Rebel vs. K2.....FIGHT!

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
I received some K2 emitters, and Rebels today. I decided to do a direct comparison since the bins of each that I got are very close, so they should have very similar performance.

I received PW12-S00 K2 emitters, bin SYAH, manufacture date 07/03 (week 3, 2007) (SYAH = 51-67 lumens @ 350mA, 3.03-3.27 Vf)

The rebels I received are PWC1-0050, bin HWOD, manufacture date 07/09 (week 9, 2007) (HW0D = 50-60 lumens @ 350mA, 3.03-3.27 Vf)

Here are the contenders:

k2rebel1.jpg


The rebel is extremely tiny, and difficult to work with.

Here's a closer image:

k2rebel2.jpg


The entire top surface of the rebel is covered in phosphor. If you shine a blue LED at a rebel, the entire thing glows yellow. With a K2 (and all other Luxeons), only the die glows yellow, as that's the only place where phosphor is applied.

The heat sink and electrical contacts are both on the bottom of the rebel. This almost requires reflow soldering to achive thermal and electrical connections.

I decided to take a different approach. I made a rebel "star", by cutting out an 0.8" circle of 1/16" aluminium, then cutting a notch in the circle for the electrical contacts. I then used arctic alumina thermal epoxy to attach the rebel to the heat sink:

k2rebel3.jpg



Here's the backside, showing the electrical connections:

k2rebel4.jpg



I did this mostly for testing purposes, but maybe it'll give people ideas about alternate ways to mount the rebel to a heat sink without requiring reflow soldering.

Both the rebel and K2 were mounted to a 2"x2" heat sink with a fan blowing on it during all testing.

I mapped out Vf and brightness across a wide range of currents, and used my lumen integration rig to estimate luminous output. I measure the brightness at regular intervals, from on-axis, to 90 degrees to create a beam profile, then use some geometry and the definition of brightness in lux to convert bare-LED measurements to an approximate lumen output.


Here are the results:

First the Vf curves:

k2rebel5.png


The rebel maintains a lower Vf at high currents. I wouldn't be surprised to the the K2's Vf drop after running for some time.

And brightness:

k2rebel6.png


The K2 maintains brightness at high currents a bit better. This could be because of a slightly better thermal path - for this testing, the K2 was mounted directly to the heat sink, whereas the rebel was mounted to a "star" with thermal epoxy, which was then mounted to the heat sink. So the extra thermal interfaces could be accounting for the performance difference at high current.

The K2 and rebel give very similar levels of performance. Given that both were manufactured about the same time, it wouldn't surprise me if they were using the same dies, just in different packages.

The K2 has the advantage that it's in a package that's easy to handle, one that we're used to using.

The rebel definitely is made for more specialized applications, where space is a premium, and cost is a factor. They are more suited for a manufacturer with part placement and reflow soldering capability.

Future Electronics has the rebels for $2 each when you buy 2 or more, whereas the -12 S-bin K2s are $4 each.

If you can deal with the package and mounting, the rebel is a pretty good deal. But, for a small amount more, the S-bin K2s perform equally as well, and are much easier to use.
 
Last edited:

Led_Blind

Enlightened
Joined
Nov 22, 2004
Messages
633
Location
Sydney, Australia
Thanks, thats very useful information.

I have been thinking about getting a bunch of the rebels to use as home lighting. How are they for colour and colour temp?
 

Calina

Enlightened
Joined
Jul 26, 2006
Messages
955
Location
Longueuil, Québec
Wouldn't it be possible to solder the rebel from the top if you scrape the phosphor?

On the other hand, I'm not sure anything useful would be achieved by doing this.
I guess it would depend on the projected use. If top soldering is indeed possible maybe you could get better heat sinking since it would provide a larger contact surface with the heat sink, provided it is properly insulated.
 
Last edited:

MikeSalt

Flashlight Enthusiast
Joined
Jan 10, 2007
Messages
1,948
Location
Stoke On Trent, Staffordshire, UK
Great, insightful reviews. But hang on a minute, wasn't the Rebel supposed to be Lumiled's 'Cree/Seoul killer'? If it is not better than the K2, it stands no chance against the Cree. Whatever happened to the '500 lumens at 2A' we were promised. Have I missed something here?

Thanks
 

yellow

Flashlight Enthusiast
Joined
Oct 31, 2002
Messages
4,629
Location
Baden.at
hehehe, right
but that test is awsome, I like it.

One question I never understand is the "reflow soldering" the Rebel (and the Cree f.e) are planned for:
what happens to the slug then?
I always thought that it is not possible to solder aluminium, or what material is the slug?

If the slug were possible to solder, most of (my) problems would disappear: make some notches into the large heatsink and place contacts (wires, copper sheet, ...) there, isolated from the sink, put three solder blobs (contacts and slug), heat everything till the solder melts and press in place.
Nevertried that as for the "slug soldering problem"

possible?
 

Robban

Enlightened
Joined
Jul 6, 2004
Messages
849
Location
Sweden
MikeSalt said:
Great, insightful reviews. But hang on a minute, wasn't the Rebel supposed to be Lumiled's 'Cree/Seoul killer'? If it is not better than the K2, it stands no chance against the Cree. Whatever happened to the '500 lumens at 2A' we were promised. Have I missed something here?

Thanks
IIRC the Rebel is not that LED. The 500lum at 2A LED will be released later and I think it's refered to as "GenX". I could be wrong though, it's been a while since I read the huge thread that was made when the Rebel surfaced.

More reading material regarding the Rebel:
http://candlepowerforums.com/vb/showthread.php?t=150005
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
nein166 said:
Its called vaporware.


No, it's not. The parts exist, and at least one CPF member has a reel of them - he got VXXH binned parts (113-147 lumens @ 350mA).

I suspect the wait is more of a business decision than a technical one. Lumileds supplies bazillions of LEDs to all sorts of sectors, not just flashlight makers.
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
yellow said:
hehehe, right
but that test is awsome, I like it.

One question I never understand is the "reflow soldering" the Rebel (and the Cree f.e) are planned for:
what happens to the slug then?
I always thought that it is not possible to solder aluminium, or what material is the slug?

If the slug were possible to solder, most of (my) problems would disappear: make some notches into the large heatsink and place contacts (wires, copper sheet, ...) there, isolated from the sink, put three solder blobs (contacts and slug), heat everything till the solder melts and press in place.
Nevertried that as for the "slug soldering problem"

possible?

The slug is not aluminium. In fact, it's been years since any luxeon had an aluminium slug. They've all been silver coated copper. The K2 was the first luxeon that could be reflow soldered, and that was due to changes in the rest of the package to make that possible.

The rebel is built on a ceramic package. The heat sink and electrical contacts are copper pads on the bottom of the device. The heat sink pad really isn't necessary except as a solder mounting point. It doesn't enhance the thermal transfer characteristics at all - all the heat sill travels through the thickness of the ceramic package.

Here's the underside of the rebel:

k2rebel7.jpg
 

nein166

Flashlight Enthusiast
Joined
Feb 16, 2006
Messages
1,575
Location
New York
OOO your right they are popping up ARC Mania got some
http://candlepowerforums.com/vb/showpost.php?p=1926105&postcount=353


evan9162 said:
No, it's not. The parts exist, and at least one CPF member has a reel of them - he got VXXH binned parts (113-147 lumens @ 350mA).

I suspect the wait is more of a business decision than a technical one. Lumileds supplies bazillions of LEDs to all sorts of sectors, not just flashlight makers.
 

mdocod

Flashaholic
Joined
Nov 9, 2005
Messages
7,544
Location
COLORado spRINGs
hmmmm... Your tests are very revealing and sad. I'm saddened that lumileds would ship a new LED, AGAIN with no guaranteed improvements in efficiency. They should have been hiding those bins that perform no better than the K2/LuxIII in an underground warehouse somewhere, to be melted down and made into different LEDs later on. It's a disappointment if you ask me.
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
I'm sure the rebel is more about the package and price than it is any improvement in performance. Of course, you can't stop companies from making big performance claims as part of their marketing.
 

liveforphysics

Enlightened
Joined
Dec 4, 2006
Messages
343
No need to mess with the reflow or slotted bases. 2 passed with a razor over the top permits them to solder normally.


sidebyside2.sized.jpg


BTW- I don't know where you got your rebels, but I bought a few strips of both 95lumen @ 700mA (correct old mistake where I listed 350mA) and 105lumen @ 700mA (corrected old mistake where I listed 350mA) rebels. I would like to see them compaired with the CREE die equiped LEDs.

Thank you for the neat output and VI graphs.

Best Wishes,
-Luke
 
Last edited:
Top