Luxeons: 1W vs 5W, heat and permanent lumen loss (and failure)

[hotfoot]

Hi all,

Soon we will have the choice of either the 5W or the 1W luxeon, but in the flashaholics neverending quest for brightness, runtime, reliability, cost-effectiveness, price and long service life - many of these requirements will always be at odds with one or more of the others.

Putting purchase price and running costs aside, which would *you* pick if you were building a flashlight and you knew these facts about the luxeons:

- 1W luxeons tolerate higher flashlight case temps than the 5W (105C vs 75C at metal-core PCB)

- 1W luxeons will experience less permanent lumen-loss over time than 5W in the first 42 days of continuous operation at spec (~1% vs ~10%)

- a single underdriven 5W luxeon is more efficient than one or more overdriven 1W luxeons if you are striving for a certain amount of light

- 2 (or more) underdriven 1W luxeons will also be more efficient than a single overdriven 1W luxeon if you are striving for a certain amount of light

- a 1W white LS+optics retails for about $14-16

- a 5W white LS (no optics) retails for about $20-27 (optics are about $1.50 to $3.50 apiece)

Your choices are:

1) Single overdriven 1W LS
2) Dual underdriven/spec-driven 1W LS
3) Single underdriven 5W LS

IMHO, this is a tough decision - but I'm inclined towards (2), and less so to (3). I wish I could provide more info to help with the choices, but that's all I have myself too.

So, what would your choice be?

 

[MR Bulk]

Hmmm, although I'm inclined to agree with some of the above, I can't agree with all of it, mainly because I've not seen it myself (light loss greater with 5W over first 42 days, etc.). Wondering where I might be able to peruse this info.

 

[Doug S]

Hmmm..., are you a durian fancier? Yet another food most folks here have never seen. I've never smelled a ripe one by I'd like to based on their reputation. I would like to try one but they are hard to find in the US.

 

[hotfoot]

MR Bulk,

I got the info from the datasheets. 42 days actually works out to about 1000 hours and in both the 1W (lumen-maintenance) and 5W datasheets, they've plotted out the permanent lumen-loss percentages. Y'know - the more I read those datasheets, the more I realize that these things have some ways to go. Can be a wee bit demoralizing at times, especially since I design a lot around luxeons myself.

Doug S - you're a very adventurous sounding guy! Brave, even! Yes - I LOVE durians, as most locals here do and I'd pay anything from US$20 upwards for a single excellent fruit. I wouldn't think they're easy to find in the US at all, primarily because of their smell. The aromas vary from fruit to fruit and several mainstream varieties are available, all with their own taste and olfactory personalities. I'd gladly Fedex you one to try, but it'd never make it past US customs!

If you are ever in Singapore during durian season, I'll bring you on a durian tour and feast

 

[APW]

hotfoot, I like the bitter sweet one!
How do you say it? Malay way?
Do Ri Ang
Due ri en
Liu Lian

 

[Doug S]

Originally posted by hotfoot:
Doug S - you're a very adventurous sounding guy! Brave, even! Yes - I LOVE durians, as most locals here do and I'd pay anything from US$20 upwards for a single excellent fruit. I wouldn't think they're easy to find in the US at all, primarily because of their smell. The aromas vary from fruit to fruit and several mainstream varieties are available, all with their own taste and olfactory personalities. I'd gladly Fedex you one to try, but it'd never make it past US customs!

If you are ever in Singapore during durian season, I'll bring you on a durian tour and feast

<font size="2" face="Verdana, Arial">Hotfoot: That is certainly a kind offer you have extended. Yes, I fear that were you to ship one, the authorities in biohazard suits would likely surround it, suspecting a terrorist attack

 

[hotfoot]

Originally posted by APW:
hotfoot, I like the bitter sweet one!
How do you say it? Malay way?
Do Ri Ang
Due ri en
Liu Lian

<font size="2" face="Verdana, Arial">APW, "durian" *is* Malay!

It literally means "thorny". Ah... I'm craving for some Tangkak or D28 bittersweet durians now too. Can you find any in the UK?

Doug S - you're most welcome. The tour and feast offer stands anytime you're here.

And I've now just adopted Darell's habit of going off-topic too - I thought we were talking about luxeons here!

 

[APW]

Originally posted by hotfoot:
Can you find any in the UK?

<font size="2" face="Verdana, Arial">Yes, in Chinese Food Shop but they're overpriced!
Andy

 

[hotfoot]

Originally posted by APW:
</font><blockquote><font size="1" face="Verdana, Arial">quote:</font><hr /><font size="2" face="Verdana, Arial">Originally posted by hotfoot:
Can you find any in the UK?

<font size="2" face="Verdana, Arial">Yes, in Chinese Food Shop but they're overpriced!
Andy</font><hr /></blockquote><font size="2" face="Verdana, Arial">Unbelievable!

The world truly is getting smaller...

 

[hotfoot]

EDIT: *bump* - Doug S - hope you see this - for some reason, this post fell off the charts

Also took the opportunity to clarify the post a little and provide more detail.

Doug S,

My most alarming 'lab' result so far...

Test conditions:
=================

- Multi-emitter configuration (x3). No optics.

- less than 25mm spacing between emitters at center

- Direct Arctic Alumina bond onto primary aluminum heatsink with mass and exposed surface area similar to that of 5 mounted luxeons ie. LS/O or LS/C

- Secondary heatsink with mass equivalent of roughly 10 standard luxeon PCBs, however total surface area of all exposed heatsink material would equal perhaps 7 or 8 mounted luxeons. Total thermal resistance yet to be calculated accurately. Heatsinks are bonded to each other with Arctic Alumina at all points of contact.

- Emitters were driven in series to electrical spec by Xitaniums (350ma measured)

- Closed volume steel enclosure, about 2 cubic feet. All heat radiating sources are thermally insulated from direct contact with steel walls/floor with closed-cell foam rubber matting.

- Ambient temp 30C

- Luxeon light-source and transformer are operating within the same closed volume. Thermometer and light probe placed inside and nothing was moved throughout test. Temp monitoring was conducted via a 5 mm gap left in the lid of the enclosure, light readings could be seen via external LCD.

- Pre-test shock: Luxeon unit was overdriven at double rated input current (700ma) using a different-rated Xitanium for 1/2 hour continous under same conditions. This is my most harsh accelerant method.

- Test duration: 24 hours continouous

Results:
========

- Max recorded ambient temp within enclosure (measured at around 12 hours): 48C

- Max recorded temperature of heatsink, probe located 15mm away from emitter slug: 89C (reached about 12 hours into test)

- Typical stable temperature: 86.5C (based on random checks throughout test)

- Relative light output at 5cm from light sensor (taken before pre-test shock): 100%

- Light loss after test: 25-27% (permanent loss) (!!!) Lamp was allowed to cool to ambient temp and restarted. Measurements were then taken after 15 mins of running (warm-up). Even though the lamp did not reach the same during that 15 mins, light output levels did not return to pre-test levels.

CONCLUSION: I have run the same fixture to spec on other occasions, but MINUS the 700ma pre-test overdrive and have not yet experienced permanent light loss, only reversible heat-resulant light loss. I thus concluded that overdriving the luxeons was the primary reason for the permanent loss.

 

[hotfoot]

Elektrolumens,

My fear is the same as Lucien's. A severely overdriven luxeon will probably experience noticeable lumen-decay much sooner than 1000hours, much less 10,000 hours as suggested. Mileage may vary depending from luxeon to luxeon due to their inconsistent production, though.

*EDIT:* In my harsh-condition accelerated test above, even 30-mins of use a day would result in that level of light loss in about 2 months.

There have already been some CPF posts about commercial overdriven 5mm nichia LED-based lights losing their oomph less than a year after their purchase.

This info may affect those who use their lights very regularly and those who use them as lanterns, which typically run on extended duties.

Flashaholics who merely whip out lights to momentarily blind others for a split second may be exempt somewhat

 

[Lucien]

Well, I haven't had a chance to play around with any LSs yet, so my take is all conjecture...

IMHO it really depends on the job at hand. The dual 1W definitely wouldn't fit in something like a 2AA Mini-Mag, leaving just a single 1W or 5W. Given the thermal limitations of using a 5W in that case, it would then depend on how much light output that could be gotten out of one while keeping the heat level managable. If it were possible I want to try one

, but after factoring cost into the equation, I'd only find it worth it if light output was ~3x that of a 1W.

Though I probably wouldn't want to overdrive the 1 W unit anyhow if it meant the light output of the unit would decay faster. Which does bring up a question: is "lumen decay" reduced when underdriving a unit? If not, then its a bit of a waste doing so.

Given a larger light with adequate heatsinking, I'd then be looking at what kind of beam I'm looking for, flood or spot / whether variable focusing can be done. A number of (slightly)underdriven 1W units becomes an attractive option for a nice flood beam.

If I'm only limited to 2x1W, then its a question of which option (2 or 3) is brighter. So far I have no reason to believe that a flashlight can be "too bright"...

 

[ElektroLumens]

I have modified flashlights in all of the combinations you mention. The 1W overdriven, a dual 1W, a triple 1W. The 5W at 6 volts underdriven, and the 5W slightly overdriven at 7.5 volts (5AA).

The dual 1W is a very nice setup. It is slightly brighter than a single overdriven 1W, but each emitter is working easier. I usually overdrive the dual just as the single, but each emitter is only getting 1/2 the current (parallel setup here).

For brightness, even if the 5 watt loses some of it spark, the 1 watt could not compare. Remember, we do not really notice a difference in emitted light unless there is huge difference. If you compare a light with another that is 10% less, you would not be able to discern the difference.

For battery life, you better forget about the 5 Watt, unless you seriously underdrive it. I can get 24 hours of very bright light from a single 1 watt, from 3 D cells. With the 5 watt, using 5 D cells, you might get 12 hours. Of course, I still need to do a lot more testing with the 5 watt.

With 5 AA batteries and the 5 watt, I get just under 2 hours of bright light. Somewhere between 1.5 and 2 hours, it drops off rapidly to a low light. How long that goes, I don't know, as I haven't tested this yet.

Wayne J.
www.elektrolumens.com

 

[shipinretech]

One problem with the 5W parts is that they are like firearms, if you can see the muzzle opening, you are in grave danger. I really love the light that kicks out of the 5W part, but I don't see it as a general purpose part in the same way that the 1W is. Putting it differently, the 1W LS we use in the InReTech adapter produce a light that is clearly better than the incandescant light that comes with the Mini-MAG. That makes it "good enough" for most uses that people have for a general purpose flashlight. The 5W part produces a light that is beyond the need of most "nice to have" lights. As lovely as that intense a light is, it is hard to justify having a lot of them.

 

[hotfoot]

Interesting replies from one and all. But no one has queried the (to me, anyway) worrisome *permanent* lumen loss.

There may be somewhere along the respective 1W and 5W lumen-maintenance curves where a pair or trio of 1W luxeons actually exceeds the output of a 5W. I guess I'm alluding to the lifespan of the flashlight here.

How would the rational person in you (yes - not many of us here have that ;D) see this? Would you buy a premium flashlight that was built on the premise of a lifetime of service, and then have to shelve it after as little as 3000 hours (or about only 4 months) of continuous use because its "lost its shine"?

I'm still an solid-state lighting advocate, but sometimes it seems like a heavy price to be on the cutting edge...

 

[Doug S]

Hotfoot, like you, I think that longterm degradation is a significant issue. I am biased by the perspective of engineering stuff that must perform to spec for many 1000's of hours. I think that you and I are in the minority among the LED fans on this forum. The more common view is to get maximum output, efficiency and life be damned.
I am curious about the basis for the following statement in your original post:

- 1W luxeons will experience less permanent lumen-loss over time than 5W in the first 42 days of continuous operation at spec (~1% vs ~10%)

What conditions are you considering "at spec".
The current datasheets for the 1 and 5W Luxeon give identical luminous degradation graphs with a specified Tj of 70C for the InGaN LEDs [a condition very hard to maintain if operating anywhere near 5W].

 

[ElektroLumens]

Originally posted by shipinretech:
One problem with the 5W parts is that they are like firearms, if you can see the muzzle opening, you are in grave danger. I really love the light that kicks out of the 5W part, but I don't see it as a general purpose part in the same way that the 1W is. Putting it differently, the 1W LS we use in the InReTech adapter produce a light that is clearly better than the incandescant light that comes with the Mini-MAG. That makes it "good enough" for most uses that people have for a general purpose flashlight. The 5W part produces a light that is beyond the need of most "nice to have" lights. As lovely as that intense a light is, it is hard to justify having a lot of them.

<font size="2" face="Verdana, Arial">I have to agree with you here. The 1 watt is adequate light for most purposes. The StarLux flashlight I will be selling in a few weeks, using the 1 watt, slightly overdriven, will fill the bill for most uses. I will also put in the 5 watt emitter, when white is available, for those of us

who 'NEED' more light!

At only $49.99, it is in the affordable range of most people, too.

StarLux, The 3 'D' Cell Luxeon Star Flashlight

Wayne J.
www.elektrolumens.com

 

[hotfoot]

Doug S,

Good spotting!

My statement was based on data picked out of:

- "Luxeon 5-watt Star", DS30, page 7 (see graph at page bottom)

- "Application Brief AB07 Lumen Maintenance of White Luxeon Power Light Sources", page 4 (see graph at page bottom)

If you compared the 2 documents above, you'd see why I said what I said.

But, I can see where the confusion is coming from: DS23 (the Luxeon 1-watt datasheet) *does* indeed show an identical lumen-decay graph for the 1W.

AB07 is an older document and the info represented within was comparing lumen-decay of luxeons vs typical 5mm white LEDs. Given the very early stage of the luxeon's introduction at the time that document was prepared, data may have been extrapolated or guesstimated. Hmmm - that would mean that even the 1-watters now will sag in brightness pretty badly...

 

[hotfoot]

With regards to overdriven Luxeons, I suppose I could live with them if I could easily replace them when I feel that their performance is no longer "like new".

But I'm still more than a little uncomfortable at the thought that these incandescent-replacements will just get dimmer and dimmer. At least with a bulb, everytime you pop one, you get a fresh new and *bright* replacement.

This rant doesn't seem to be going anywhere because I'm just wondering how many folks out there realize that many of their LED barndoor blasters simply won't be so hot anymore beyond a coupla thousand hours. And the sickening thing is that it doesn't matter whether you're driving it to spec or not, except that overdriven ones dim faster (I figure).

 

[Doug S]

Hotfoot, there certainly is a big discrepancy between the AB07 and the datasheets on this point!
BTW, the AB07 page 4 figure does not explicitly identify the power LED tested as a Luxeon. The graphs on page 5 are identified as Luxeons.