spot that is not super high K color). Quite good brightness! Think I paid $19.95 for it. That was the shelf price, but I have to check the receipt to make certain.
A pair of these above my bench would be more than adequate.
Would be nicer to get this sort of output in MR-16 size, but I can wait. This will be sufficient. No sign on the box or unit where it was made! (ha!) Wait, assembled in Mexico (but no idea where the parts are from). House brand product for Home Despot.
Dimmable too it says.
Can't imagine making my own comparable unit from parts...
Report from the field.
_-_-bear
Pretty darn good. Spot. Good color light 3000K (like a halogen spot that is not super high K color). Quite good brightness! Think I paid $19.95 for it. That was the shelf price, but I have to check the receipt to make certain.
A pair of these above my bench would be more than adequate.
Would be nicer to get this sort of output in MR-16 size, but I can wait. This will be sufficient. No sign on the box or unit where it was made! (ha!) Wait, assembled in Mexico (but no idea where the parts are from). House brand product for Home Despot.
Dimmable too it says.
Can't imagine making my own comparable unit from parts...
Report from the field.
_-_-bear
CPC has them cheaper Andrew, and they have a 9W due in (GU10) but that is 25 quid.
I've used LED lighting for ages where I can, my kids and grandkids have only ever learned to turn lights on... In fact that applies to every electricaly powered device
and I'd rather have 8 GU10 downlights with 4W bulbs than 50W halogen.
I've used LED lighting for ages where I can, my kids and grandkids have only ever learned to turn lights on...
In fact that applies to every electricaly powered deviceand I'd rather have 8 GU10 downlights with 4W bulbs than 50W halogen.
A cheapskates led lighting set up can be built out of christmas tree leds, drill lots of holes in some thing wood or Ali, insert leds and wire up...200 leds approx 4w, with all the leds you got a wide coverage of light, not super bright, but good coverage. After a few months the odd LED started failing...it gave me somthing to do with the old christmas tree lights.
On a more serious note, I have had a high wattage bulb fail (4W), in a lamp holder where a lower powered multi LED one worked for a couple of years (and still does). I wonder whether some of the lighshades designed for standard bulbs trap the heat and dont let the led bulbs cool to full efficiency?
On a more serious note, I have had a high wattage bulb fail (4W), in a lamp holder where a lower powered multi LED one worked for a couple of years (and still does). I wonder whether some of the lighshades designed for standard bulbs trap the heat and dont let the led bulbs cool to full efficiency?
Over here the brightest LED bulb to be sold at department stores and supermarkets is the Philips myVision 9 W/470 lm for about €20.
Ikea sell a 8.1 W/400 lm LED bulb for €10.
The Philips "squashed orange" 12 W/806 lm bulb isn't yet available at these kinds of stores. And in stores where they are available, they're still too expensive (€50 or more) to be economically competitive with CFLs. I was also rather surprised to see that the number of specified on-off cycles is as low as 20,000. In a CFL, I would find that number better than average, but in a LED bulb I would have expected many more...
Ikea sell a 8.1 W/400 lm LED bulb for €10.
The Philips "squashed orange" 12 W/806 lm bulb isn't yet available at these kinds of stores. And in stores where they are available, they're still too expensive (€50 or more) to be economically competitive with CFLs. I was also rather surprised to see that the number of specified on-off cycles is as low as 20,000. In a CFL, I would find that number better than average, but in a LED bulb I would have expected many more...
Most of the chinese import CFLs I have are both dirt cheap ($1USD on closeouts to <$2.50 ea) and usually last 1-2 years max in normal service. I fail to see any energy savings given the cost per bulb, and the increased cost of mfr compared to a tungsten bulb, vs. the minor amount of electric bill savings.
Where they might save is in terms of the load back at the generating plant given the millions of houses that use them...
_-_-bear
Where they might save is in terms of the load back at the generating plant given the millions of houses that use them...
_-_-bear
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.