Want to create LED Fog lights, sugestions?

[mds82]

So i'm looking at creating some custom LED fog lights for my 2006 BMW 3 series. I was thinking about running 10 or 11 XR-E's all at 1000ma, all stuck to one large heatsink.

For power i was thinking about drivers, and i was thinking about using 3 sharks each running 3-4 LED's

The only issue is waterproofing, because water does get inside th bumper of the car, so is it safe jut to coat the Shark with some silicone once its all set and wired?

For optics, i was going to use a combo of Oval and wide angel optics to give plenty of spill. The LED's would not be mounted completely flat to make sure light goes out into all angles.

 

[LEDninja]

The blue of normal LEDs get refracted/reflected by the fog making them useless as fog lights. Make sure you buy amber LEDs. Or at least warm white LEDs.
XR-E Flux and Warm White Tint Bin Selection:
N2 Flux 8B Tint(2720k to 3050K)
P3 Flux 8A Tint(2900k to 3100K)
P4 Flux 7A Tint(3200k to 3500K)

 

[TorchBoy]

Wide angel. Heh heh.

 

[VanIsleDSM]

I'm planning to do something very similar with LEDs.. but I'm not sure yet how I want to get the yellow light.. I don't think amber LEDs are the answer, and I've tested the 8B tint 8A and the 7A tint CREE LEDs.. None were much what I was looking for.

What does looks best so far is 5000k white CREE Q5s through a thin yellow filter.. and you appear to lose less light this way than the warm white LEDs.. the N2 is quite dim.

Something else I want to try is mixing in some green to go with the blue already in the white LEDs and possibly adding a bit more blue if needed.. maybe even a bit of red for better CRI.

Possibly a little combo of both adding some greed and a very slight yellow filter.. I haven't done much work with coloured LEDs.. but I'm looking forward mixing them up and trying to get the colour I'm going after.

 

[TorchBoy]

maybe even a bit of red for better CRI.

That would make for worse CRI as the large red spike would mean it was further away from a blackbody radiation curve of the same colour temperature. Add them, sure, but CRI isn't the be-all and end-all of accurate colour recognition!

Using white LEDs is (still) more efficient than using a checkboard of green and red, right?

 

[VanIsleDSM]

Since I'm aiming for yellow light in the first place red isn't that far.. and adding a 'bit' of red.. there would be no large red spike.. plus red is also a longer wavelength better in the fog..

Did you see those over 100lm/w warm white houselamps that use CREE white and golden dragon (I think) reds for a warmer colour and better CRI? I can't remember the name of the bulb I'll try to find it... not production yet though.



Using just white is more efficient yes.. but great results can be achieved by supplementing white with a bit of RGB.. and hopefully I can tune in enough yellow that way.. But I might end up using a bit of yellow filtering... we'll see.

 

[TorchBoy]

Another thought - if red LEDs were used in the array they'd have to be completely diffused to be legal, and thus showing no "red" light to the front of the vehicle. VanIsleDSM, your idea of a light yellow filter over neutral white and green LEDs sounds easier.

 

[mds82]

Well i know this sounds a bit silly, but i am going to stick with a 6000k color. I know in true fog it isnt effective at all, but there are maybe 5- in 365 days in my are that are foggy. I'm more interested in them matching up with my regular driving lights, ( 6000k Xenon bulbs).

As for powering them, will a shark handle more then 4 amps of ouput? I would like to get away with only 2 shark's per side, but i know i will be pushing a lot of power through them.

10 LED's x 1000ma each = 50 watts correct? 1 XR-E at 1000ma is about 5 watts - correct?

 

[TorchBoy]

3.7W according to the spec sheet (ie, 3.7V @1A), but I've measured a Vf as low as 3.3V at 1A, which would be 3.3W.

 

[mds82]

OK, thanks
!!!

 

[frenzee]

It appears that the effectiveness of yellow/amber/red or any othe color in fog lights is all but questionable at best:

http://www.newton.dep.anl.gov/askasci/phy99/phy99xx4.htm
http://www.gi.alaska.edu/ScienceForum/ASF5/593.html

Also, adding red or green LEDs to white LEDs may decrease the overall efficiency of your light, since generally they both have lower lm/W numbers (compared to cool whites). I have done some experimental lights using reds, greens and whites (although Lumileds not Crees), but for an entirely different reason, which we won't go into here. If you do decide to go this route regardless, let me just caution you that once you do put the light assembly together, assuming everything is connected in series, despite what theory says, don't be surprised if your red LEDs light up happily while the green ones - sometimes, not always - just sit there in the dark. Is the Vf variation the culprit? I don't know and have not found a reasonbale explanation for it, but have come across it enough times to know that it can and does occur, and does so randomly for all I can tell. Reds in series, no problem. Greens and ambers...same. Mix them up and spooky things start to happen. So FWIW, I would suggest isolating different color banks and driving them using separate drivers. Another FWIW: It takes 4 reds to 1 green, OR 2 reds, 1 green and 1 cool/neutral white to come up with a reasobale approximation of yellow/amber, assuming everything is driven at the same current. Again this is based on experimental lights using Lumileds emiiters, Rebels to be specific.

And I also disagree with the notion of CRI/blackbody radiation having anything to do with automotive lumination, unless you were planning to take photos. (sorry Torchboy, I love your posts otherwise). The primary function of foglights should be discerning objects and judging their distance and relative motion. What color they are could only be secondary or tertiary concern.

No offense intended, of course. I respect everyone's posts here immensely and deeply and their points of view as well. Just voicing a different opinion.

 

[TorchBoy]

Great information there frenzee, great post. Thanks especially for the info on the number of emitters required for yellow light. I guess I'll add http://www.danielsternlighting.com/tech/lights/light_color/light_color.html which also considers the way our eyes work, especially with blue light. (It had to be mentioned at some point. )

assuming everything is connected in series, despite what theory says, don't be surprised if your red LEDs light up happily while the green ones - sometimes, not always - just sit there in the dark.

Were there equal numbers of each colour in each series? For example two red, one green, one white? Or was each string just one colour? If the latter it would be very hard to match the total Vf of the strings.

I'm really not a great fan of CRI. I think it's horribly misleading, even if you are trying to take a photo. The thinking behind my post was a) it's not worth going for, and b) it's possibly the wrong way to get it anyway. Personally I'd rather go with your point of using another higher efficiency white emitter instead.

 

[VanIsleDSM]

I've read before the yellow light doesn't help for whatever scientific reason.. but the fact is in the real world it does help.. I can notice a difference easily... here's a couple pics courtesy of a guy from hidplanet..

Good info about the number of emitters required.. I didn't think 3 red/1 green would make yellow? :thinking: I thought you needed mostly blue/green.

 

[TorchBoy]

Well, it looks like it takes out a lot of the glare. If you have a CRT screen handy have a close look at which phosphors are lit up by the 2nd pic. (Red and green, making yellow.)

 

[VanIsleDSM]

If it does not glare as much (reflect back) then does it not pierce the fog better?

In either case.. it's more useful for fog.... and all bad weather conditions.


...How do I check the phosphor levels individually?

 

[mds82]

I am aware that yellow is much better for fog lights, because it reduces the amount of glare you get, however i'm not really concerned about that.

My main question is driving these with the car electrical system.

How many XR-E's can the shark handle when the LED's are driven at 1000ma each, asuming the input voltage is 12v( or more likely 14.4v)

 

[-Virgil-]

It appears that the effectiveness of yellow/amber/red or any othe color in fog lights is all but questionable at best:

http://www.newton.dep.anl.gov/askasci/phy99/phy99xx4.htm
http://www.gi.alaska.edu/ScienceForum/ASF5/593.html

H'mm. I see problems with both of these. The first one does a good job explaining why the Rayleigh Scattering "explanation" does not apply to fog lamps, but both it and the second one are based around a bit of a straw man: They both set out to prove that yellow light does not penetrate fog better than white light. That's easy to prove, because it's true. The common knowledge that yellow fog lamps "penetrate fog better" is an accurate description of a perception, but not an accurate explanation for why it is so, which is related more to how human eyes interact with light, not so much how light interacts with fog. One of these also brings up the issue of light losses via filtration, but in an overly simplistic fashion. Certainly any kind of filter in the system will block some of the light — that's what filters do, it's how they work. But the total light loss depend on the SPD of the light source and the transmissivity curve of the filter (i.e., what's being produced vs. what you're filtering out). Filtering an incandescent source to obtain yellow gives very small overall light loss indeed (much smaller losses than filtering for blue or "extra white" light, for example). I'd post more info on blue vs. yellow filtration of incandescents to demonstrate the point, but that would pull the thread even further off topic. Furthermore, objecting to yellow fog lamps on grounds of filtration losses makes the assumption that all light of any/every wavelength is equally useful for the task at hand, and that's not so. For whatever reason (the way the light interacts with the atmosphere, the way our eyes interact with the light, the way the super-secret space aliens have programmed our universe to work, whatever), SPD can be optimized for whatever task is at hand, and there's not single "best" light color for any given job.

Here's a paper (pdf) out of the Lighting Research Center at RPI in New York, which suggests there is a benefit to "yellow light" (in quotes because it's more complicated than that) for driving in inclement weather. Torchboy's link to the Stern page about yellow vs. white is also worth a read.

So...how best to achieve yellow light with LEDs? I'm honestly not sure. It is frequently stated that Selective Yellow lies outside the RGB colorspace, but I have seen variable-color LEDs emitting light that certainly appears to be an unmistakable Selective Yellow.

 

[-Virgil-]

Well i know this sounds a bit silly, but i am going to stick with a 6000k color. I know in true fog it isnt effective at all, but there are maybe 5- in 365 days in my are that are foggy. I'm more interested in them matching up with my regular driving lights, ( 6000k Xenon bulbs).

This sounds as if you're not interested in making fog lamps, but in making toy "lookit me!" lites to go with your toy headlites (6000K = less light and illegal light color).

Do you plan to do anything at all to control glare from these "fog lamps" you're going to make...?

 

[mds82]

a stock 4350k Xenon bulb produces 3200 Lumen where as a 6000k Xenon bulb produces 3000 Lumen, so there is not much of a light decrease at all. Also, 6000k is still well within the legal limit of coloring. 8000k or 10000k is where you start to have problems with both decreased light and illegal color temperatures.

I am looking to build this more or less for fun.

I was planning on putting optics on each led and aiming the everything downwards so that very little light would be aimed upwards into oncoming drivers eyes. I plan to use 10 led on stars with optics, all pointing in a slightly different direction so as to make sure there is a fairly even light output as well as keeping all the light aimed low to the ground

 

[-Virgil-]

a stock 4350k Xenon bulb produces 3200 Lumen where as a 6000k Xenon bulb produces 3000 Lumen

Incorrect. The Philips 6000K Ultinon or Osram 5400K D-HC produces 2,800 lumens. That's a loss of 400 lumens, which is significant.

6000k is still well within the legal limit of coloring.

Incorrect. The 5400K Osram D-HC and 6000K Philips Ultinon are both outside the SAE/ECE white boundary (too blue). Don't know where you were getting your info — maybe from whoever sold you your 6000K bulbs?

I was planning on putting optics on each led and aiming the everything downwards so that very little light would be aimed upwards into oncoming drivers eyes.

Well, that'll be a start...you'll need to watch for primary, secondary, and higher-order reflections causing upward stray light.