Lights

[Deleted member 4366]

It's pointless getting hung up on the Lumens, just use them as a guide once you have one type of LED then you can then use the rating to decide if it will be brighter or not
I hope the the 9 LED monster has a good sized battery pack, that sucker will draw some serious current.

It uses three 5aH Trustfire 26650 batteries. I think that they can give 15 amps at 12.6v. They'd probably make a good ebike battery.

 

[Old_Dave]

Forget the hype, luxes, lumens, watts and candle powers (which was in the old money before we went metric )

The only scale to use is the Fing rating

Examples of usage are ..

Thats Fing bright = very good , or Fing hell my retinas are burnt = excellent, the other end of the scale approximates to.. that's Fing useless = not good

As to power consumption the same Fing scale can be used, as in..

That didn't last Fing long = bad, or that lasted Fing ages = very good.

Value of products can also share the same scale, as in.. thats Fing cheap.

 

[Jeremy]

It's pointless getting hung up on the Lumens, just use them as a guide once you have one type of LED then you can then use the rating to decide if it will be brighter or not

Very true. The BS that gets spouted about light brightness is near-universal, because few of them ever get properly tested by an accredited lab. I doubt many realise that a few hundred lm is a heck of a lot of light, and a few thousand is really massive over kill, just a waste of power really.

The photo below shows some of my LED bits with the smallest torch I've made so far. This has a single Cree LED (by coincidence the same ones that d8veh linked to earlier) and I over-run it a bit at around 900ma. It's a U2 bin cool white and puts out around 280lm or so after the losses in the optics (maybe 320 to 340 lm before the optics). With that 20 deg collimator it is blindingly bright, far brighter than any small incandescent torch of a similar size. I did use the same optics and LED (you can see some spare Cree LEDs in that photo) for a bike light, but the beam is too narrow. I'm currently making a three LED bike light, using a linear array of three Cree LEDs in a block, with a mix of optics to get both good long distance penetration and a wide spread at ground level in front of the bike.

 

[Deleted member 4366]

This has a single Cree LED (by coincidence the same ones that d8veh linked to earlier) and I over-run it a bit at around 900ma.
View attachment 4584

I don't think it's the same one. The T6XML runs up to 3000ma - more than three times as much current!

 

[morphix]

On the subject of lights, the "Flashlight" app on my HTC Desire puts out an astonishing amount of light for a single LED on the phone!! I used it last night and it actually illuminates the pathway or even a road, quite well, better than my bicycle light does!! I like how it has a touch sensitive brightness control too. I bet it drains the battery down very fast, but still very handy.I used it to read the other night on low

Edit: just had a thought..with a Speedict setup with the HTC on a handlebar mount, you have effectively got a very powerful forward facing touch screen bike light, should you need it ;-)

 

[Jeremy]

I don't think it's the same one. The T6XML runs up to 3000ma - more than three times as much current!

Well, I can assure you, as I happen to have one of them in my hand, that it IS the same one. I'm not yet blinded by the thing such that I can't read the part number!

Sure you can over-run XM-Ls, or any of them come to that, as long as you can get the heat out of the puck. They are actually rated at only 700mA though - the output for higher currents is calculated, if you note the data sheet, and you need to take account of the thermal issues when running them at very high current.

In my case, the puck is mechanically clamped to the alloy front block of that little torch (which is only 25mm diameter), with a bit of heatsink compound on the back. I didn't bother with finning, as this is only a 3 and bit watt torch, and as a result it gets a bit hot to touch after it's been on a while. There's no way I could run it at the absolute maximum rating of 3000mA, it would last just a few seconds before the SM LED die melted the solder off the star substrate (I've had this happen on the bench, it doesn't take much to unsolder them).

The bottom line here is that this is a poor LED in terms of efficiency, just 100 lm/W, so it gets very hot when you over-run it to get high light levels. If you can get the heat out, perhaps by moving away from the common 20mm star mount and using a good cooling block, then you can increase the brightness. However, even 9 of these LEDs running at 3000mA each with proper coolers (they'd probably need heat pipes and fans on the heatsink) are still only going to give you (theoretically) less than 10,000 lm into the optics at the very most, maybe 9000 lm output. TBH, I doubt very much that this is that realistic in a torch body, as getting rid of the heat would be a significant issue, especially when you add in the driver loss, which is going to add at least another 10 to 15W of heat inside the torch.

It seems that even a very optimistic assessment indicates that 11,000 lm output is BS. As mentioned above, I doubt it matters though, as even a few thousand lm is a heck of a lot of light in practical terms, especially when focussed to a tight beam.

 

[themutiny]

I have a pair of lumicycle 4si, one flood and one spot.
On boost, they kick out 1120 lumens each. I can't conceive of ever needing anymore, and you can't run them on boost if you're stationary, due to the heat (they automatically switch down). Lumicycle are now selling 3XML lights @ 3000 Lumens each, but I can't see the point, personally?

 

[Deleted member 4366]

I have a pair of lumicycle 4si, one flood and one spot.
On boost, they kick out 1120 lumens each. I can't conceive of ever needing anymore, and you can't run them on boost if you're stationary, due to the heat (they automatically switch down). Lumicycle are now selling 3XML lights @ 3000 Lumens each, but I can't see the point, personally?

I'm glad that at least one other person gets it.

 

[Jeremy]

I'm glad that at least one other person gets it.

There's no problem with "getting it" at all. The spec of 11,000 lm is unachievable, so is BS, plain and simple. The very best output is likely to be around 9,000lm, and that's probably not sustainable for more than a few seconds before the dies overheat. This thermal case is EXACTLY the point I made earlier, in fact, here's chapter and verse on my rationale for limiting the current in the pictured torch for which you questioned my ability to read a part number for earlier:

The LED is mounted on to a good 20mm alloy thermal substrate (the star puck that’s used in the majority of torches). These have a thermal resistance from LED die mount to the back plate of 1.2 deg C/W. Combined with the junction to case thermal resistance of 2.5 deg C/W this gives a total thermal resistance from the junction to the point where the puck connects to the heatsink of 3.7 deg C/W.

The nose part of my small alloy torch, with the puck mount, has a calculated thermal resistance (assuming ideal conditions) of 14 deg C/W. In practice I suspect it may be a bit worse than this, even though the heatsink is integral to the nose machining. The total thermal resistance from the LED junction to the ambient air is at best about 17.7 deg C/W, but realistically I expect it’s closer to 20 deg C/W, judging from the temperature the case reaches.

Maximum junction temperature allowable is 150 deg C, but if the LED is allowed to run this hot the light output drops by about 28%, reliability suffers and the waste heat to be dissipated increases further. If we accept that a 10% drop in light output is acceptable, then the maximum junction temperature allowable is 75 deg C.

100 lm/W luminous efficacy is an efficiency of 14.64% (683lm/W is the physical upper limit). For every W used by the LED, 0.8536 W of waste heat is generated. Vf for the Cree XM-L at 3000mA is typically 3.35V. At 3000mA the LED junction will be creating about 8.58W of waste heat initially (3.35 x 3 x 0.8536) when the LED is cool (it’ll generate more heat as the efficiency drops with increasing temperature).

With a total thermal resistance of 17.7 deg C/W and an ambient temperature of 15 deg C, the LED junction will reach around 167 deg C if run at 3000mA, which is too hot, in fact it's into the area where thermal runaway may be a problem, as there would be 30% or so more heat generated when the junction was that hot.

To keep the LED junction temperature at our selected 10% efficiency loss figure of 75 deg C we need to ensure that the LED waste heat generation doesn’t exceed about 3.4 W (for 15 deg C ambient and a cool LED, less if the ambient temperature is higher or the LED has warmed up).

3.4W dissipation with a LED efficiency of 85.36% equates to a current of around 1.18A as the maximum that this LED can be used at in this application. I chose to play safe and run it at 900mA, to allow for the fact that the torch would be used in ambient temperatures higher than 15 deg C at times and to allow for the increase in waste heat generated as the LED junction temperature increases.

Simples, as Aleksandr Orlov would say....................

 

[Scimitar]

It's pointless getting hung up on the Lumens, just use them as a guide once you have one type of LED then you can then use the rating to decide if it will be brighter or not

That's the whole point of measurement - if the makers are lying, we need to know how much by. Using the numbers as a guide is what it's supposed to be about, but that's totally undermined by unscrupulous marketing weasels, who wouldn't know a lumen if it bit them on the bum.

If I might add - the reason I get hung up about this is simple - lumens are important to me, as a keen photographer and fiddler with things, I like to know how much light there is and for that matter, what characteristics it has. I get thoroughly pis sed off with sales liars, too, so there's another reason for my annoyance.

 

[Jeremy]

That's the whole point of measurement - if the makers are lying, we need to know how much by. Using the numbers as a guide is what it's supposed to be about, but that's totally undermined by unscrupulous marketing weasels, who wouldn't know a lumen if it bit them on the bum.

If I might add - the reason I get hung up about this is simple - lumens are important to me, as a keen photographer and fiddler with things, I like to know how much light there is and for that matter, what characteristics it has. I get thoroughly pis sed off with sales liars, too, so there's another reason for my annoyance.

Exactly.

It's clear that you cannot get the claimed light output from many of these lights, even for a short burst, and the advertisers should be stamped on for misleading advertising. What amazes me is that even when there is strong evidence that advertising claims are patently wrong, some will still buy these things and promote them to others using the same inaccurate information that the advertisers have used.

In this case the engineering evidence is clear. The LEDs used are rated at an absolute maximum light output (and in reality this is a very short burst rating) that is much lower than the claimed total light output, so the claim is plain BS, pure and simple. At best this "11,000 lm" light is going to deliver an output of maybe 9,000 lm for a few seconds before the LED dies heat up and the output starts to drop, then the output will drop further as the temperature rises further. The continuous output is most probably going to be around 3,000 to 5,000 lm, primarily due to the thermal problem.