LED durability or ARC vs Peak

I am reading through the LED archives and it is really amazing what I learnt the last few weeks.
Right now I am choosing between the Peak Matterhorn and the new Arc-AAA. I realise that I will be happy with both of them, but somehow my techno-geeky-me insists in having a rational reason for choosing one or the other. So here's my question for all you LED pros:
Is it correct that the Arc is achiving the higher light output by 'abusing' the LED much more and therefore reducing the lifespan of the LED substantially below the few thousand hours life expectation?

Here's my reasoning: It seems that for a happy long life the recommended load on a LED suppose to be somewhere around 70 mW (20 mA * 3.6V). Increasing this load decreases the life span fast.
From Luxman (Arc AAA Vs. Peak beamshots, Jul 09, 2005) I see that the Peak 1 LED "U" 1xAAA consumes somewhere around 200 mW and the ARC-P 1xAAA approx. 400 mW. The increased power consumption of the Arc corresponds well with the much higher brightness (Flashlightreview.com).
So looking at these number it is very obvious for me that I will be much longer happy with a Peak Matterhorn.

Is this correct? Am I missing something? Please enlighten me.

AFAIK, your theory is quite right. Comparing the new arc AAA with the older one and one can see exactly how the abusive current has such a bad effect to the led. Of course, the new leds is another factor; however, comparing arcs among similar rev/batch, with one is a user and the other is not, one can see the difference in brightness resulting from the high current drive.

That said, these decrease in brightness might not matter much since by the time you notice the decrease in brightness, it is probably time to change your keychain light with nextgen stuff /ubbthreads/images/graemlins/wink.gif. Like now, the new arc AAA can be viewed as that next gen compared to old arc AAA.

As for me, i dont like arc's too much overdriving the led. At that current, arc better off replacing that led with 0.5W high power led (or a MJLed /ubbthreads/images/graemlins/icon3.gif) Peak overdrives them high enuff, but arc is just too much, IMHO.

Just one thing to add. Ensure that you use LED power consumption vice battery power consumption. The ArcAAA drives the LED at around 45mA with a Vf of about 3.6V. That puts power consumption at around 160mW. Battery power consumption will be higher due to converter losses.

I have found that the new Nichia LEDs used in the ArcAAA hold up very well at high currents. On a test I did, the Nichia only dropped about 4.5% in brightness after 147 hours at 57mA current. That's at least 30 battery changes.

Here is an efficiency graph made with Doug Owen's data on one of the early ArcAAAs. Peter may have used newer components with higher efficiencies on the new ArcAAAs.


Paul

Thanks for the test, chimo
That pretty much explains it, you will lose a little over time so if you want the Arc to be the last light on your keychain, and you use it for hours per week... better look somewere else.
Then again, with the semi-regulation of the Arc and Peak, the max drive will not be applied too long with alkaline batteries. If you run lithiums, the LED will suffer a lot more so keep that in mind.
Since we are on the subject, if you use the light for long periods in HOT environments...the LED will degrade faster.
In two or three years, an new wazoo LED will be out and your eye will wander yet again. Maybe a variable output 1-30 lumen Arc AAA will show up so you need a reason to get another one.
I view it as a question of whether you run lithiums, what beam quality do you prefer and how many LEDs do you want.
Isn't having a choice great?! /ubbthreads/images/graemlins/happy14.gif /ubbthreads/images/graemlins/smile.gif /ubbthreads/images/graemlins/cool.gif

Good discussion. If I may, I would like to add some quick points to this thread:


- The latest Arc-AAA is driving at the same average current as the original design made 4 years ago.

- We have never observed any problems over an extended period with driving a Nichia at those current levels. The newer Nichias are even better at handling heat.

Of course, we would replace the light if any were returned with a failed LED.

- Our tests of a few other LED brands have shown in some cases a dramatically reduced lumen maintenance compared to the Nichia brand. Those tests were conducted at 50 hours.

- Any color change observed from previous Arc-AAAs in the new LEDs is a result of that LED's color bin.

- I would like to see an independent 5k hour test with our light compared to other brands!

Thanks.

Peter

Well, what would be a failed led? It wouldn't stop working, just degrade, wouldn't it? How can one measure this, let alone prove it?

What I'd like to see is a simple test.
Take two arc aaa's that show a smiliar reading on a lightmeter.
Put one aside, let the other run through 20 or 30 cells.
Compare readings again.

Repeat

This would be enough to estimate the amount of light lost after one or two years.

Chrisse

Chrisse, see the Nichia CS vs 26k mcd head to head test in the electronics forum for a test sort of like what you're describing.

I've begun to feel that any light this expensive should have a reasonable way to upgrade the LED (having to unsolder the old one is ok).

[ QUOTE ]
chrisse242 said:
Well, what would be a failed led? It wouldn't stop working, just degrade, wouldn't it? How can one measure this, let alone prove it?

What I'd like to see is a simple test.
Take two arc aaa's that show a smiliar reading on a lightmeter.
Put one aside, let the other run through 20 or 30 cells.
Compare readings again.

Repeat

This would be enough to estimate the amount of light lost after one or two years.

Chrisse

[/ QUOTE ]

I have great confidence in the Nichia LEDs.

I volunteer to do the test. You just have to convince Peter to send me two ArcAAAs. /ubbthreads/images/graemlins/smile.gif

Paul

Thanks for the excellent information chimo!

Hi drmaxx,
I did talk to Peak LED after my previous measurements...and they said I might already have the new version LED and, if so, probably is the snow.
Well...I just had to know for sure what SNOW looked like and ordered another one (1xAAA 1xLED)... /ubbthreads/images/graemlins/smile.gif
Just rcvd it yesterday..and....its very nice...but I think I may like my "older" Peak color better.
Intensities are close, IMO the new one may have a slightly smaller spot and therefore appear a little brighter at the center. Both beams are very clean and smooth.

The old Peak fairly closely matches the color of the Ultimate 60, which I consider to be my color "reference".
The NEW Peak light is IMO more white and has VERY slightly more blue tint - mostly just looks whiter and is very nice...and still quite less blue than the ARC-P. I think "Snow" is a good name for this Peak's color...

But, I wondered if I would see much difference in battery current between the old vs new light....
Used the same new battery for all measurements with V(no load) = 1.543V (Fluke 87, meter verified against 89 and 8060A /ubbthreads/images/graemlins/smile.gif).
I cleaned both Peak lights contacts with a used, but clean, 2000 grit paper...just slightly to make them shinny. Did not clean the ARC-P...sponge in the way and it looked shinny anyway.
Interestingly both Peak light currents increased...new one by 32 and the old one by 36 mA.
Also Note: These are crude measurements taken with probes by holding the Bat(+) against the head
contact while the other probe pushes against the head threaded area. Used 10A input on the meter.
[It would be nice to have a fixture for proper current conduction..oh well...]

Results:
Old Peak "U" 1xLED 1xAAA = 174 mA
New Peak "U" 1xLED 1xAAA = 130 mA /ubbthreads/images/graemlins/cool.gif
ARC-P ........ 1xLED 1xAAA = 289 mA
Measurements were repeatable.

Copied from my old beamshot thread:
Peak 1 LED "U" 1xAAA => 144 mA. (AAA => 1.523 V no load)
Peak 1 LED "H" 1xN => 93 mA. (N => 1.549 V no load)
ARC-P 1xAAA => 290 mA! (hard to keep stable contact)(AAA => 1.529 V no load)

Notice the Old Peak current increased from 144 to 174 mA (30mA)!
probably since I have cleaned the contact.
Considering Old and New Peaks are about the same intensity, I wonder if the New Peak's lower current is due to LED version or Lottery..? /ubbthreads/images/graemlins/confused.gif
LM

Thanks for the kind words.

Paul

Very intersting! Thank you guys for all the interesting facts and contributions. I also figured that a typical AAA battery has somewhere around 1200 mAh (Li-Enegizer), so that the average comsumption over 5 h can not exceed 250 mW. So the 57 mA current (200 mW) through the LED seems much more realistic.
My conclusion: This obviously I need to find an other criteria to choose between the two. It might boil down to availability /ubbthreads/images/graemlins/wink.gif

drmaxx,
I was interested to see what the battery voltage would be under load, especially with the ARC.
[measured current and voltage under load]
As I was experimenting taking many measurements, I noted the ARC-P maintained I=289mA while the Peak's current dropped from 130 to 111mA. [Battery no load V after test = 1.497 V.]

Through the narrow voltage range I measured, seems like the ARC-P circuit is acting like a constant current source...while the Peak's current was continuously decreasing with a small change in V more like DD.
[IMO the light intensity also would drop accordingly]

I understand your concern about LED life...but thought you may find this interesting and wonder how important maintaining relatively constant brightness is in your criteria? /ubbthreads/images/graemlins/smile.gif

Really good point. It seems that the Arc has the better circuit. Thanks for pointing that out. I just hope that the Arc's are available soon (the date seems to shift weekly /ubbthreads/images/graemlins/smile.gif)

Luxman,
That would explain why Peaks have such a serious runtime, much less current drawn! Since t.i.n did all those measurements with lithium AAA batteries, I ordered 16 of them off the net and hope they don't get caught up in customs.
By the way, it has been a long time since I heard someone reference a Fluke 8060A True RMS multimeter /ubbthreads/images/graemlins/happy14.gif I still have one from the 80's and it still works fine. Generally, I use my Fluke 179 True RMS meter but it is good to go back and push those side selector switches.
Back in the day...

BentHeadTX,
The 8060A brings back memories /ubbthreads/images/graemlins/smile.gif. I loved that meter when it first came out [talk about a clicky]. /ubbthreads/images/graemlins/grin.gif ..I think I had an 8020A? then and sold it then get the 8060A ...still works great!
89IV--1.587 or 1.5868 V
8060A 1.587 or 1.5872 [not bad for about a 20 yr old meter]
87 ----1.587 /ubbthreads/images/graemlins/grin.gif
/ubbthreads/images/graemlins/happy14.gif

drmaxx,

...don't forget the color difference...the Peak has less of that "LED look" and certain objects might be more comfortable to look at... /ubbthreads/images/graemlins/smile.gif
Depending on your situation you may need a longer runtime etc...

Like everyone says...."Get Both"..!! /ubbthreads/images/graemlins/cool.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif

I am aware of the colour and decided that I do not mind the blue tint. I somehow like the fact, that the Arc switches over to a moonlight modus. In this way it is very clear that the battery is empty and I can replace it with a good conscience. In other lights I tend to postpone the battery change, even if I am sick of the light output /ubbthreads/images/graemlins/jpshakehead.gif. And as a long lasting light I got the Infinity....

Be aware, drmaxx, that the switch-over to moon mode is not always, maybe never, a sudden evident change. My old Arc AAA seems to slide gradually into moon mode. I don't recall ever seeing it drop abruptly into moon mode. This, to me, is not a problem or concern.

Geoff

I voluntarily did not receive the ARC-P paperwork since I had ordered a pre-release version...and did not read the ARC FAQ...until now...
[I kinda thought all "good" lights would have similar regulation]
Although I like the color and smooth beam of the Peak better....But now...since I have done these simple measurements AND read the ARC FAQ, I have a new appreciation of the ARC-P and like it even more! /ubbthreads/images/graemlins/grin.gif
I think I appreciate the current regulation much more and especially to get "more" out of my battery at a higher intensity and know better when to change it instead of the light getting dimmer and dimmer and wondering if I should change the battery and waste whats left....Nice Light.. /ubbthreads/images/graemlins/smile.gif

[ QUOTE ]
Like everyone says...."Get Both"..!!

[/ QUOTE ]
As a newbie there are lot's of niiiiice lights out there. Next on the list is a eternalight. It seems that this is a very nice safety/camping light. So, right now two AAA light are just too much (...convincing my wife that I really need two lights for one keychain /ubbthreads/images/graemlins/tongue.gif).