AR Sponsor
Posted: 8/6/2014 4:05:24 PM EDT
New video blog explains and demonstrates why the highest-output flashlight may be the least reliable: An expanded explanation and more information is available here: http://www.ELZETTA.com/blog/lumenwars/ |
|
Yes, your emitter and optics can generate 1100 lumens otf. What does that prove about reliability? You say higher output lights generally aren't as reliable and mention limits of components and margins of safety, but don't actually explain them. All the video shows is the emitter being over-driven a little bit.
Driving components to their absolute limits does shorten their lifespan undoubtedly. Especially if heat management is not properly addressed. Is the emitter in the Elzetta light given a direct thermal path to the heatsinking of the head? Or is there an insulative [electrical or thermal] layer of plastic between the metal the die solders to and the mounting pad inside the light? That is the primary downfall to over-driving emitters, is the thermal management. Currently there are mounting pads that provide a direct thermal path to the light body, and those can enable the standard Cree XM-L2 to generate 1600 lumens without seriously compromising diode lifespan. 1100 lumens from a XM-L2, (which is what the AVS uses, correct?) is a perfectly healthy output if the heat is managed correctly. Granted its a bit much for a pocket light to be able to deal with, but you don't make the point in the video. A more...honest explanation, and one that is completely valid, would have been to open-up the Elzetta head and put it next to a commercially-available Chinese light that was actually driven hard enough to generate 1500 lumens or so from a comperable emitter, then take IR temperature readings of the aluminum surrounding the emitter at various points of time, and show how much-more heat-soaked the over-driven die is. Or you could discuss safety more by comparing amperage draws that both lights would generate and discuss the limits of certain cells, thermal runaway, ect. Or, given that a large part of your market are going to put these lights on rifles, you could mention that too much light destroys one's natural night vision. another perfectly valid counterpoint to higher-output flashlights. You could do more than just say that harder-driven lights are less reliable and not provide any points to back it up. |
|
Overdriving an LED does not mean it's going to fail. There are a LOT of factors to consider, heat sinking is the major factor.
A Vendor over on Candlepowerforums, vinhnguyen54 HEAVILY modifies LED lights and his are damn reliable! But, he does do it correctly. MASSIVE copper or aluminum heats sinks. Tailored control circuitry, ect. I have a Fenix TK41 he modded that is kicking out 1,500 Lumens+. My Fenix TK75 he modded and de-domed is outputting 4,400 Lumens. I can't even begin to describe how bright and how far the throw is. |
|
Quoted:
Overdriving an LED does not mean it's going to fail. There are a LOT of factors to consider, heat sinking is the major factor. A Vendor over on Candlepowerforums, vinhnguyen54 HEAVILY modifies LED lights and his are damn reliable! But, he does do it correctly. MASSIVE copper or aluminum heats sinks. Tailored control circuitry, ect. I have a Fenix TK41 he modded that is kicking out 1,500 Lumens+. My Fenix TK75 he modded and de-domed is outputting 4,400 Lumens. I can't even begin to describe how bright and how far the throw is. Let's be completely fair, that's turbo mode, and 1800+ lumens from a XM-L2 is pretty harsh even with the best heatsinking. You're getting right towards the top of its output curve, not to mention horrible efficiency. |
|
Quoted: Currently there are mounting pads that provide a direct thermal path to the light body, and those can enable the standard Cree XM-L2 to generate 1600 lumens without seriously compromising diode lifespan. I wonder what exactly the lifespan is on an XM-L2 being run at 19+ watts, factor in the shock from recoil and I can't imagine those bond wires holding up for very long. Yeah you will definitely need some heat sinking given that, what, ~85% of that wattage is just heat and not to mention any additional heat cooked up by the driver -another limiting factor in lights this size (who wants to stack 14 AMC7135's?) Plus you have the issue of actually powering the whole thing, even the latest and hottest unprotected 18650's can't yet maintain that kind of discharge current for more than a few minutes. Maybe Elzetta kept in vague so they wouldn't overwhelm people with the vast number of rather technical reasons for not pushing their ANSI FL1 rated lights to compete with the claims of the Chinese competitors who, often quite generously, estimate their light's theoretical output at the emitter instead. I'm just waiting for some new optics for the XP-L to come out- forget trying to push a single XM-L2 at 20+ watts when I can get the same 1,600 lumens from 9 watts and a triple XP-L. 
|
|
Quoted:
I wonder what exactly the lifespan is on an XM-L2 being run at 19+ watts, factor in the shock from recoil and I can't imagine those bond wires holding up for very long. Yeah you will definitely need some heat sinking given that, what, ~85% of that wattage is just heat and not to mention any additional heat cooked up by the driver -another limiting factor in lights this size (who wants to stack 14 AMC7135's?) Plus you have the issue of actually powering the whole thing, even the latest and hottest unprotected 18650's can't yet maintain that kind of discharge current for more than a few minutes. Maybe Elzetta kept in vague so they wouldn't overwhelm people with the vast number of rather technical reasons for not pushing their ANSI FL1 rated lights to compete with the claims of the Chinese competitors who, often quite generously, estimate their light's theoretical output at the emitter instead. I'm just waiting for some new optics for the XP-L to come out- forget trying to push a single XM-L2 at 20+ watts when I can get the same 1,600 lumens from 9 watts and a triple XP-L. Quoted:
Quoted: Currently there are mounting pads that provide a direct thermal path to the light body, and those can enable the standard Cree XM-L2 to generate 1600 lumens without seriously compromising diode lifespan. I wonder what exactly the lifespan is on an XM-L2 being run at 19+ watts, factor in the shock from recoil and I can't imagine those bond wires holding up for very long. Yeah you will definitely need some heat sinking given that, what, ~85% of that wattage is just heat and not to mention any additional heat cooked up by the driver -another limiting factor in lights this size (who wants to stack 14 AMC7135's?) Plus you have the issue of actually powering the whole thing, even the latest and hottest unprotected 18650's can't yet maintain that kind of discharge current for more than a few minutes. Maybe Elzetta kept in vague so they wouldn't overwhelm people with the vast number of rather technical reasons for not pushing their ANSI FL1 rated lights to compete with the claims of the Chinese competitors who, often quite generously, estimate their light's theoretical output at the emitter instead. I'm just waiting for some new optics for the XP-L to come out- forget trying to push a single XM-L2 at 20+ watts when I can get the same 1,600 lumens from 9 watts and a triple XP-L. As I stated in my post, a 19 watt LED torch isn't ideal for weapons mounting anyway, its blinding overkill. Plus, the Elzetta lights don't have the heatsinking for that. My issue was the lack of explanation in the video. No reasons, not even a list of bullet points, just "High output lights are bad, mkay?" And a triple XP-L will be awesome, just a shame it'll be no good for throwers. And my 1600 lumen torch (more like 1400 lumens otf) is a thrower toy. Look into IMR batteries. Less capacity than ICR cells, but they'll do 20A-30A discharges like a walk in the park. |
|
Quoted:
As I stated in my post, a 19 watt LED torch isn't ideal for weapons mounting anyway, its blinding overkill. Plus, the Elzetta lights don't have the heatsinking for that. My issue was the lack of explanation in the video. No reasons, not even a list of bullet points, just "High output lights are bad, mkay?" And a triple XP-L will be awesome, just a shame it'll be no good for throwers. And my 1600 lumen torch (more like 1400 lumens otf) is a thrower toy. Look into IMR batteries. Less capacity than ICR cells, but they'll do 20-30 discharges like a walk in the park. Agreed, a bit more detail on their reasoning probably could have helped. I actually was looking at IMR cells when estimating run time, I figured the XM-L2 would need at the very least 4.8 amps at 3.7v to hit 1,500lm at the emitter, with a generous 96% total system efficiency and (an unlikely) zero voltage loss through a linear regulator, even the very latest Keeppower IMR18650's wont maintain that voltage for more than 10 minutes. A triple XP-L will definitely be all flood if the optics for it follow the current trends, but pair it with a new 70N02 FET based driver and an IMR like the one mentioned above and hot damn if you won't be knocking on the door of 3,000 lumens on startup. Kind of amazing to think about how far we have come, it was only 6 years ago that Surefire released the original 6P LED that managed a whopping 80 lumens -now just about everyone scoffs at anything under 500lm stock for that size and modders aren't happy with anything under 1k. |
|
Thanks for watching/reading the blog. There is only so much detail that can be packed into a 5-minute video or 500-word blog. The main point Mr. Barnett was trying to make was that, for a given system, pushing the limits of various components (more components than the LED are involved) compromises reliability. The Elzetta AVS Head could easily be made to produce 1100 lumens or more, especially as defined by the FL1 Standard (see last month's video blog below). Increasing output in this manner would not add a penny of manufacturing cost and would undoubtedly play well for sales and marketing. By limiting the output to 900 lumens, however, we produce a safer and more reliable product for the client and that is what drives our company. |
|
How much margin is built in? How do you decide how much to guardband? What is the maximum lumen level just before failure? Do you measure thermals with a FLIR or thermocouple?
Quoted:
Thanks for watching/reading the blog. There is only so much detail that can be packed into a 5-minute video or 500-word blog. The main point Mr. Barnett was trying to make was that, for a given system, pushing the limits of various components (more components than the LED are involved) compromises reliability. The Elzetta AVS Head could easily be made to produce 1100 lumens or more, especially as defined by the FL1 Standard (see last month's video blog below). Increasing output in this manner would not add a penny of manufacturing cost and would undoubtedly play well for sales and marketing. By limiting the output to 900 lumens, however, we produce a safer and more reliable product for the client and that is what drives our company.  http://www.youtube.com/watch?v=pQfSuVxF87Q&list=UUcjecWKnVv_91HFkVfS7xRQ
|
|
Quoted:
How much margin is built in? How do you decide how much to guardband? What is the maximum lumen level just before failure? Do you measure thermals with a FLIR or thermocouple? I'm guessing that the emitter in the Elzetta light employs a standard aluminium mounting pad, which are constructed in such a manner that there is no direct thermal path for the emitter's heat to conduct into the head. In such a circumstance the emitter will heat soak and reach the apogee of its output curve at 5000mA amps, giving off a bit more than 1200 lumens. I'll be optimistic and guess the Elzetta optics have only a 10% efficiency loss. That means that for the light to generate 900 lumens out the front, the emitter itself needs to produce 1000 lumens, consuming roughly 2900mA, which is a reasonable demand to put on a stack of 3 CR123's. This means that the Elzetta light is driven 42% less than the theoretical maximum for its emitter. Not a whole lot, but its still only 3 amps, which is just fine for a standard-mounted XM-L2. Now, when you give the emitter a direct thermal path through an aftermarket copper mounting pad [Sinkpad, Noctigon], the emitter requires a much, much higher level of current flow in order to heat-soak and reach the top of its output curve. At the same 5000mA where an aluminium-mounted XM-L2 peaks at 1250 lumens, a copper-mounted XM-L2 is pumping out a hair shy of 1600 lumens, and at that given point the slope of its curve is the same as the aluminium-mounted emitter at 2900mA. My guess is that a copper-mounted XM-L2 could theoretically digest a full ten amperes and generate shy of 2000 lumens before it heat soaks. But that amount of current might fry the bond wires in the emitter itself. http://budgetlightforum.com/node/19331 These tests were conducted by Match of the budgetlightforum, using an emitter mounted to a chunk of copper barstock as a heatsink, measurements taken after 30 seconds using an integrating sphere over a bare emitter. |
|
Quoted:
I'm guessing that the emitter in the Elzetta light employs a standard aluminium mounting pad, which are constructed in such a manner that there is no direct thermal path for the emitter's heat to conduct into the head. In such a circumstance the emitter will heat soak and reach the apogee of its output curve at 5000mA amps, giving off a bit more than 1200 lumens. I'll be optimistic and guess the Elzetta optics have only a 10% efficiency loss. That means that for the light to generate 900 lumens out the front, the emitter itself needs to produce 1000 lumens, consuming roughly 2900mA, which is a reasonable demand to put on a stack of 3 CR123's. This means that the Elzetta light is driven 42% less than the theoretical maximum for its emitter. Not a whole lot, but its still only 3 amps, which is just fine for a standard-mounted XM-L2. Now, when you give the emitter a direct thermal path through an aftermarket copper mounting pad [Sinkpad, Noctigon], the emitter requires a much, much higher level of current flow in order to heat-soak and reach the top of its output curve. At the same 5000mA where an aluminium-mounted XM-L2 peaks at 1250 lumens, a copper-mounted XM-L2 is pumping out a hair shy of 1600 lumens, and at that given point the slope of its curve is the same as the aluminium-mounted emitter at 2900mA. My guess is that a copper-mounted XM-L2 could theoretically digest a full ten amperes and generate shy of 2000 lumens before it heat soaks. But that amount of current might fry the bond wires in the emitter itself. http://budgetlightforum.com/node/19331 These tests were conducted by Match of the budgetlightforum, using an emitter mounted to a chunk of copper barstock as a heatsink, measurements taken after 30 seconds using an integrating sphere over a bare emitter. Quoted:
Quoted:
How much margin is built in? How do you decide how much to guardband? What is the maximum lumen level just before failure? Do you measure thermals with a FLIR or thermocouple? I'm guessing that the emitter in the Elzetta light employs a standard aluminium mounting pad, which are constructed in such a manner that there is no direct thermal path for the emitter's heat to conduct into the head. In such a circumstance the emitter will heat soak and reach the apogee of its output curve at 5000mA amps, giving off a bit more than 1200 lumens. I'll be optimistic and guess the Elzetta optics have only a 10% efficiency loss. That means that for the light to generate 900 lumens out the front, the emitter itself needs to produce 1000 lumens, consuming roughly 2900mA, which is a reasonable demand to put on a stack of 3 CR123's. This means that the Elzetta light is driven 42% less than the theoretical maximum for its emitter. Not a whole lot, but its still only 3 amps, which is just fine for a standard-mounted XM-L2. Now, when you give the emitter a direct thermal path through an aftermarket copper mounting pad [Sinkpad, Noctigon], the emitter requires a much, much higher level of current flow in order to heat-soak and reach the top of its output curve. At the same 5000mA where an aluminium-mounted XM-L2 peaks at 1250 lumens, a copper-mounted XM-L2 is pumping out a hair shy of 1600 lumens, and at that given point the slope of its curve is the same as the aluminium-mounted emitter at 2900mA. My guess is that a copper-mounted XM-L2 could theoretically digest a full ten amperes and generate shy of 2000 lumens before it heat soaks. But that amount of current might fry the bond wires in the emitter itself. http://budgetlightforum.com/node/19331 These tests were conducted by Match of the budgetlightforum, using an emitter mounted to a chunk of copper barstock as a heatsink, measurements taken after 30 seconds using an integrating sphere over a bare emitter. Thank you for the informative post. It sounds like what Elzetta is trying to say is that they are not exceeding the maximum operating current specification that the LED mfg lists in their datasheet, which is 3.0 A. 900 lumens is just under 3 A in the link you provided. I bet you are not too far off on your 10 A estimate. The LED mfg is recommending 3 A so as not to exceed their maximum junction temperature and therefore allow for reasonable longevity. If you just throw it on a bench and try to see how much current will kill it on the spot, it'll probably be about 10 A if you have good heat sinking. Usually 3x the current is a good estimate. If you get 10 W gives you a junction temperature rise of 125 degrees, then 30 W will probably give you a rise of 125*3 degrees and you're getting pretty close to turning semiconductor into smoke. Bottom line: Sounds like Elzetta is doing due diligence as they say in their video. You can't just throw out all kinds of technical terms like watts, amps, thermal resistance, etc. to 99.9% of your customer base. |
|
Yes, batteries are an important limiting factor. Panasonic (the only U.S. manufacturer of CR123A batteries) specifies 1.5 amps as the maximum current draw from CR123A cells (3.5 amp pulse allowance). Unfortunately, many/most flashlight manufacturers ignore this specification and over-draw batteries to achieve high lumen outputs. Elzetta Modular Flashlights never draw more than 1.5 amps from the batteries. This month's blog covers this topic with this supplemental video: |
| The alternative of course is the employment of a lithium-ion rechargeable battery. Alot of the high-output lights on the market are advertised as working with either two CR123's or a single 18650 li-ion battery. Fenix, Olight, ect, all of their 2xCR123 lights with high-outputs also accept a 18650 battery. FourSevens high-output cr123 lights only put out that high-rating for the 60-second 'burst mode'. You only need a three-amp draw to get 1000lumens from a XM-L2 emitter (more to make is 1000lm otf), and three amps is something well within all good-quality 18650's (Panasonic, Sanyo, Samung, LG) performance envelope. |
AR Sponsor