Hacking the LR6/CR6 Lighting Modules

Joined
Jul 15, 2007
Messages
590
Blastermann, I think you've met your match for your lamps. The LR6/CR6 lighting modules are pretty darn cheap (they have to be- the entire lamp is 50$ at HD) and they give excellent color.

The DR-1000 units I have (and I believe all units) have a feedback sensor in them to monitor both color and luminance. The former is a supposition, but given conversations I've had I'm pretty certain it's true.

Now the question is: Can you override the sensor to achieve higher performance?

The sloppy (dare I say that) production of the heat sinks leaves ALOT to be desired. If I had the capability and time I'd take it in and get the face on both units milled flat... and still use the thermal gasket. The outside of the unit is neutral aluminum- I'd have it painted flat flat black.

And after that, looking through the face of the unit you can see what appears to be XPG LEDs. There are about 8 traces going in and out, so if there are rows... that might explain it.

When I get some time (hah!) I'll do a tear down of the units I have for photos.

I'm tempted to remove the optical diffuser and replace it with some optical glass I've got. More transmittance, less diffusing. I'd even go with modifying it so it wasn't so much a 'deep recess' fixture- I didn't want Deep recess, I wanted 1000 lumens. No product offering.

Just some ideas as I wait for the next 400$ of lighting to arrive...
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
Now the question is: Can you override the sensor to achieve higher performance?
I'll go one better-would it be possible to tweak the sensor to change the color temperature? I unfortunately don't have any recessed lighting to try the LR6/CR6 in. However, if I did, that's the first thing I would want to do, with the goal of bumping it from 2700K up to 4500K-5000K.

I'll hazard a guess that by directing a yellowish source, perhaps an amber LED, into the sensor you would fool the controller into thinking the light was a lower CCT than it actually was. In turn, the controller would adjust the CCT upwards. Only question is will the white balance still be OK?

If I could find a spare $50 I'm tempted to get one just to play with.
 
Joined
Jul 15, 2007
Messages
590
They do make 3500K and 2700K units, so what I would have to find out is what the ratio of yellow to red dies are. The 3500k unit is a nice pattern, so I'm guessing there is only a couple of physical differences.

When talking with cree the reason they don't go more 'natural' daylight is because they start losing out in efficiency. TrueWhite is what I think they call it.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
The thing that has me facinated and yet baffled is why Cree *is* really using the color mix of LEDs in these lights in the first place. It simply adds to the price of the units in some form.

Yes, it increases CRI, but that's not a mandatory requirement.

It allows for precise color contol, but what is the variable here in the first place? Does it allow Cree to use a wider assortment of base color LEDs so that they can be 'dialed in' with the red/amber?

Or, is the circuitry correcting something as the fixture ages? If this is the case, why are we getting color shifts in the first place? Is the thermal load too high for the design, or is this a pre-emptive strike for the aging of the LEDs? I've seen Cree warm-whites shift in intensity and color a bit over their first few thousand hours with even low die temps, but these were older LEDs and thought the problem was fixed.

What I'd really like to see is Bridgelux's Helion compared to the Cree fixture in a 1 on 1.
 
Joined
Jul 15, 2007
Messages
590
The thing that has me facinated and yet baffled is why Cree *is* really using the color mix of LEDs in these lights in the first place. It simply adds to the price of the units in some form.

Yes, it increases CRI, but that's not a mandatory requirement.

It allows for precise color contol, but what is the variable here in the first place? Does it allow Cree to use a wider assortment of base color LEDs so that they can be 'dialed in' with the red/amber?

Or, is the circuitry correcting something as the fixture ages? If this is the case, why are we getting color shifts in the first place? Is the thermal load too high for the design, or is this a pre-emptive strike for the aging of the LEDs? I've seen Cree warm-whites shift in intensity and color a bit over their first few thousand hours with even low die temps, but these were older LEDs and thought the problem was fixed.

What I'd really like to see is Bridgelux's Helion compared to the Cree fixture in a 1 on 1.

I don't know if it allows for precise color control, but I can figure out a few experiments to test it.

I think they've overdesigned the device so much so that they will have 0 failures in that realm for some time. The thermal load is very low- although I haven't measured it- 12 watts and that size of heat sink? Wowzers.

They have the mixture of color and yellow LEDs to allow them to hit their lumens/watt. That makes sense- white isn't that efficient yet at the color temperatures they want. I personally would like to see some 5000K LEDs with some of the spectra filled in with amber and such, but that is more cost. Right now it's just yellow and red leds and a sensor to drive them.

I wish I knew someone that had the 2700K unit and we could both get some internal photos.

I have some CC filters- may run the unit with that over the sensor and see what happens ;)
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
12 watts and that size of heat sink? Wowzers.

Yeah, probably because they're thinking worse case scenario; customers installing them with a foot of insulation sitting on top of the fixture, etc. In an ideal installation scenario and decent breathing room double the lumens should be possible.

So, you say this unit is using both amber and red LEDs along with white?

The lack of a higher CCT unit intrigues me. It's actually more efficient in terms of light production, but it might not work in this design.
 
Joined
Jul 15, 2007
Messages
590
I was a little confused- did you mean the fact they don't offer HIGHEr than 90 CRI confusing?

I did a teardown of the one unit today before I installed it. I still say they've got some pretty sloppy production going on. There were gaps I could see through the heat sink and into the light.

There are 36 LEDs in this light. That means each is run at least 1/3 of a watt... but that also means it could go up to 36 Watts.... with proper heat sinking.

I'll get some photos up later. I'd do it now but I'd rather finish my wine, eat some bread, and have some brie.
 
Joined
Jul 15, 2007
Messages
590
CREE_DR100020100827-110836-IMG_6943.jpg


CREE_DR100020100827-111406-IMG_6952.jpg


CREE_DR100020100827-111411-IMG_6954_1.jpg


CREE_DR100020100827-111504-IMG_6962_2.jpg


CREE_DR100020100827-111534-IMG_6963_3.jpg


CREE_DR100020100827-111619-IMG_6966.jpg


CREE_DR100020100827-111628-IMG_6967.jpg


CREE_DR100020100827-111941-IMG_6972.jpg


CREE_DR100020100827-112153-IMG_6975.jpg
 
Joined
Jul 15, 2007
Messages
590
There are 4 header pins out the top in the middle of the unit. I have a feeling they're used for programming something.... or perhaps monitoring.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
did you mean the fact they don't offer HIGHEr than 90 CRI confusing?

No, higher CCT / color temp. Obviously the higher the color temp the greater the efficiency, so I'm still trying to figure out why Cree is so 'bugga-boo' about a cooler version of the fixture.

Questions: Are those red LEDs mixed with the neutrals, or amber? (or both)

Also, it looks likes theres' a little sensor in the far left middle of the array. That might be our photo detector.

The sink also looks like it's cast metal. It is aluminum though, right?

In any respect, thank's for putting all this work into the tear down! It's a facinating product.
 
Joined
Jul 15, 2007
Messages
590
No, higher CCT / color temp. Obviously the higher the color temp the greater the efficiency, so I'm still trying to figure out why Cree is so 'bugga-boo' about a cooler version of the fixture.

Questions: Are those red LEDs mixed with the neutrals, or amber? (or both)

Also, it looks likes theres' a little sensor in the far left middle of the array. That might be our photo detector.

The sink also looks like it's cast metal. It is aluminum though, right?

In any respect, thank's for putting all this work into the tear down! It's a facinating product.

I was told the reason they don't go cooler is because they take a hit on efficiency- whether that is lumens/watt or what, I don't know. Given what I see though I find it hard to believe they couldn't gin up a third path.

The sink is cast aluminum- and quite beefy. Like I said I"m sorely disappointed by the finishing- they both should have been though a mill or had a flattening operation performed on them.

The electronics look like they are separated from the light engine by a plate- I didn't tear it down far enough to check- but I'd guess it's copper in the middle... and that is what conducts heat to the sides. I'm guessing there.

There is a couple of 293 linear regulators and a few other chips I didn't recognize. I expected to see that triac chip that came out about a year ago that has dimmer capability for CC devices, but it could have been on the backside of a board.

I believe the LEDs to be Red and Yellow, as per the documentation. They were labeled R and Y.

I could see some significant power being done with this guy with a few small engineering changes- including the use of heat pipe tape or a few spreaders and some case pipes into the aluminum shell.

Either way, add a fan and this guy could run easy at 30 watts.... muhahahaha.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
Those are XP-G's, correct? At least the white ones.

Which means they are being severely under-driven, which is fine given this means they will be running at peak efficiency, but still.....

It would still seem that having fewer, higher driven LEDs would be significantly less costly than 34 emitters driven at very low current. It's all going through a diffuser anyways.

Or, this says something about the actual production cost of LEDs :whistle:
 
Joined
Jul 15, 2007
Messages
590
Those are XP-G's, correct? At least the white ones.

Which means they are being severely under-driven, which is fine given this means they will be running at peak efficiency, but still.....

It would still seem that having fewer, higher driven LEDs would be significantly less costly than 34 emitters driven at very low current. It's all going through a diffuser anyways.

Or, this says something about the actual production cost of LEDs :whistle:

Agreed. At 100ma (and I don't have their spec sheet) I expect them to be extremely efficient.

I think they're XPGs. They look like xpgs... although I have never seen a red xpg to know what it looks like, but it seems to be that way.

I'm hoping the guy that bought the CR6 would crack it open for photos.

But yeah- 34 LEDs with the average selling price of 6$ a pop says that is the significant cost here. I think this would make an upfiring luminaire like you've done in the past blow everything out of the market- and I know you're against fans, but you could smoke 30+ watts out of this beast.

The diffuser is odd. I'm reminded of a holographic grating while looking at it. I'd love to put it on the microvision and see what it looks like and what it does to point light sources. You'd have to see it up close to understand, but I get the impression it is either very finely concentric rings or triangles - some sort of pattern to it.
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
Just a thought here but those might be XP-Es, not XP-Gs. Same form factor but different die. Also, they might be XP-Es which fell short of the lowest bin rating, hence making them unsellable through normal channels. However, by running them at ~100 mA, the efficiency would be good enough. In short, Cree is basically turning a sow's ear into a silk purse here. It's not uncommon for manufacturers to find some way to use their factory rejects. By reject I mean that these LEDs might just miss the minimum brightness bin, but otherwise there's nothing wrong with them. Or perhaps they might be a less desireable tint bin. Cree can compensate for that with the colored LEDs.

And yes, this just proves how little these LEDs cost to produce. I'll bet good money even XP-Gs at this point have a production cost well under $1.

It looks like they're using 16 whites and 18 colored. Did you count how many of the colored LEDs are red and amber?
 

LEDninja

Flashlight Enthusiast
Joined
Jun 15, 2005
Messages
4,896
Location
Hamilton Canada
Just a thought here but those might be XP-Es, not XP-Gs. Same form factor but different die.
...by running them at ~100 mA, the efficiency would be good enough.
XP-Cs?
The EZ600 dies are rated:
DC forward current 400 mA.
Peak orward current 600 mA.
http://www.cree.com/products/pdf/CPR3EE.pdf

For XP-Es
The EZ900 dies are rated:
DC forward current 1000 mA.
Peak orward current 1200 mA.
http://www.cree.com/products/pdf/CPR3DX.pdf
This would be overkill.

Sorry but I can't find the data for the LEDs themselves. Terrible at searching for things.
I figured what dies go into which LED by comparing the traces on top of the dies with this picture.
xpg_xpe_xpc_diecompare.jpg


EDIT
purduephotog, can you get a good look at the LEDs and compare them with the picture above.
 
Last edited:
Joined
Jul 15, 2007
Messages
590
XP-Cs?
The EZ600 dies are rated:
DC forward current 400 mA.
Peak orward current 600 mA.
http://www.cree.com/products/pdf/CPR3EE.pdf

For XP-Es
The EZ900 dies are rated:
DC forward current 1000 mA.
Peak orward current 1200 mA.
http://www.cree.com/products/pdf/CPR3DX.pdf
This would be overkill.

Sorry but I can't find the data for the LEDs themselves. Terrible at searching for things.
I figured what dies go into which LED by comparing the traces on top of the dies with this picture.
xpg_xpe_xpc_diecompare.jpg


EDIT
purduephotog, can you get a good look at the LEDs and compare them with the picture above.

They definitely looked like XPG- I was surprised (I know I've seen pics before) but the whole surface looked like it was the emitter. The Reds must be XPEs then- they had a slightly smaller die. Definitely not the XPC.

If I wasn't so busy I'd pull one down and put a multimeter across them... but I am, so I won't.

I'll see if I can get a dimming video at some point.
 
Top