Preview of updated K2: testing of "new" K2 sample

[chris_m]

I just think there needs a little perspective to the "love-in" which seems to happen on here with every Luxeon pre-release announcement. I sometimes wonder whether if Luxeon released a part which was miles behind the current state of the art efficiency but still a little better than what they had released before there would still be people on here swooning over it because it was a drop in replacement for lights they bought back when Luxeon USED to be the performance leader.

I'll give you that certainly the Rebel is a good performing part, but there is as yet no evidence that we will see K2s released in the near future which even do as well as that. Meanwhile some of my irritation is that your graphs seem to show that this K2 you have performs as well as a Cree and the Rebel better, and it does NEED to be pointed out that the Cree part in your graphs is a really old bin, so it's not really a very current comparison. Is all very well you commenting on the bins listed in Luxeon's bin structure - actually seeing such parts in the real world is another matter (noting that Cree has R4 bins in it's bin structure, which is 130lm at 350, so ~290 at 1A).

jeff - I've looked again at that graph, and it still doesn't show the K2 on test performing any better than the Cree P4.

 

[Kiessling]

The difference between a P4 and a Q4 Cree isn't all that exciting IMHO. However, being able to use all the stuff (reflectors, optics, heads, bezels, drivers, ...) we were using back then with the Luxeons again with the K2 is worth a lot. There is experience, the wheel does not need to be re-invented as it had to be withthe Cree and Rebel.

bernie

 

[evan9162]

I just think there needs a little perspective to the "love-in" which seems to happen on here with every Luxeon pre-release announcement.

There's the same amount of "love-in" over Cree announcements and new XR-E bins that are released. CPF is an enthusiast site. People get excited over new releases, increased competition, and more choices. Sheesh. Why do you even visit this place if you don't understand that?

I sometimes wonder whether if Luxeon released a part which was miles behind the current state of the art efficiency but still a little better than what they had released before there would still be people on here swooning over it because it was a drop in replacement for lights they bought back when Luxeon USED to be the performance leader.

Luxeon is a part name. Lumileds is a LED manufacturer.

Anyways, I very seriously doubt it. The V-bin K2s have been available in quantity for about 6 months now, and no one gives a crap because how far they are behind the curve in terms of efficiency. No one is swooning in this thread. The only thing that people are getting interested in is the fact that Lumileds just might get back into competition in terms of efficiency, where Cree and Seoul have been leading for a long time. I thought more choices were a good thing?

I'll give you that certainly the Rebel is a good performing part, but there is as yet no evidence that we will see K2s released in the near future which even do as well as that.

We have no idea when these are coming out. But since people have been getting samples, and the parts have even shown up temporarily on Future's parts listings, so it sounds like it could be soon. Lumileds is notorious for dragging their feet on new releases, a fact that universally annoys everyone.

Meanwhile some of my irritation is that your graphs seem to show that this K2 you have performs as well as a Cree and the Rebel better, and it does NEED to be pointed out that the Cree part in your graphs is a really old bin, so it's not really a very current comparison.

Well ex-****ing-cuse me for not owning all the latest and greatest LEDs on the planets that meet your standards for comparison. I didn't have anything but a P4 XR-E when I did this test. How am I supposed to include data for a part I don't have? Why not truck on over to my XR-E Q4 review to see it compared there?

It's people like you and responses like this that discourage people from doing these kinds of investigations and sharing information, when the only thing that happens is that they get sniped at because someone doesn't like some insignificant detail about how things were done. It's definitely goes against the spirit of CPF when you behave like this.

This was an opporotunistic test of an unreleased part. It's normally something people look forward to. People don't look forward to someone threadcrapping all over everything just because they feel like making snarky comments that lower the level of discourse.

Is all very well you commenting on the bins listed in Luxeon's bin structure - actually seeing such parts in the real world is another matter (noting that Cree has R4 bins in it's bin structure, which is 130lm at 350, so ~290 at 1A).

ARC Mania had -200 parts on hand a while ago. The current Rebel-100s hit the mid 230s at 1A. There's no reason that -200 parts shouldn't be in abundant supply whenever they get around to releasing these. Even -240 parts should be reasonable, again, given the current performance of Rebel-100 parts.

 

[Anglepoise]

Keep up the good work Evan. Much appreciated .

 

[Calina]

Keep up the good work Evan. Much appreciated .

+1

 

[jtr1962]

There's no reason that -200 parts shouldn't be in abundant supply whenever they get around to releasing these. Even -240 parts should be reasonable, again, given the current performance of Rebel-100 parts.

One thing which concerns me isn't that Lumileds can't make a competitive LED. They obviously can with the recent Rebel 100s. Their biggest problem is production. Rebel 100s aren't currently available. They only were for perhaps a month, and even then I couldn't get an answer on buying a specific bin as with the Crees. There may indeed be a -240 K2 soon, but if it's only available for a short time, and not again for weeks or longer, then what good is it? At least when Cree releases a new bin, it'll be out there in large quantity, even if admittedly a little expensive/hard to get in the beginning. I really hope that the K2 -200s and -240s are available in large quantity soon, and at competitive prices. The more players in this ballgame, the better it'll be for everyone.

And thanks for the great review!

 

[Doh!Nut]

I am sure chris m would love to send evan a Q5 to update his tests

 

[Grox]

Keep up the good work Evan. Much appreciated .

What Anglepoise said :twothumbs

I enjoy all your reviews.

Thank you for the effort.

 

[evan9162]

One thing which concerns me isn't that Lumileds can't make a competitive LED. They obviously can with the recent Rebel 100s. Their biggest problem is production. Rebel 100s aren't currently available. They only were for perhaps a month, and even then I couldn't get an answer on buying a specific bin as with the Crees. There may indeed be a -240 K2 soon, but if it's only available for a short time, and not again for weeks or longer, then what good is it? At least when Cree releases a new bin, it'll be out there in large quantity, even if admittedly a little expensive/hard to get in the beginning. I really hope that the K2 -200s and -240s are available in large quantity soon, and at competitive prices. The more players in this ballgame, the better it'll be for everyone.

And thanks for the great review!

Yes, I've noticed this too, and it's rather puzzling. Huge availability first, then none anywhere. Their production ability / supply chain is very wierd. Perhaps they're diverting their thin film die production to the new K2 ramp. Or perhaps they're just having problems, and can't produce fast enough. Who knows? From what we can see, Lumileds seems to lack production consistency (not in terms of product quality, but in terms of output volumes). Definitely not the kind of behavior you expect from a big player in the LED market.

A lack of availability of parts definitely hurts them, and makes potential customers switch to another available option, even if it means re-working their design for a new package or beam profile.

 

[coolwaters]

dude a arms race with LEDs...

anyways didnt know that rebels can go up to 1.8 amps...

where and how much did u get them for? i want to try.

 

[evan9162]

dude a arms race with LEDs...

anyways didnt know that rebels can go up to 1.8 amps...

where and how much did u get them for? i want to try.

They really can't. Lumileds only rates the rebels up to 1A of current. 1.8A is severly pushing things.

The Rebel-100s don't seem to be available anymore, which is pretty stupid. At the time I got them (mid-July), they were $6.60 each.

 

[Kiessling]

Not that it seems to be necessary any more right now ... but I still will say it ... Gentlemen ... please be civil and respectful.
Thank you
bernie

 

[evan9162]

Well, I got actual shipping parts today. I ordered 3 -160 and 3 -180 parts.

First, the okay part - these things perform pretty much like they should. No surprises here:

	Current(mA)
		100	350	700	1000	1250	1500	1750	2000
----------------------------------------------------------------------------------------
K2 4-180 #1	26	79	138	179	210	235	256	273
K2 4-180 #2	27	81	141	183	213	237	259	276
K2 4-180 #3	26	79	138	179	210	236	258	276
K2 4-160 #1	24	73	128	168	197	222	244	262
K2 4-160 #2	27	79	137	180	210	234	254	271
K2 4-160 #3	26	79	136	174	201	223	241	256

Yep, definitely won't win any brightness contests with these.

And now the really ugly part:

	Current (mA)								
		20	100	350	700	1000	1250	1500	1750	2000
-------------------------------------------------------------------------------------------------
K2 4-180 #1	2.66	3.26	3.74	4.12	4.37	4.56	4.70	4.82	4.93
K2 4-180 #2	2.89	3.23	3.69	4.06	4.28	4.44	4.57	4.70	4.80
K2 4-180 #3	2.89	3.23	3.70	4.08	4.31	4.48	4.62	4.75	4.85
K2 4-160 #1	2.96	3.28	3.64	3.92	4.11	4.23	4.35	4.46	4.55
K2 4-160 #2	2.95	3.27	3.61	3.85	4.01	4.13	4.22	4.30	4.38
K2 4-160 #3	2.86	3.18	3.56	3.85	4.03	4.16	4.28	4.40	4.50

No, those are not typos. The Vf on these is terrible. The sample I tested had a pretty high Vf already - but these are just through the roof. At 1.5A, there's at least 0.5-0.7V "extra" creeping in somewhere, translating into 0.7-1W of extra heat to be dissipated. That 5.5C/W sure comes in handy to get rid of all that extra heat.

These were so far off, I used two different power supplies and two different DMMs just to verify the results.

It looks like these K2s are starting to live up to the legacy of their earlier ancestors - ho-hum performance and terribly high Vf.

The relatively low brightness of these is one thing, but having such a high Vf will undoubtely be a deal breaker for using these is any battery powered application.

 

[mofiki]

This is bad news considering the Fedex guy just delivered 6 of these K2 0180 today. Oh well, another $40 spent towards the cause. I've kind of learned not to jump on band wagons over the years or the old saying if everybody is jumping off the bridge, you know? So when people here are pushing Cree or Seoul I didn't pay much attention. Maybe I should have listened. I wanted to use these in a battery application. I wonder if there would be a better use for them like illuminating the inside of my clothes dryer.

 

[evan9162]

If possible, would you be able to map out the Vf curve like I have done? It would be good to know if all of them are behaving the same way.

 

[mofiki]

My lack of the proper equipment would make my results unreliable. I am a noob to this hobby of light building but I must say I'm getting hooked.

I only have a BK Precission 1651 power supply which doesn't seem to give more than 600mA regulated current. I've been looking for a good current meter but haven't picked up anything yet. I don't think I would invest in to much more than that except maybe a light meter if they're not to much. Even though I'm enjoying building some lights and messing with leds I can't see that kind of equipment being that useful. I want to get more into machining parts.

I wish I could help. If there is anything you think I can do I'd be glad to help.

 

[Curious_character]

. . .

It looks like these K2s are starting to live up to the legacy of their earlier ancestors - ho-hum performance and terribly high Vf.

The relatively low brightness of these is one thing, but having such a high Vf will undoubtely be a deal breaker for using these is any battery powered application.

Having a high Vf is not, in itself, a bad thing. In fact, it could make flashlights more efficient. Here's why: The Vf of most present LEDs falls in the middle to top of the range of voltage that a single Li-ion cell has during its discharge. This makes it relatively difficult for a regulator, which has to be able to reduce the voltage (buck) when the cell is fresh, and change to a boost mode somewhere in the middle of the discharge if it's to regulate during the whole discharge period. The resulting buck-boost regulator ends up being more complex and generally less efficient than a straight boost or buck design. If an LED has a voltage that's reliably higher than the cell voltage, a plain boost regulator can be used. This is the case now with lights running from one or two Alkaline or NiMH AA or AAA cells. So a higher Vf could bring us more efficient, less expensive lights which run from a single Li-ion cell such as an RCR-123, 17670, or 18650 and be well regulated during the whole discharge period. (A lot of lights now regulate well only during the first or last part of the discharge because the regulator can only buck or boost but not both. For example, an MRV bucks but doesn't boost, so it regulates only during the very first part of an 18650 discharge. A P1 boosts but doesn't buck, so it regulates only at the very end of the discharge period when using an RCR-123. The Tiablo A8 is unusual in that it has a buck-boost regulator which regulates over the whole discharge period of an 18650 cell. But it's not very efficient.)

What we have to do is look at the LED efficiency in terms of lumens per watt, not lumens for some given current which isn't really a measure of efficiency. If two LEDs supply the same number of lumens per watt, one with Vf a bit greater than 4.2 volts would have a definite advantage for single Li-ion cell lights. But of course, the LED output is often shown for lumens out for a given current. If two LEDs produce the same number of lumens at the same current, the one with a lower Vf has higher efficiency than one with higher Vf because it's consuming less power. But if they produce the same number of lumens out with the same power in, the one with the higher Vf is equally efficient and might be easier to use.

My point is while it's certainly fair to consider an LED to be inferior if its efficiency (lumens per watt) isn't as good as another, having a higher Vf, by itself, isn't necessarily a disadvantage and can actually be an advantage. These LEDs do have poorer efficiency than other modern LEDs. But I'd like to see ones with the same efficiency as other LEDs and with the higher Vf that these devices have.

c_c

 

[PhantomPhoton]

Well there goes my hopes for retrofitting some lights with Lumileds for the next 6 months.
Guess I can keep anticipating SSC releasing a V bin or something.
Once again thank you for the info Evan9162.

 

[mofiki]

Thanks for that info Curious. Your point of view makes sense and is something I didn't even consider all the while I was complaining about finding a power source, especially when trying to power four leds.

 

[Curious_character]

Sorry, I inadvertently left out the essential word not in the first sentence which I've now corrected by editing. The remainder of the posting was intended to support the statement with the not included. I apologize for the sloppy proofreading.

c_c