CREE home lighting project. Pics.

[VanIsleDSM]

I got my LEDs from cutter today and quickly soldered up a little trio to run from a linear current reg using an LM317 for trial purposes.. I needed about 13v so I grabbed the 14.4v drill battery and rigged it all up for comparison...

The kitchen has 300w of MR16 halogen lighting (6 bulbs) and wasting that amount of power is driving me nuts.

I ordered some 3200-3500k CREE LEDs from cutter.. I'm going to be using a switching 48v PSU to power the LED driver.. then each LED driver powers 9 CREE in a series.. the PSU is 84% efficient, and the LED driver is 95%, each LED runs at 700mA, or 2.45w. 3 LEDs per bulb.. 6 bulbs, 18 total LEDs

This means I have 2.45w x 18 = 44.1w of LED light for the whole kitchen instead of 300w.

If you factor in PSU and driver losses....

0.84 x 0.95 = 0.798, take the inverse 1/0.798 = 1.25... 1.25x 44.1w = 55.2w total consumption... barely over 1 original bulb..

I unplugged 2 of the 3 halogens.. to compare 1 to 1.. pics are a little fuzzy because I had to use a slow enough shutter speed to get a life like comparison.. and I didn't want to fish out the tripod..

Here they are... as you'd assume the 3200-3500k LEDs are slightly whiter than halogen.. They look great.. still nice and warm but you can see the higher CRI, the colours appear much more brilliant with the LED lighting... I'm very pleased and can't wait to finish the project.


HALOGEN:

LED:

Let me know what you think!

 

[Darell]

Looks GREAT! Moved to "fixed lighting."

 

[Juctuc]

Looking good!!..You dont have any optic in the "bulbs"?..In my opinion its better in kitchen without optic,just asking..

How you will handle the heat?..Im planing my own project as well. Im going to use three cree as well, but @350mA. I want low watt with maximum lumens. I was also planing to "underdrive" more then 3 leds, so i might get high lm/w..but lets see.. i did not find good driver for "underdriving"

 

[VanIsleDSM]

no optics.. not needed.

The LEDs will be mounted on a big heatsink made in a lathe.. it'll be the size of an mr16 bulb.

It's LED lighting just as much as it is fixed lighting.. really a toss up where to put it.

 

[DaFABRICATA]

You're really tall or your sink/counter top is very low

Looks good! cant wait to see the finished product!

 

[holeymoley]

VanIsle, can you post some pics of the assembly once its finished? I might try this one, great idea.

Also, part numbers for the PSU and driver?

 

[jrminky]

Looks good. I'm curious about the exposure durations of the photos. You mentioned that you had to pick a longer shutter duration for good pictures. Are the shutter times for the MR16 pics the same as those of the CREE pics?

 

[Newuser01]

Since OP hasn't answered the question: here is the info from the las picture

ExposureTime - 1/3.2 seconds
FNumber - 2.80
ExposureProgram - Shutter priority
ISOSpeedRatings - 160
LightSource - Auto
Flash - Flash not fired, compulsory flash mode
FocalLength - 6.00 mm
CustomRendered - Normal process
ExposureMode - Auto
White Balance - Auto
FocalLengthIn35mmFilm - 36 mm
SceneCaptureType - Standard
GainControl - Low gain up
Contrast - Normal
Saturation - Normal
Sharpness - Normal

PS. Just FYI. used irfanview with exif plugin (freeware) great progy!

Enjoy and happy newyear everyone!!

 

[eyeeatingfish]

Have a picture of the assembly?

 

[jrminky]

Thanks for the tip on the EXIF data stuff. Very useful.

So it looks like the camera picked ISO 200 for the MR16 pics while it picked ISO 160 for the Cree pics. Shutter speeds were the same. Thus the Cree fixture is actually a little brighter than the pictures suggest.

 

[VanIsleDSM]

I can tell you for sure the CREE trio was much brighter.. even brighter than the picture shows apparently.

I put the camera on shutter mode, I thought that was good enough, I have since learned that it needs to be on competely manual for perfect comparison..

 

[VanIsleDSM]

I just received the lathe yesterday that I'm going to use to make the aluminum heatsinks.. this should all come together in the next month or so.

 

[VegasF6]

Wow, this is beautiful. I stumbled upon this while trying to come up with some vehicle lighting, boy I am impressed with the results. Good luck with your experiment, I can only imagine a lot of people would, or should have these in there house.

I have switched to CFL in as many places as possible, which is fine for general lighting. But, of course dimming CFL's suck, and are way to expensive and unreliable. And for all my spot lighting I am running halogen. Low voltage when possible, but again, dimming is a problem. My over cabinet lighting is 110 halogen on a standard dimmer and that is where I would really love LED instead.

Any plans for a dimming function on this?

 

[Newuser01]

Any more updates? Especially the finished product.

Regards.
Noob

 

[VanIsleDSM]

Funny you should ask today!

It's been a lot of planning (and money spent on tools)

...But I just finished the prototype heatsink on my lathe tonight.. just undergoing some thermal testing right now... so far I've tried it in the worst case scenario position where the convection cooling would be least efficient

A little hotter than I was hoping for.. but I knew I was going to have to cram quite a bit of heat into a small area to match a 50W MR16.. I'm still well within spec; the heatsink reaches 64C max at 22C ambient right under one of the LEDs... it's running in the "optimal position" right now to see the difference.. but it hasn't quite leveled out yet.

Running 2.45W per LED.. about 80% heat I have about 2W of heat per emitter... @ 8C/W directly mounted onto the heatsink with high pressure.. I'm sure I'm probably under 9C/W to the heatsink.. but using 9C/W @ 2W each I've probably got die temps around 83C.. sound about right?

Cree says their 50,000 lifetime is accurate to 80C die temps..

In the good convection position I seem to be getting about 58C.. but it may creep up a little more yet.

I'm going to add some fins to this design in the other direction as well.. being MR16 is will most likely be on an angle a lot of the time.. so a peg board or "pin fin" layout will allow the air to convect through it at any angle.. I'll report my findings after I've done that.

This is bare aluminum though.. I will be anodizing these.. newbie showed that high emissivity can reduce temps by up to 20C.. and there will also be another thermal path for heat to escape through the fixture once installed.. so I should be in good shape.

 

[VanIsleDSM]

Well I figured I better post a some pics..

..and a fast shutter to see the front.

Haven't put the pins in the end yet for going into the socket... also didn't have a proper countersink for the screws.. and they were a little tall so I had to sand the tops down.. I suppose that's what prototypes are for

Final version will look cleaner.. and hopefully I can tweak it to work better as well.

 

[jtr1962]

I'm going to add some fins to this design in the other direction as well.. being MR16 is will most likely be on an angle a lot of the time.. so a peg board or "pin fin" layout will allow the air to convect through it at any angle.. I'll report my findings after I've done that.

A pin fin design will definitely work very well here. It will allow air flow regardless of orientation. As I'm sure you've already figured out, with fins in only one direction you risk warm air being "trapped" between fins unless there is a way for it to escape. Going by the size of your prototype I think you'll probably be able to get your temperature rise to half what it is now with pin fins. Very interesting project.

Not as nice as yours but I did this a few months ago:

It's basically just a retrofit of an incandescent fixture using a few microprocessor heat sinks for the P4 Crees. The LEDs run at 700 mA. The heat sinks get warm but not hot, perhaps around 40-45°C. It's been running continuously over the bar in the basement since September 2007 with no apparent loss of output.

 

[Newuser01]

Very very nice !! (both of your work.)

I wish that I could do this but budject and skill sets prevent me from ever trying it. But you give me hope!

Keep them coming and hope that your house is lit by all led's and forever!

Regards.
Noob

 

[VanIsleDSM]

When doing the test first I either didn't leave the test going long enough or I didn't have the probe secured well enough.. the temperature still slowly creeps up even after it's been on for over an hour..

So I never really did get a good measure of it before.. and I didn't want to leave it on for so long again to see how high it would go... I really didn't think I'd have this much trouble getting rid of ~6watts of heat..

I cut it vertically so it has the 'pin fin' design and left it for an hour.. 80C, and probably still on a slow rise :shakehead

I've now painted it flat black and am trying again, but I'm doubting it will be very much better. I'll report back in an hour or so.

I don't really know how better to design the heatsink within my capability? ..any ideas are welcome, I may end up lowering the drive current to 600mA if I have to..

 

[jtr1962]

You might want to take a look at this. The fins are running the opposite of the way yours are. I'll also add that the temperatures you're seeing in your tests are probably higher than you'll get in practice. Sitting on a table there's little room for a convection flow since the table is blocking air coming in from the bottom. Once mounted in the fixture this problem won't exist. Also, the fixture itself will conduct away a little heat, and you'll probably have some airflow in the room just due to natural convection patterns or from the HVAC system. Even normal room air flow of a few inches per second can cut the temperature rise in half. I think you'll be fine. And yes, it often takes a few hours for the temperatures on these things to stabilize. You're dealing with a pretty massive hunk of metal and a fairly low wattage heat source.