PhotonFanatic was kind enough to send me a sample of a 'new' version of the K2. The part number is LXK2-PWC4-0160. Using the K2 part number decoder ring tells me that this should be a 160 lumen at 1A part. Notice the lack of an alphabetical bin marker - it looks like Lumileds is moving to the Rebel part numbering system for K2s. It's likely they will also move to the rebel binning structure for Vf and Luminous flux as well.
These parts don't seem to be available anywhere yet, nor are they listed anywhere either. Given the recent release of higher brightness Rebel parts, using the thin film die and new phosphor, it's reasonable to assume that this new version of the K2 is also using the same construction for the die/phosphor.
First up, a visual comparison between the existing K2 (PW12-S00) on the left, and the 'new' K2 on the right:
Head on (apply directly to the forehead!):
You can see that the new K2 has a different phosphor, since you can see the dot pattern of the die through the phosphor now, much like the difference between the 100 lumen rebel and the older rebel. In addition, the silicon submount is a gold color - however, I think I've seen K2s with gold colored silicon submounts before, so I'm not certian that it's an entirely new feature.
And from the side:
The die/phosphor combo is shorter, again like the difference between the new and old rebel. This should make the beam project relatively more forward (less side emitted light) compared to the old K2. So it looks like indeed, the new K2 is using the thin film technology, which eliminates the sapphire substrate part of the die.
Let's check out the beam profile:
The new K2's beam profile is a little different, projecting less out the side as predicted. This might affect the beam slightly when used with reflectors.
Here is the raw performance data for the new K2:
This part is supposed to be 160-180? lumens at 1A. It looks like this part is performing at the high end of the bin range.
Compared to other power LEDs in the same class:
First, Vf:
The Vf of this sample is higher than most of the others. This will cause it to take a hit in the efficiency department.
Next, output:
This sample is performing about equal to the Cree XR-E P4 bin. Hopefully better performing versions of the new K2 will be available when the update to this part is announced. It would seem rather silly that the Rebel would out-perform the K2 - Lumileds can produce very high performing LEDs in the rebel line - there doesn't seem like much reason to keep those improvements from getting into other products immediately.
Efficiency:
Even though this sample about equals the XR-E in output, it lags in efficiency, due to the higher Vf.
Finally, "droopyness"
If the rebel didn't have "anti-droop" technology, then the new version of the K2 also doesn't. The new K2 scales pretty much the same as all the other parts.
While this was just a sample part, it's nice to see Lumileds finally get their buts in gear and get the K2 performing in the range of the Cree and Seoul parts. This sample definitely wasn't a performance leader - hopefully the final release parts perform better. Lumileds is going to have to stay on the ball with this if they are going to keep up with the pace that Cree has set in the last year for performance increases.
Of course, we'll have to see how long it is before Lumileds decides to make these new K2 parts available for everyone.
The time frame, performance, and pricing of these new K2 parts will all effect how willing people are to use them. If we have to wait to long to get the improved parts, people will begin to fear "another K2", where parts were promised but never delivered. The new K2s will have to perform much better than this sample for people to decide to use them vs. the higher performing Cree parts (Q4, Q5, and better).
These parts don't seem to be available anywhere yet, nor are they listed anywhere either. Given the recent release of higher brightness Rebel parts, using the thin film die and new phosphor, it's reasonable to assume that this new version of the K2 is also using the same construction for the die/phosphor.
First up, a visual comparison between the existing K2 (PW12-S00) on the left, and the 'new' K2 on the right:
Head on (apply directly to the forehead!):
You can see that the new K2 has a different phosphor, since you can see the dot pattern of the die through the phosphor now, much like the difference between the 100 lumen rebel and the older rebel. In addition, the silicon submount is a gold color - however, I think I've seen K2s with gold colored silicon submounts before, so I'm not certian that it's an entirely new feature.
And from the side:
The die/phosphor combo is shorter, again like the difference between the new and old rebel. This should make the beam project relatively more forward (less side emitted light) compared to the old K2. So it looks like indeed, the new K2 is using the thin film technology, which eliminates the sapphire substrate part of the die.
Let's check out the beam profile:
The new K2's beam profile is a little different, projecting less out the side as predicted. This might affect the beam slightly when used with reflectors.
Here is the raw performance data for the new K2:
Code:
Current (mA) 0.1 30 130 310 670 980 1260 1570 1930
---------------------------------------------------------------------------------------------
Vf 2.53 2.75 3.05 3.25 3.47 3.60 3.68 3.78 3.89
Watts 0.40 1.01 2.32 3.53 4.64 5.93 7.51
Lumens 33 75 141 187 221 257 292
Lumens/W 84 74 61 53 48 43 39This part is supposed to be 160-180? lumens at 1A. It looks like this part is performing at the high end of the bin range.
Compared to other power LEDs in the same class:
First, Vf:
The Vf of this sample is higher than most of the others. This will cause it to take a hit in the efficiency department.
Next, output:
This sample is performing about equal to the Cree XR-E P4 bin. Hopefully better performing versions of the new K2 will be available when the update to this part is announced. It would seem rather silly that the Rebel would out-perform the K2 - Lumileds can produce very high performing LEDs in the rebel line - there doesn't seem like much reason to keep those improvements from getting into other products immediately.
Efficiency:
Even though this sample about equals the XR-E in output, it lags in efficiency, due to the higher Vf.
Finally, "droopyness"
If the rebel didn't have "anti-droop" technology, then the new version of the K2 also doesn't. The new K2 scales pretty much the same as all the other parts.
While this was just a sample part, it's nice to see Lumileds finally get their buts in gear and get the K2 performing in the range of the Cree and Seoul parts. This sample definitely wasn't a performance leader - hopefully the final release parts perform better. Lumileds is going to have to stay on the ball with this if they are going to keep up with the pace that Cree has set in the last year for performance increases.
Of course, we'll have to see how long it is before Lumileds decides to make these new K2 parts available for everyone.
The time frame, performance, and pricing of these new K2 parts will all effect how willing people are to use them. If we have to wait to long to get the improved parts, people will begin to fear "another K2", where parts were promised but never delivered. The new K2s will have to perform much better than this sample for people to decide to use them vs. the higher performing Cree parts (Q4, Q5, and better).
Last edited:
