Let's design a road front light beam
- Thread starter Bandgap
- Start date
Help Support Candle Power Flashlight Forum
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
I joined the forum (reader only before), cleared out my temp folder, and that helped the loading.
The side cross section diagram showing mirror/hood placement and the spread of light from the MC-E's is very helpful.
Nice useable low beam road light in keeping with this thread that could be scaled to different cases and LEDs.
The side cross section diagram showing mirror/hood placement and the spread of light from the MC-E's is very helpful.
Nice useable low beam road light in keeping with this thread that could be scaled to different cases and LEDs.
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
Pepko: Exactly that.
It is as if you and LukeA collaborated:
https://www.candlepowerforums.com/threads/214561postcount=6
You added a few tweaks like a mirror and a down angle. Idea could be used in front of existing single LED lights and flashlights too, with appropriate mods. With a translucent hood on the traffic side, side visibility could be increased while cutting off that light.
Other than modded lights that had low beam optics, this appears to be the first idea in the thread, to go prototype, and now DIY light.
Photographers lights have "barn doors" that do much the same thing, so the idea isn't new, but the adpatation is. We have more lumens now, and can afford to throw some away to get the beam we want.
It is as if you and LukeA collaborated:
https://www.candlepowerforums.com/threads/214561postcount=6
You added a few tweaks like a mirror and a down angle. Idea could be used in front of existing single LED lights and flashlights too, with appropriate mods. With a translucent hood on the traffic side, side visibility could be increased while cutting off that light.
Other than modded lights that had low beam optics, this appears to be the first idea in the thread, to go prototype, and now DIY light.
Photographers lights have "barn doors" that do much the same thing, so the idea isn't new, but the adpatation is. We have more lumens now, and can afford to throw some away to get the beam we want.
panicmechanic
Newly Enlightened
I have been reading here mostly and absorbed ideas, but I think I can give some back now about an MC-E and aspherics.
Please excuse if this posting is a bit confused, I may rethink and rework it later. Photos are not what I do best, and the weather here currently does not allow outdoor shots (it's all white).
I came across the beam patterns of aspherics on an MC-E that trout photographed perfectly. But it was znomit with his idea of tilting the lens and his wiring pattern that helped a lot on my way.
This is what I have now:
The DX 50mm sku.12834 gives a two-lane flood with enough sidespill for curbs and pedestrians. The tiny blue 'flames' on top are enough to light up roadsigns far ahead. The camera does not show the structure completely - there are different brightness levels even where the picture says 'white')
Outdoor comparison. Crappy pics, I was riding at 20km/h, camera in one hand. The bike is equipped with a Nexus hub dynamo. First shot: b&m Cyo Speed 60lux, second a Paniclight prototype using the DX 28mm lens sku.5297 and znomit's wiring.
I had an MC-E mounted on a heatsink and different aspherics that I clamped to a homemade magnetic stand, allowing them to be pushed around freely on a steel base.
-> Photo to follow. Or better, just design your own test rig, as mine is not well thought out and hard to read angle and distance values from.
When I pushed the lens sideways from the axis of the MC-E I noticed the brightest part of the beam concentrates on one side of the outer rim, while still giving a lot of low-level spill on the rest of the beam. This needs some experimenting to find the perfect shape, you will almost certainly have to tilt the lens and/or move it out of the beam axis. With some lenses also tilting the led might be necessary.
The distance to led has to be shorter than the focal length, so you start with that well known sharp-edged disc-shaped beam.
The whole setup is a matter of millimeters, so it's unlikely you will even get proper beamshots if you simply hold the lens by hand. I'm also not giving too many details about tilt angles and offset values, as these will vary a lot between lenses. You might even discover that lenses shipping now are different from what I got. And I would like those who are interested to come up with the findings that work best for them, so your collaboration is more than welcome.
Just a short overview on lenses that show the effect and those that do not:
-DX 28mm http://www.dealextreme.com/details.dx/sku.5297 works well, has a dark area between high and low pair of chips. In a different arrangement, the dark spot is gone but the area close to the bike will be a bit too bright. Might be mended by underpowering the upper pair of chips.
-DX 50mm http://www.dealextreme.com/details.dx/sku.12834 works well, haven't explored further because of weight and size. Definitely better beam shaping quality than the above.
-DX 22mm*7.15mm - http://www.dealextreme.com/details.dx/sku.13570 too shallow, not showing the desired effect. Might be useful as additional concentrator close to the emitter (not tested).
-DX 50mm*22.5mm - http://www.dealextreme.com/details.dx/sku.13667 the stripes prevent any effect. Not useful here.
- DX 44mm - http://www.dealextreme.com/details.dx/sku.4558 too high, the desired pattern doesn't show. Optical quality is miserable, doesn't focus at all and comes scratched by default. Not even useable in a straight setup.
- The one that syc has from surplusshed looks good from here, definitely worth trying.
Some general thoughts on the design:
The lens will almost certainly have to be tilted downwards against the usual position, so it will 'look down' to the ground. But I'm open to surprises here.
The vertical offset can be up or down by a few mm, ore none if you're lucky.
Quite important for making a light out of this is finding a geometry where the led can sit in the center of the body, looking straight ahead as in a flashlight.
And then you'd probably want a light whose body points in the direction of movement, just for viewers' habits.
Electrical stuff:
Compared to 3 leds in series, znomit's idea drops the forward voltage by another 0.25V, as the parallel pair runs at half the current. I measured 2,8V only - may be wrong. To dim this pair further, I want to hook a taillight led + resistor in parallel to it. Taking away 20mA from these two MC-E chips has no visible effect, might be necessary to feed 60-80mA into the taillight if you want to dim the nearfield a bit more. Or, switch the pairs separately, which looks quite promising as a high-low setup. Sadly, I don't have pictures yet.
Optical stuff:
It may sound confusing, but it's in fact the upper two chips that illuminate the near field.
The emitter sits square, rotating by 45° did not turn out well (for me).
A 28mm lens will actually sit close enough to capture most of the light, of course some will be wasted. Have played with small mirrors, but did not manage to add the reflected light to any interesting part of the beam.
Any comments?
Please excuse if this posting is a bit confused, I may rethink and rework it later. Photos are not what I do best, and the weather here currently does not allow outdoor shots (it's all white).
I came across the beam patterns of aspherics on an MC-E that trout photographed perfectly. But it was znomit with his idea of tilting the lens and his wiring pattern that helped a lot on my way.
This is what I have now:
The DX 50mm sku.12834 gives a two-lane flood with enough sidespill for curbs and pedestrians. The tiny blue 'flames' on top are enough to light up roadsigns far ahead. The camera does not show the structure completely - there are different brightness levels even where the picture says 'white')
Outdoor comparison. Crappy pics, I was riding at 20km/h, camera in one hand. The bike is equipped with a Nexus hub dynamo. First shot: b&m Cyo Speed 60lux, second a Paniclight prototype using the DX 28mm lens sku.5297 and znomit's wiring.
I had an MC-E mounted on a heatsink and different aspherics that I clamped to a homemade magnetic stand, allowing them to be pushed around freely on a steel base.
-> Photo to follow. Or better, just design your own test rig, as mine is not well thought out and hard to read angle and distance values from.
When I pushed the lens sideways from the axis of the MC-E I noticed the brightest part of the beam concentrates on one side of the outer rim, while still giving a lot of low-level spill on the rest of the beam. This needs some experimenting to find the perfect shape, you will almost certainly have to tilt the lens and/or move it out of the beam axis. With some lenses also tilting the led might be necessary.
The distance to led has to be shorter than the focal length, so you start with that well known sharp-edged disc-shaped beam.
The whole setup is a matter of millimeters, so it's unlikely you will even get proper beamshots if you simply hold the lens by hand. I'm also not giving too many details about tilt angles and offset values, as these will vary a lot between lenses. You might even discover that lenses shipping now are different from what I got. And I would like those who are interested to come up with the findings that work best for them, so your collaboration is more than welcome.
Just a short overview on lenses that show the effect and those that do not:
-DX 28mm http://www.dealextreme.com/details.dx/sku.5297 works well, has a dark area between high and low pair of chips. In a different arrangement, the dark spot is gone but the area close to the bike will be a bit too bright. Might be mended by underpowering the upper pair of chips.
-DX 50mm http://www.dealextreme.com/details.dx/sku.12834 works well, haven't explored further because of weight and size. Definitely better beam shaping quality than the above.
-DX 22mm*7.15mm - http://www.dealextreme.com/details.dx/sku.13570 too shallow, not showing the desired effect. Might be useful as additional concentrator close to the emitter (not tested).
-DX 50mm*22.5mm - http://www.dealextreme.com/details.dx/sku.13667 the stripes prevent any effect. Not useful here.
- DX 44mm - http://www.dealextreme.com/details.dx/sku.4558 too high, the desired pattern doesn't show. Optical quality is miserable, doesn't focus at all and comes scratched by default. Not even useable in a straight setup.
- The one that syc has from surplusshed looks good from here, definitely worth trying.
Some general thoughts on the design:
The lens will almost certainly have to be tilted downwards against the usual position, so it will 'look down' to the ground. But I'm open to surprises here.
The vertical offset can be up or down by a few mm, ore none if you're lucky.
Quite important for making a light out of this is finding a geometry where the led can sit in the center of the body, looking straight ahead as in a flashlight.
And then you'd probably want a light whose body points in the direction of movement, just for viewers' habits.
Electrical stuff:
Compared to 3 leds in series, znomit's idea drops the forward voltage by another 0.25V, as the parallel pair runs at half the current. I measured 2,8V only - may be wrong. To dim this pair further, I want to hook a taillight led + resistor in parallel to it. Taking away 20mA from these two MC-E chips has no visible effect, might be necessary to feed 60-80mA into the taillight if you want to dim the nearfield a bit more. Or, switch the pairs separately, which looks quite promising as a high-low setup. Sadly, I don't have pictures yet.
Optical stuff:
It may sound confusing, but it's in fact the upper two chips that illuminate the near field.
The emitter sits square, rotating by 45° did not turn out well (for me).
A 28mm lens will actually sit close enough to capture most of the light, of course some will be wasted. Have played with small mirrors, but did not manage to add the reflected light to any interesting part of the beam.
Any comments?
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
A lot of food for thought here. :rock:lovecpf
Sorry to play cheerleader only as my LED's are BO'd.
The inversion of the dies means the focal point falls between the lens and the LED. If you wanted a sharper cutoff on the left, a shade on the right at, or just behind the focal point should do the job, as this is how the low beam in car projector style headlamps work or the shades in commercial theatre spot lights. Want top of beam cut off? Then use a shade across the bottom. Another fine variable to play with, as if mounting angle and offset weren't enough!
Yeah, the 'albino cat on snow' pics won't help much. Be about as good as the pics of a 'black cat at midnight, deep in the unlit coal mine'.
Your pics give a decent idea that you are getting close. How about how it looks to a motorist or pedestrian? Snow won't hurt that test overmuch. Frostbite maybe but we can stand it.
The sad news is that it is very lens-led-mounting specific and someone else can't be sure they will get exactly the same lens from DX with the next order. Trial and error can get costly with a lot of lenses.
Assuming the lens you have is very similar to the ones the reviewers on the DX page received, it sounds like this is a Plano Curved (PCX) lens that is nearly spherical, if not spherical in front profile, and flat or plano on the back. Surplusshed has some small lenses that look to be similar for $4 each that might be XP-G friendly. By description of lens type, they are used "to fill in the corners'. That sounds like a rectangular 'even' hot spot to me. A feasible plan 'B' with some basis in a degree of success. :thumbsup:
Sorry to play cheerleader only as my LED's are BO'd.
The inversion of the dies means the focal point falls between the lens and the LED. If you wanted a sharper cutoff on the left, a shade on the right at, or just behind the focal point should do the job, as this is how the low beam in car projector style headlamps work or the shades in commercial theatre spot lights. Want top of beam cut off? Then use a shade across the bottom. Another fine variable to play with, as if mounting angle and offset weren't enough!
Yeah, the 'albino cat on snow' pics won't help much. Be about as good as the pics of a 'black cat at midnight, deep in the unlit coal mine'.
Your pics give a decent idea that you are getting close. How about how it looks to a motorist or pedestrian? Snow won't hurt that test overmuch. Frostbite maybe but we can stand it.
The sad news is that it is very lens-led-mounting specific and someone else can't be sure they will get exactly the same lens from DX with the next order. Trial and error can get costly with a lot of lenses.
Assuming the lens you have is very similar to the ones the reviewers on the DX page received, it sounds like this is a Plano Curved (PCX) lens that is nearly spherical, if not spherical in front profile, and flat or plano on the back. Surplusshed has some small lenses that look to be similar for $4 each that might be XP-G friendly. By description of lens type, they are used "to fill in the corners'. That sounds like a rectangular 'even' hot spot to me. A feasible plan 'B' with some basis in a degree of success. :thumbsup:
panicmechanic
Newly Enlightened
Thanks for cheerleading!
About the focal point: I'm definitely closer than the focal point of the straight lens, maybe there is a second one when tilting?
I do not especially like shades if the result could be achieved in another way, since I'm planning for a hub dynamo and want as much light as I can have.
Yes, all 'good' lenses seem to be pcx and nearly spherical, maybe a bit parabolic. Hard to measure.
To prepare for variations in lens types I'd suggest making light components adjustable: the led on a threaded pill, as in a flashlight, will allow for a few mm of play lenghtwise. The lens holder could be sliding, so only the angle has to be fixed.
I'll go make coffee and try better, battery-powered outdoor shots later. This needs some preparation. Can't take the Cyo, as running it from a battery is unfair. It consumes less than 200mA at the max. recommended 7.5V.
What I can say from testing: no one complained, the beam just stops below car rear windows/mirrors, just like driving a car. Less bright though...
About the focal point: I'm definitely closer than the focal point of the straight lens, maybe there is a second one when tilting?
I do not especially like shades if the result could be achieved in another way, since I'm planning for a hub dynamo and want as much light as I can have.
Yes, all 'good' lenses seem to be pcx and nearly spherical, maybe a bit parabolic. Hard to measure.
To prepare for variations in lens types I'd suggest making light components adjustable: the led on a threaded pill, as in a flashlight, will allow for a few mm of play lenghtwise. The lens holder could be sliding, so only the angle has to be fixed.
I'll go make coffee and try better, battery-powered outdoor shots later. This needs some preparation. Can't take the Cyo, as running it from a battery is unfair. It consumes less than 200mA at the max. recommended 7.5V.
What I can say from testing: no one complained, the beam just stops below car rear windows/mirrors, just like driving a car. Less bright though...
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
Yep. Any discarded light with a generator hub is highly undesireable.
Good news on the beam height and acceptance in traffic. :thumbsup:
Nice to know it's the PCX lens type: Plan B is viable. I was planning an 11 degree optic and maybe some angled hoods reflecting light down and right. I also want to see what a modified grid shade (from photography) does to get a narrower more precise cut-off beam. The 11 degree optic turns out to be 26 degrees (not close even in horseshoes) so it needs help. If those tests fail or if I am still searching for a better solution: plan B.
A triple CREE XP-G layout in an MR11 shell makes things a bit tight but I can get 2 mm depth easy and another 8 with increasing difficulty. If I can put the PCX lenses on the front 'glass' or in a circular holder that replaces the front glass, then I will have room. Projecting forward, they could help the aero compared to a flat lens and be a tribute to Douglas Adam's girl from Eroticon 6. :naughty:
Good news on the beam height and acceptance in traffic. :thumbsup:
Nice to know it's the PCX lens type: Plan B is viable. I was planning an 11 degree optic and maybe some angled hoods reflecting light down and right. I also want to see what a modified grid shade (from photography) does to get a narrower more precise cut-off beam. The 11 degree optic turns out to be 26 degrees (not close even in horseshoes) so it needs help. If those tests fail or if I am still searching for a better solution: plan B.
A triple CREE XP-G layout in an MR11 shell makes things a bit tight but I can get 2 mm depth easy and another 8 with increasing difficulty. If I can put the PCX lenses on the front 'glass' or in a circular holder that replaces the front glass, then I will have room. Projecting forward, they could help the aero compared to a flat lens and be a tribute to Douglas Adam's girl from Eroticon 6. :naughty:
Nice job! My feeling about the aspherics with a MCE is that a noticeable fraction of the light escapes the edge of the aspheric I tested, even with the short focal length, so there would be even more lost lumens if it were tilted off axis. Then again, maybe a troutie style liberator housing would reclaim them for lighting up the near field. Very little light escaped the aspheric when using the XRE - too bad the XPGs are closer to the MCE than the XRE in terms of native beam pattern.
I've recently moved, and all my gear is in boxes, but hopefully I'll be able to dig it out and work on some stuff in a few weeks. Bending acrylic turns out not to be very difficult, and I found modeling clay that is cheap, and should be easy to work with to form a mold for acrylic sheet. At this point my feeling is that either trying to clone a half-parabola like the B+M reflectors, or using an elliptical optic + a reflective hood (almost exactly like the diagram that pepko posted) are the best way to get a gradient beam without wasting any lumens.
panicmechanic
Newly Enlightened
Back from the cold. The real problem comes with sorting the pics...
Now this is what the beam would look like from the front (minus the starry pattern). The light is mounted above the handlebar, at ca. 110cm, the camera starts from 165cm and goes down to 15cm above ground.
Cyclist's view. With best regards to the people at b&m (they have the nicer tunnel :wave
:
While in the tunnel I shorted the upper pair of chips and got this (Light not horizontal):
Now, out in the fields. Can you spot the snow rabbit?
All pictures taken @ 500mA, F4, ISO100. 6 seconds for normal view, 1/3 second for frontal view.
Last edited:
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
Yeah. When you back the top ones off a tad for balance, you are there. :twothumbs
I think your tunnel beam pic is nicer even if the tunnel isn't quite as picturesque. Great choice.
Maybe a fill just in front, if that is too dark after backing the tops down. I'd love to match it. If you aren't losing much if any light with the angled lens as Syc discussed, then great. Maybe you got a slightly different lens with the same SKU from DX.
I need to order some lenses so I can place/mount the leds to use them.
I think your tunnel beam pic is nicer even if the tunnel isn't quite as picturesque. Great choice.
Maybe a fill just in front, if that is too dark after backing the tops down. I'd love to match it. If you aren't losing much if any light with the angled lens as Syc discussed, then great. Maybe you got a slightly different lens with the same SKU from DX.
I need to order some lenses so I can place/mount the leds to use them.
panicmechanic
Newly Enlightened
Thanks, Brian!
Just consulted the camera manual and found my setting makes for ISO100 But then, the snow is bright enough already, and a rumour says Casios are 'brighter' than other cameras. :thinking:
I'm hoping I can convince more people that beamshaping is priceless fun, much more so than buying optics and shelving them if they don't behave as expected. That makes it cheaper, too, in the long run, if your design allows for some adjustment.
A loss of light is definitely there, as I noticed a fair amount of light being reflected to the ceiling (without the housing). Let's assume the optical efficiency were a lousy 60%:
The M-bin MC-E will give 107lm/die at 350mA, that's 288lm per pair at 500mA. Multiplied with 0.6 there should be 172 lumens for the top of the beam, and another 60 for the rest if I go to 150mA there.
The good news is that all of them are friendly lumens
I'm pretty happy with the current status, will only need to think about connecting a standlight circuit that can run from 2.8V and help reduce current for the near zone. Ideas? Well, may be offtopic here.
For the future, if not Cree or another company 'forces' me to buy a new emitter, I'd like to increase throw a little, by narrowing the top to, say, two thirds. Awaiting input there as well...
And there's so much to try, only I'm slow with that. Single emitters, for instance, or the LED-TECH triple XP-G.
And I need to bring the weight down from the current 92 to 60g, so I could use standard brackets.
Regards, Martin
Just consulted the camera manual and found my setting makes for ISO100
But then, the snow is bright enough already, and a rumour says Casios are 'brighter' than other cameras. :thinking:I'm hoping I can convince more people that beamshaping is priceless fun, much more so than buying optics and shelving them if they don't behave as expected. That makes it cheaper, too, in the long run, if your design allows for some adjustment.
A loss of light is definitely there, as I noticed a fair amount of light being reflected to the ceiling (without the housing). Let's assume the optical efficiency were a lousy 60%:
The M-bin MC-E will give 107lm/die at 350mA, that's 288lm per pair at 500mA. Multiplied with 0.6 there should be 172 lumens for the top of the beam, and another 60 for the rest if I go to 150mA there.
The good news is that all of them are friendly lumens
I'm pretty happy with the current status, will only need to think about connecting a standlight circuit that can run from 2.8V and help reduce current for the near zone. Ideas? Well, may be offtopic here.
For the future, if not Cree or another company 'forces' me to buy a new emitter, I'd like to increase throw a little, by narrowing the top to, say, two thirds. Awaiting input there as well...
And there's so much to try, only I'm slow with that. Single emitters, for instance, or the LED-TECH triple XP-G.
And I need to bring the weight down from the current 92 to 60g, so I could use standard brackets.
Regards, Martin
Last edited:
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
I have some flexibility yet and can always redo the heat sink/mounting. Since the LED's aren't here yet, and no holes are drilled for the leds.
I was planning to mount 3 XP-G's as wide apart and as close to the body as 10 mm optics and MCPCB's permit for heat management. So after these latest posts I thought: three smaller PCX lenses to fix the beams which are no where near narrow enough.
OR I can mount the 10 mm MCPCB's as close to the center as possible, with two mounted lower so they will be top of the beam of a single PCX lens for power at distance and throw, and one led for the lower/closer part of the beam, like a three die MCE spaced out.
There is some play there to mount the lens eccentrically in the bezel (especially a lens smaller than 30 mm) and/or angle either the single or multiple lenses -assembly in the bezel.
Martin:
I see the advantage of honing the light pattern if it doesn't suit. Fortunately, I was planning these lights as an experimental base and designed in some flexibility. Maybe not enough, but some. Looks like they will be a work in progress for a while.
Syc: You mentioned short focal length: how short is short?
Looking at 29-34 mm diameter lenses (to fit the bezel/body) on surplusshed and the shortest FL's range from 28-44 mm or 100-130% of their diameter.
I assume the minimum distance from LED to lens needs to be at least the FL of the lens. Correct?
I might rework to get 28 mm, but would need to set any longer FL lens ahead of the front of the bezel at its loosest, if the above is true. How much space do you guys have between the DX lens and the LED? The 28 mm DX lens is half the cost and no shipping but would require the wait. Worth it?
I can find two 7 and 9 mm lenses with focal lengths of 7 and 10 mm one for each single led whether they have a collimator lens or not.
With the right O-rings, the bezzle could "adjust focus" over 2-2.5 mm.
Thoughts?
I was planning to mount 3 XP-G's as wide apart and as close to the body as 10 mm optics and MCPCB's permit for heat management. So after these latest posts I thought: three smaller PCX lenses to fix the beams which are no where near narrow enough.
OR I can mount the 10 mm MCPCB's as close to the center as possible, with two mounted lower so they will be top of the beam of a single PCX lens for power at distance and throw, and one led for the lower/closer part of the beam, like a three die MCE spaced out.
There is some play there to mount the lens eccentrically in the bezel (especially a lens smaller than 30 mm) and/or angle either the single or multiple lenses -assembly in the bezel.
Martin:
I see the advantage of honing the light pattern if it doesn't suit. Fortunately, I was planning these lights as an experimental base and designed in some flexibility. Maybe not enough, but some. Looks like they will be a work in progress for a while.
Syc: You mentioned short focal length: how short is short?
Looking at 29-34 mm diameter lenses (to fit the bezel/body) on surplusshed and the shortest FL's range from 28-44 mm or 100-130% of their diameter.
I assume the minimum distance from LED to lens needs to be at least the FL of the lens. Correct?
I might rework to get 28 mm, but would need to set any longer FL lens ahead of the front of the bezel at its loosest, if the above is true. How much space do you guys have between the DX lens and the LED? The 28 mm DX lens is half the cost and no shipping but would require the wait. Worth it?
I can find two 7 and 9 mm lenses with focal lengths of 7 and 10 mm one for each single led whether they have a collimator lens or not.
With the right O-rings, the bezzle could "adjust focus" over 2-2.5 mm.
Thoughts?
Here's an image of the MCE dies focused fairly clearly on the ~38mm diameter, ~20mm focal length aspherics I found. Surplusshed may not be selling them anymore.
A 40mm diameter lens 20mm above the emitter is getting all the light that is +/- 45 degrees from the axis of the LED. Here's the radiation pattern per jtr1952's testing:
Granted that you can bring in the lens closer to the emitter to defocus the spot a little for actual use, but at sharpest focus/most intense brightspot, the entire area under the curve to the right of the 45 degree mark on this graph is lost (maybe 25% loss?). So, not counting any losses due to flaws in the lens, you've already given away one quarter of the lumens coming off the emitter.
If you position the lens further from the emitter, you will capture an even smaller proportion of the light coming out of the emitter. Long focal length aspherics are a waste of lumens - you may get better results with a real spot optic that collects a much larger proportion of the total light.
Maybe you can combine a TIR optic and an aspheric or a reflector and an aspheric, but the beam patterns seem to get wonky.
Steve
panicmechanic
Newly Enlightened
I would recommend making a separate lens holder and a heatsink for the led's first and find your settings before you change your housing.
Or buy one :nana:
(At least an inspiring design)
The XP-G will lose more light if lenses are small, due to it's wider beam angle. Anyway, you might start with a single XP-G and see how it does. A triple lens setup in one light will get complicated.
I don't have 10mm boards here, only single 21mm boards and the triple that I linked to above. Must dig for it and re-solder the R4's, or some XP-E R2s, will take photos when I have that. Guessing the beam will be wider at the top if exchanging it 1:1 against the MC-E.
I measured FL of 19.6mm with the MC-E and 20.9mm with the XP-G on the DX sku.5297. Someone else in the DX forum reported 16mm from the dome top with a Cree, most likely an XR-E. Lens dimensions are 28.1mm*9.63mm. In my light the distance must be around 12mm, going straight ahead from the dome base until you hit glass.
For the MC-E and the 50mm*18.6mm sku.12834 FL is 26mm. Please note that I'm measuring from the top of the domes and adding 3mm for MC-E and 1mm for XP-G dome heights.
At Surplusshed some lenses look promising, is anyone willing to take a chance?
http://www.surplusshed.com/pages/item/l3461.html
http://www.surplusshed.com/pages/item/l3786.html
http://www.surplusshed.com/pages/item/l3953.html
http://www.surplusshed.com/pages/item/l10171.html
http://www.surplusshed.com/pages/item/pl1004.html
And then there are replacement lenses for led spotlights, like these, but without dimension and geometry data. Just keep your eyes open...
For first experiments, the 28mm DX is cheap and highly recommended. Just leave some for me...:candle:
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
SYC: Thanks for confirming the focal length issue and hammering home how important it is to output as well as beam shape. The perfect shape at 90% loss ratio isn't good. This ties into the light bench mockup:
No sweat to build. Hard to justify to buy. Thanks for the pic to go by. Wanted to test run the driver boards anyway. So I can set up two mounting patterns and have at it.
:touche:
I thought that the XP-G was slightly on the MC-E side of halfway between the MC-E and the XP-E R2 in beam angle. I agree about complicated.
Likely, and maybe lobed with a wide darker streak. Might be able to trim some to get them tighter, but would lose ability to use any other lens. That's what the test rig's for!
So am I right that you are running the MC-E about 8 mm inside of the focal length? :thinking:
I got two of those on the lens selector, was a little tighter on focal length not knowing what short was. The first and the fourth might be fun to play with while I wait for (one, not all) the sku.5297 to come in as a back up in case the surplusshed ones don't work well.
Order tomorrow. :tired:
No sweat to build. Hard to justify to buy. Thanks for the pic to go by. Wanted to test run the driver boards anyway. So I can set up two mounting patterns and have at it.
:touche:
I thought that the XP-G was slightly on the MC-E side of halfway between the MC-E and the XP-E R2 in beam angle. I agree about complicated.
Likely, and maybe lobed with a wide darker streak. Might be able to trim some to get them tighter, but would lose ability to use any other lens. That's what the test rig's for!
So am I right that you are running the MC-E about 8 mm inside of the focal length? :thinking:
I got two of those on the lens selector, was a little tighter on focal length not knowing what short was. The first and the fourth might be fun to play with while I wait for (one, not all) the sku.5297 to come in as a back up in case the surplusshed ones don't work well.
Order tomorrow. :tired:
panicmechanic
Newly Enlightened
I had an XR-E TIR optic on the XP-G, and it was disappointing compared to the same on an XR-E. I was only estimating that this was due to the optic only being made for collecting ~ 90° from the XR-E.
And in a setup with a tilted aspheric I would worry about heavy losses. But - it's always better to try than to discuss, as I'm the Forrest Gump of optical design...
But while we're at it, I hope to find a meniscus lens where the concave inner surface might be more effective.
Could someone please define focal length in a tilted setup :huh:
A circle around where the emitter chips must be, with the focal length as radius, will intersect the lens somewhere above the center from the flat side, and near the bottom at the bulged side.
Last edited:
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
I have some optics for the XP-G coming, and they are derived from XP-E, and they do give a wider output from the wider input, and there is a thread in the MTBR forum DIY light section on the XP-G optics. I hope to get a reasonable high beam.
http://forums.mtbr.com/showthread.php?t=575994
Looking this up in Wiki to refresh my freshman college physics, suggests to me, that since these are relatively thick lenses with an infinite radius (flat side), that it would be dumb for the manufacturer to use the center of the lens definition also known as the thin lens approximation for focal length. So the focal length for these lenses is likely the distance from the front apex of the lens to the focal point. In other words, the lens thickness is included in the focal length. That is where I went wrong. I was thinking the middle of a thin lens, but actually adding the distance to the back of the lens. :shakehead This helps a lot with 20-30 mm FL and lens thickness of 15 mm or more in physically installing it on my light. :thumbsup:
As I understand this (I may have misunderstood), regardless of the exact definition of focal point in these lenses, the focal point won't change by tilting. By the equations, it is a property of the lens curvature(s) and thickness. If you keep the focal point of the lens directly ahead of the intersection of the dice, then you will change the relative distance of the upper and lower LED dice from the focal point and so the amount of spread of the light by the lens will change for closer versus farther dice by tilting the lens.
:candle:
http://forums.mtbr.com/showthread.php?t=575994
Looking this up in Wiki to refresh my freshman college physics, suggests to me, that since these are relatively thick lenses with an infinite radius (flat side), that it would be dumb for the manufacturer to use the center of the lens definition also known as the thin lens approximation for focal length. So the focal length for these lenses is likely the distance from the front apex of the lens to the focal point. In other words, the lens thickness is included in the focal length. That is where I went wrong. I was thinking the middle of a thin lens, but actually adding the distance to the back of the lens. :shakehead This helps a lot with 20-30 mm FL and lens thickness of 15 mm or more in physically installing it on my light. :thumbsup:
As I understand this (I may have misunderstood), regardless of the exact definition of focal point in these lenses, the focal point won't change by tilting. By the equations, it is a property of the lens curvature(s) and thickness. If you keep the focal point of the lens directly ahead of the intersection of the dice, then you will change the relative distance of the upper and lower LED dice from the focal point and so the amount of spread of the light by the lens will change for closer versus farther dice by tilting the lens.
:candle:
Last edited:
BillyNoMates
Newly Enlightened
Using a positive meniscus lens is something I have considered as well, but I have come across a few sticking points.
1) Generally available meniscus lenses have relatively long focal lengths, therefore the light source must be placed inside the focal length if enhanced light collection is required.
2) Since a single will not collimate the light (because of point 1 above) a multiple lens set-up will be required. When I've played with multiple lens solutions I find I lose a lot of light through reflections off the lens surfaces. This means AR-coated (more expensive) lenses are required in order to maximise light transmission.
3) My sketchy knowledge of optics is likely to cost me a lot of money if I keep experimenting as keep having daft ideas and I'll end up with a lot of lenses that are no use to me.
To try to increase my knowledge I've trying to write an elementary optical ray-tracer. It is more-or-less working now and I've found it very helpful (especially with multiple lens configurations). However to use it properly I need to know the lens parameters (surface curvature, aspheric coefficients etc). It is possible to use the lens-maker's equation to derive some of the parameters but I need manufacturer data for aspherics. Fortunately, good optical suppliers provide this information (I get most of my info from ThorLabs.
Hopefully, I'll be able to try-out different lens configurations to get some estimate of which schemes are likely to work before I commit to buying.
By the way, in my experiments with aspherics, I noticed that the glare from a cheap lens can still be a problem for on-coming motorists. AR-coated lenses should have a much lower off-axis glare - has anyone on here tried that?
BrianMc
Enlightened
- Joined
- Nov 4, 2009
- Messages
- 940
[FONT=Arial, Helvetica, sans-serif]L4056[/FONT] [FONT=Arial, Helvetica, sans-serif]PMN [/FONT][FONT=Arial, Helvetica, sans-serif]28.4 [/FONT][FONT=Arial, Helvetica, sans-serif]28 [/FONT][FONT=Arial, Helvetica, sans-serif]Y [/FONT][FONT=Arial, Helvetica, sans-serif]$4.00[/FONT]
[FONT=Arial, Helvetica, sans-serif]L10016[/FONT] [FONT=Arial, Helvetica, sans-serif]PMN [/FONT][FONT=Arial, Helvetica, sans-serif]30.0 [/FONT][FONT=Arial, Helvetica, sans-serif]26 [/FONT][FONT=Arial, Helvetica, sans-serif]Y [/FONT][FONT=Arial, Helvetica, sans-serif]$4.00[/FONT]
[FONT=Arial, Helvetica, sans-serif]L8739[/FONT] [FONT=Arial, Helvetica, sans-serif]PMN [/FONT][FONT=Arial, Helvetica, sans-serif]29.0 [/FONT][FONT=Arial, Helvetica, sans-serif]28 [/FONT][FONT=Arial, Helvetica, sans-serif]Y [/FONT][FONT=Arial, Helvetica, sans-serif]$4.00[/FONT]
Searched surplusshed for PMN (positive miniscus) from 28 to 34 mm, FL 20-30 mm. Assuming the dome is 15-20 mm (no pics or specs to determine), this would equal 6-13 mm from plane of bottom of lens (mounting point) to focal point, correct?
I assume the PMN lense with short FL let you do a single lens setup?
As a stab at likely lenses for the job, what do you think?
If you think you'd spend your money trying one of these, which or would you try them all? I will share what I find.
I agree that all permutations and combinations render a random two lens approach several lifetimes' work and a lot of money if done trial and error.
Yeah, a way to narrow the choices would help a lot.
In the case of these surplus PMN lenses, they have AR coating and no more expensive than PCX lenses without. Nice to know it could help.
As to some side spill still searing someone's sight, I think we hit a point of diminishing returns and for me, exasperation. I rarely got highs with the HID moved more at the ditch when vehicles approached as it had no other setting than off. Of course riding the side spill wasn't the best. Fortunately, pedestrians at night here are extremely rare or they'd have gotten a tan. Any response that looks like dimming gains a lot of goodwill from motorists.
Because my low beam is a work in progress, and so of dubious effectiveness until it's 'perfected', I have options. In addition to reducing power, aiming, or shutting off the high beam, I have kept three additional fall-back options for the low beam: I can swivel the light more to the right and down as before, drop its output, or both. Riders used and still use halogen MR11 dual headlight systems that are no where near as traffic-friendly as that, so I think we can get a bit too focused on cut-offs (though on a pretty girl...but I digress).
Here, manufacturers can't place headlights over a certain height from the road, but the low beams of modified 4WD sport-utility vehicles are often at driver height and seem to be aimed right at you, as a motorist. So as the cartoon character says: "It could be lots worser!" and too often, it is. Motorists here are often ecstatic that you aren't riding ninja! So well conceived and operated bike lights will be appreciated by most.
I will test tolerable settings for my low beam and hold to them as much as possible. Any motorist who hits their highs then, might find I fire back up a bit too soon and see the power of my fully armed and operational 6 R5's at 1 amp and 1300 lumens! (Cue 'Empire Theme' here.) Fire at will (or wilma, I'm not sexist!.)
:devil:
Last edited:
