Designing good daytime rear commuter lights

[BrianMc]

For those wishing to see 6 XP-E orange reds in a blinky you can serch YouTube for 'DTK10W LED tail light comparison Superflash'. This video reminded me that my reflector tailights do not show from the side at all except for the red patch on the ground and lit up signs. So I added an extended plastic bubble lens repurposed from a cap for a pressurized caulk can. One light moddified:



I videoed it tonight and it shows very well and does not diminish the rear output noticeably. Unfortuantely, I forgot to rotate the lens so the video was sideways. So video will await the delivery of the new blinky tail lights.

[1 what]

Hi Guys,
Good to see this developing....I still think the tail light is the most important light on the bike!
BrianMc - I could not agree more with your thoughts about the need for side visibility - your current model (above) is looking like the "benchmark" All you need is a Chinese driver with lots of funny flash modes to attract even more attention.
I've still not been pulled up by the Law for using outrageous flash sequences.
Savvas (post 119) the optic is DX SKU1916 (48.16mm 3-LED Reflector for Cree Emitters (Glass-like Plastic). The heat is managed by derving the LEDS at 50% of max forward (pulsed) current with a 10% duty cycle of a 2Hz square wave. The final result is a rapid strobe. Also the backs of the LEDS are bonded to a 40mm diameter piece of 2mm thick Al so there is a little extra metal behind them but I doubt if they need it. I know that I could drive them to max current but as I said earlier I felt they were plenty bright enough at the reduced drive and heat was not a problem.. The extra surface area really increases apparent visibility.
Having said all that.....this ongoing debate is starting to make me restless - I might have to have a think about what new ideas you chaps have missed there has to be some new design that will push the development in a different direction. I can't let BMc get too many light years ahead of me!

[pe2er]

Hi Brian,

... So I added an extended plastic bubble lens repurposed from a cap for a pressurized caulk can.

You've unleashed a brightest-taillight-contest here on CPF it seems Like your mod with the cap. Looks like visibility from the sides has improved a lot!

This video reminded me that my reflector tailights do not show from the side at all except for the red patch on the ground and lit up signs...

Don't think it is wise to only rely on the red patch on the ground and reflections off signs to be seen.

This is the most recent tail light I built. It is part of the Dutch Recumbent Trike Lighting Project posted elsewhere on this forum.

I Made it easy on myself and used the plastic lens of a Spanninga ARCUS XB taillight I found along the cycle path.

Instead of the 5mm LED, I mounted a 1 Watt red LED spot from Dealextreme.

The 1 Watt 'Luxeon' like LED, MR16 driver and lens are all glued to an aluminium plate and mounted on the trunk of my Trike. You can just see the driver to the right of the LED through the plastic in the photo above.

The effect of the lens/ reflector is a more diffuse light than I normally build with the same red LED spot and lens/diffuser from Dx. It is a more 'friendly' light, but daytime visibility is a lot less than my previous lights. Although in this picture, it looks quite impressive Side visibility is good. Power comes from two Li-Ion cells that drive all the lights on this recumbent.

[BrianMc]

Hi Guys, Good to see this developing....I still think the tail light is the most important light on the bike!

Missed you. The difference I get with overtaking drivers with weak Superflashes versus my current lights proves that good tail lights are very important. I also need powerful headlights because they dismiss me as 'only a bike' doing 5 mph otherwise with left and right hooks galore. 'Made you look' is an essential aspect of cycling safety.

BrianMc - I could not agree more with your thoughts about the need for side visibility - your current model (above) is looking like the "benchmark" All you need is a Chinese driver with lots of funny flash modes to attract even more attention.
I've still not been pulled up by the Law for using outrageous flash sequences.

I don't know about 'benchmark'. I think Leopold Porkstacker's (MTBR Commuting subforum, Another thread about lights, thread) somewhere around 600 lumens out the lenses, has that distinction. He uses a Phillips buckpuck driver so no flashing, though.

I thought retired Superflashes would help, but they are underwhelming in that role even at night. Placement is a problem as well as the extra weight for so little effect. My cruder video attempts showed so-so from 45-75 degrees off axis and nothing from the two DIY 1 watt lights from maybe 75 degrees on up other than road and sign spill. NOT adequate. I can see this one from the front angles while the healights are still visble! Not sure of daytime. Will check. The Radbots have some side output but are hardly noticeable at 100 feet away in a ride-by at night. Nada for day. I will supply video of the two treated lights, but I was really impressed with the prototype test. There are lumens enough to spread some of the light a bit more so this worked.

Input from so many here in this thread contributed to the idea. An early post showed a lipstick or phallic tapered lens suitable if I wanted a '59 Caddillac finned light system I suppose. Then we had the lantern tail light idea from NZ, but film cannisters where too opaque to mimic it. This cap is just a bit milky and not perfectly clear, which I hoped and found to be ideal or close enough. I took my light bezel and was trying different products caps for fit. That must have freaked out security watching the monitors in the store! I am doing the other and will get photos of the modding. The caulking will be used to reseal doors and windows so the cap is free.

If the output is still a bit bright for night, I am thinking of a Taskled driver to give frenetic 3 flashes per second and the ability to easily switch between a Hi and Lo mode. I'll let blinkys add crazy stuff.

Having said all that.....this ongoing debate is starting to make me restless - .... I can't let BMc get too many light years ahead of me!

Data here says that close to 70% of car-bike collisions are from the side. So our safety is our best payback. Have at it!

Hi Brian, You've unleashed a brightest-taillight-contest here on CPF it seems Like your mod with the cap. Looks like visibility from the sides has improved a lot!

What my crude video attempts told me was that I was not nearly as visible as I thought I was based on the lights up close. If other cyclists reading this check this themselves and up their side visiblility, it may save injury or worse. I was in this race only to the point of helping drivers who are actually looking, to see me. I can't fix stupid. I am very close to that visibility goal if not there. If I need more power, I know a cheap way to do that thanks to Leopold!

Don't think it is wise to only rely on the red patch on the ground and reflections off signs to be seen.

Yep. Exactly. If there were more cyclists with light here cycling at night, I'd have worked this out a lot sooner. Hard to ride and observe yourself at a distance without a video camera.

This is the most recent tail light I built. It is part of the -Dutch-Recumbent-Trike-Lighting-Project posted elsewhere on this forum. I made it easy on myself and used the plastic lens of a taillight I found along the cycle path. ... I mounted a 1 Watt red LED ... from Dealextreme. ......Side visibility is good.

A free purpose designed lens is a great way to a good cheap light. BTW for those just catching this discussion, it was Pe2er's post on this LED that was key to my <$40 twin tail lights. So someone only need make a triple with this LED that won't overheat in full on mode and they are ahead of me in this race, right 1What?

[1 what]

I don't know if day time visibility is as simple as "brightness" but I do have a qusetion.
Do you think side lighting is as important during the day?

[belfastbiker]

For side on visibility, I add multi-colour flashing spoke lights...

[BrianMc]

I don't know if day time visibility is as simple as "brightness" but I do have a qusetion.
Do you think side lighting is as important during the day?

Such a seemingly simple question. So many aspects. Spent some time thinking on it. So a long post.

If you spread out the light to wider angles without more power you drop output into the skylight range and may as well save the weight, battery power, and the rest IMHO.

Vehicular bike-vehicle accidents are about over twice as frequent (70%) from the sides including left and right hooks. So it is a topic worthy of consideration. I have not seen that broken down into day/night, or unlit bike /lit bike, or lighted street /unlit street. All of which would help us figure out a best approach.

Early video attempts here.

Until recently, I used behavioral observation to get a handle on this. (Still learning the new video camera but it looks like it will do.) Once I passed a certain lumen output in the headlamp and especially with the stacked head lamp/helmet lamp setup, I seem to mesmerize drivers in full dark with or without street lighting. They yield like I was the police! Once I kicked up the rear output so that if riding with others, no one would want to be behind me, passing cars get fully into the left lane (good because the lanes are narrow, and I am taking the lane, so if the ride on or inside the center lines they are closer than 3' and unsafe). For drivers coming to cross streets and looking long enough before rolling through the stop illegally, they see me and yield appropriately or not depending on my distance, as it should be. Ones approaching in a manner suggesting they may not see/stop and so hook me, get me looking at them with the helmet light. It is pretty narrow and brightens rapidly as I approach and works day or night. So far, five drivers who I suspected were going to blow the stop and endanger/hit me did not. Not all of them were going to, but my guess is at least 3 of the five were based on watching from my car over time. That is the problem you can't run exactly replicated trials with and without to prove it statistically. I am convinced it is an effective technique and one that requires my full attention, the best safety device a cyclist can have. Cycling safety is NOT passive.

The take home message here seems to be that good lighting at night offers me more sharing of the road and good behavior than I get in the bright daylight. Partly this is because most people at night aren't as distracted or in such a huge hurry and are aware they are not seeing as well so err on the side of caution. In part it is lower traffic to distract tehm. In part, the lights make me "something unusual' and get their attention. I may be a bit excessive with the beam of the DIY tailights (they are brighter than car brake lights juded in person ate 1/8 mile in a low sun shining on them, two observers).

In the bright daylight, nothing seems to beat a good reflective, light-colored upper wear, short of a police light bar and siren. Stick THOSE on your helmet! Shorts are not effective as the thighs are at oblique angles a lot of the time. Full pants woud be OK, but of a seasonal nature. The tops are more visible the more erect your upper body. An areo bar position is not a good one for being seen at any angle. But some days are cloudy and rainy and light help a lot with side visibility then. With a low sun behind the cyclist the bright top almost glows. Easily seen in a lousy video camera at 1/4 mile, so maybe a mile for a person. With the oncoming driver driving into the sun and making a left turn, you are in shadow. I have't videoed this yet, so this is from observation only. Firstly, I hate riding at these times with a lot of traffic. So I haven't a lot of experience by choice. I use my high level on the headlights and helmet in flash mode for this and bright daylight. Fine if you have no other traffic but you two. Assuming the oncoming driver is paying expected levels of attention to driving. A huge assumption, btw. But if you are being overtaken and eclipsed from view to the oncoming driver, it is another matter. They forget you were there with even one vehicel blocking their view. If they used their turn signal it would alert you that they are turning before they were blocked from your view, then you'd be warned and could pour it on to clear the drive/intersection just in case. A majority of drivers think this a useless gesture, so don't bother or do so too close to their turn. This means they plan to swing into their turn behind the passing van, truck, or SUV that you can't be seen behind. So the driver has little time to see that nice headlight or tail light with side output. Or if they are intent on running a stopsign and you are entering a full side view in front of them, I am not sure how much direct side lighting will help over just being visible, but that is the scenario I think it may help. We are not going to carry huge search lights warning of our presence behind vans or about to enter an intersection. I also suspect that no lighting at all is as bad in the opposite direction. The law requires nothing but reflectors, and the addition may not be overly effective, if injured with my bike as eveidence at a hearing, you admit to not seeing THIS!? Case closed.

So being able to see the side output of tail lights at slightly forward to rearward side angles may not help much in the day but if it helps once, it is worth some effort. What I have learned in video is that a driver at 100 feet and closing on an intersection, is at a very poor angle to see your headlights and tailights, unless you turn your helmet light that way, or you swerve so your headllight is aimed more at them briefly. I want a repeat at the 50 foot and stop line position video to see how things show up as they stop (well, sort-of). But if the driver is doing a fast check before rolling a stop which is very common here, then my bright head lights and tail lights don't show well and the reflectors show too late or too little if the car has its lows on, and it is dark enough. In some ways, I am still struggling with how to ask the question and I hope the camera will provide insight. Based on what I have seen by adding the bubble to the one light, I am impressed with the increased visibility. Whether it will help a driver is another question. I expect more time to chase this soon.

[1 what]

Hi Brian (and Guys),
I couldn’t agree more especially with your thoughts about “behavioural observation” and how car drivers react to bright lights at night.
I was going to build a prototype for what follows but I don’t have much time at the moment so will share it with you verbally rather than try to “one-up” your design.
The short version is that it’s dawned on me that the requirements for night time visibility and day time visibility are very different AND probably conflict.
For example a tail light bright enough for day time use will be dangerous (blinding and distracting) at night especially for drivers coming up behind the bike.
Furthermore if we consider side-on visibility the simple strategies for achieving it at night (such as extending a diffuser tube so it’s visible from the side) might represent a self defeating day time strategy.
As you said above:

“If you spread out the light to wider angles without more power you drop output into the skylight range and may as well save the weight, battery power, and the rest IMHO.”

So…..How about a tail light with 2 modes: 1 for night and 1 for day. Needless to say that I believe both modes should flash to increase their visual impact.
The night time mode should be bright but not blinding and give good sideway light distribution. We have a number of lights that do this.
The daytime mode should be very bright with a tight backwards focus so as to make maximum visual impact on vehicles coming up behind the bike especially if both the bike and the vehicle behind it are moving towards the setting sun (IMO a very dangerous time to be on the road on a bike).
I’m tempted to tell you how I see the rear light looking but won’t since I’d like to see what CPF’ers think of the idea and how they might build it.

How to deal with side on visibility in day light (if we use a tight focus rear light)?
Certainly hi-visibility clothing helps and I’ve posted elsewhere my thoughts about maximizing “passive” reflectors (especially 3M tape) on bikes at night.
BUT rather than try to make a compromise between rearward and sideways lighting from the one taillight (in daytime) why not use our front (forward facing) light(s). We already have them on the bike.
They could simply be rotated “outwards” – see diag.


You could use a Fenix style mount that lets you rotate the light or any number of simple devices. Surely a CPF’er with a lathe could make them up for us (plus make a few $).

What do you think?

[BrianMc]

We consider Jane Goodall observing Chimpanzees as valid science, yet often consider our daily observations cycling as anecdotal. They are in the tradition of Anthropology, and Margaret Mead. A perfectly valid way to asssess the effects of a change in action or equipment. We just need to think about what we observed and make sure it was what we saw and not what we wished we saw. CPF can provide the peer review, and we move forward.

OK. My back lights are mounted so I can change their side and down angles considerably. I was very pleased to see that the main beam still was bright straight back in the dark house. Real world needs to be tested. One of the many videos I want is day versus night with and without the bubbles, and with the lights straight back, say 22.5 degrees out off-axis and 45 degrees out off-axis. A driver I can change the levels and add a flash mode would be good. I need only 400 mA so a cheap Chinese driver with modes and good for 12.6 volts will do.

The 1.4 A output at night is borderline on social and I have a half power low I can easily flip between. My ride bys have convinced me that I need at least the 1.4 A output in flash mode in daylight to penetrate cell-phone fog. The Dyno light is only OK. 3A in flash might be interesting. The angled headlight idea also came from the ride-bys. I have a B&M mount widened to take a coupling nut between the ears and cantilever the lights either side of the braket. No independent aiming of the lights is possible in the horizontal plane with this mount as is. No quick release either. A new mount is called for. I think a 45 degree splay would be the most needed, but will verify with video.

I think reflective devices are helpful. Their performance has dissapointed me to this point in videos. I noted that my cars low beams come to about my top tube so my vest and reflective tape doesn't show well. In spite of a fairly wide headlight beams, I was not seeing much return. This may be a video effect, and the better camera will answer that.

The ideas have sufficient merit for further testing.

[CKOD]

Hi Brian,



You've unleashed a brightest-taillight-contest here on CPF it seems Like your mod with the cap. Looks like visibility from the sides has improved a lot!

Everyone stop beating around the bush then
http://www.luminus.com/stuff/content...lumination.pdf

rated for pulses upto 30A, and absolute max of 36A. Drop the duty cycle of the blinking to keep the power usage in the sane reigon, get some maxwell boostcaps for the high current source and make something that really shines


Mostly teasing though, I know eventually the brightness would end up more of a hazard for the cyclist then helping.

[BrianMc]

We can do power. A guy on MTB has 6-900 lumens in red narrow and red wide lights. Getting into paint peeling territory and law enforcement pulling you over, at that kind of power. I figure about brake light level or a bit above is good enough. It is power and the right spread and possibly a day and a night mode plus a ride in groups (I just unplug the big guns).

Here is the pair of DIY lights with one bobble lens mounted above Radbot 1000's (not on):



Here is the bubble lens source and the pair ot lights before the second bubble lens:



The doubly modded light sans power (if you can't intuit how to do it, don't bother):



Here they are with power and flash for the pic:



Lastly, no flash to reveal the output better:


By brianmcb at 2011-05-02

It is clear that there is a robust light beam even with the bubble lenses. If you don't mind me saying, so. THAT's what I'm talkin' about! Hope it translates inot real world. The side mounted lights allow easy angle adjustments up/down & in/out.

I have modded two Marwi handle bar mounts to fit on the B&M fork crown mount and allow angling of the twin headlights. Weather permitting, I am ready to repeat ridebys with video to test how well the bubbles work for side visibility, how well the angled headlight and tailight idea works, what angle(s) seem best, and how the new wider angle Danger Zone or more powerful Turbo blinkies work compared to the DIYs. The camera's zoom feature is working now (or I learned HOW to owrk it, is more like it) so that will help.

[BrianMc]

The winding road at the end of my drive is not the place to record the effects of angled lights. as I snake toward the 'intersection' Often prelim experiments identify factors that need to be controlled and this is one of those things.

I zoomed the camera for the angle I see glasses frame edge to frame edge, if looking straight ahead. That was good. Repeats at 50 and 25 feeet will be needed, too.

I have some observations. Both my head lights and tailights need to be splayed out at 45 degrees as 30 degrees was only a small improvement on seeing them coming at 100 feet from the intersection. The headlights in particular brighten considerably as the beam aims at the camera. There is a distance between the headlights passing and before the tail lights arrival, just before, during, and after broadside-on where no lights are seen. Straight on from front or rear suffers but oncoming is suppossed to be a lane to the left, so no issue from the front. A car overtaking and in the left lane will see the headlight well as I head for the shoulder or ditch! I tried the Radbot 1000, the Danger Zone, and the Turbo firing straight back and conclude that on an overcast day they were about the same. The Turbo is an improvement over the almost unseen Superflash, but a wider more powerful helmet blinky would be better, if not too heavy. Comparing the three 1 watt (the DZ is .5 + .5): More duration on the PDW lights, more peak flash on the Turbo but its short duration made it seem close, and the Radbot is the best of the lot from straight back IMHO. Any two of these create a visual ruckus flashing out of sync. You get redundancy and more oomph.

I mounted blinkies firing sideways from the top of the rear fender to test lights for the broadside on. Even just cruisin' in the 12-15 mph range, there are so few flashes in the 2 seconds of traversing the 'intersection' that I had to search for them frame by frame. Underwhelming. If these would work at all, it will be as steady-on side marker lights as on larger trucks and 1 watt is likley the minimum to do much in dayl;ight. Mounting My DIY lights sideways on a rack and let Blinkies handle rear duties may be an option. My leg hits them when they are at 90 degrees in their current mount point.

Using two 1225 lumen sources (CKOD) to produce two 180 degree wide (or close) by, say 5-15 degree high beams, and those beams overlapped for 30-45 degrees in the horizontal in the back leaving the same angle clear forward, with a means to tone them down for night, would be interesting. Didn't look up how much power that would take. As long as any angle was about car nrake light intensity, I can't agree with a complaint of 'too bright'. A lesser 3-500 lumen 9 dice LED was used in a flood and a medium lensed pair of lights by Leopold Porkstacker in the MTBR DIY lights, and Commuters forums. With atha much power you NEED to spread it out. Which will give you side visibility.

So I will document my lights in a more suitable spot.

[Offroad'Bent]

Anybody know if the Magicshine tail light will run off 1x 3.7v battery? I know it's meant for 7.2v, but not sure how it's regulated.
It might be a decent daylight-visible helmet rear light.

[BrianMc]

Since the cutoff voltage (to avoid damage) for a nominal 7.4 volt pack is between 5 and 6 volts, I doubt the MS tail light will run, or more to the point, shouldn't.

[pe2er]

What do you think?

I couldn’t agree more especially with your thoughts about “behavioural observation” and how car drivers react to bright lights at night.

My (admittedly subjective) observations support yours. Better visible light earns you more respect from motorized traffic. Too much (too bright) light causes irritations (and cars/ scooters flashing their lights at you).

The short version is that it’s dawned on me that the requirements for night time visibility and day time visibility are very different AND probably conflict.


For example a tail light bright enough for day time use will be dangerous (blinding and distracting) at night especially for drivers coming up behind the bike.
Furthermore if we consider side-on visibility the simple strategies for achieving it at night (such as extending a diffuser tube so it’s visible from the side) might represent a self defeating day time strategy.

I Agree with that.

So…..How about a tail light with 2 modes: 1 for night and 1 for day.

That would be a pragmatic approach. For me, I also need a brake-light function in the tail light (for use on my Strada VM).

Needless to say that I believe both modes should flash to increase their visual impact.

I Would not use a flashing light at night. I am of the opinion that although a flashing light does attract more attention, a solid light makes it easier for the observer to estimate your speed and direction of travel. Flashing lights are also not allowed by law over here (Holland).

The night time mode should be bright but not blinding and give good sideway light distribution. We have a number of lights that do this.

Yes, just a 1 Watt red LED in a standard tail light housing fulfils those requirements. Lights like the DiNotte 140R lack in the ‘good sideways light’ department.

The daytime mode should be very bright with a tight backwards focus

Yes, but mainly because it is impractical to also have good visibility from the side in daytime.

I’m tempted to tell you how I see the rear light looking but won’t since I’d like to see what CPF’ers think of the idea and how they might build it.

I Am curious.

My thoughts at the time are along these lines. Use a standard light like this to modify:

For daytime visibility, use a 35mm lens/ diffuser and a 1 Watt red LED like the lights I built before:

To the Left and Right of that lens, add three or four 1 Watt red LEDs behind the transparent (red) reflector.

One red LED behind the reflector should be sufficient at night This is what I have now on my recumbent trike. The other LEDs (two or three) would then come on when I brake.

How to deal with side on visibility in day light (if we use a tight focus rear light)?
Certainly hi-visibility clothing helps and I’ve posted elsewhere my thoughts about maximizing “passive” reflectors (especially 3M tape) on bikes at night.

Nothing beats a Hi-Vis jacket in bright sunlight.

BUT rather than try to make a compromise between rearward and sideways lighting from the one taillight (in daytime) why not use our front (forward facing) light(s). We already have them on the bike. They could simply be rotated “outwards” – see diag.

This thread is about rear commuter lights, but I do have some experience with that on the Strada VM. I Use different lights for night and daytime visibility. A Nice B&M IQ Cyo with StVZO approved cut-off above the horizon at night, and two 1 Watt LED lights with normal ‘flashlight’ patterns aimed straight forward in daytime. For me, that is easier to realize than tilting lights. Plus I can also use the daytime’ lights dimmed in the night to enhance my visibility.


Erwin, “One of the Guys”

[BrianMc]

A SUNNY day! Been very rare this spring. Over 18" of rain since April 1.

I set the camera to cover the same width as what I see straight ahead at the same position as the camera (within my glasses frames), without a head turn. Real people would turn the head, but the fisheye effect of the video kicks in if I made a wider view and it becomes unclear how well the video stands in for a driver. I made one at 50 and one at 100 feet, and they say the same thing as the 25 foot video.

The headlights and taillights are angled out as much as I could physically without new mounts (about 45 degrees).

Flashing blinkys aimed sideways were easily misssed. Moved them to the top of the rear fender to simulate a rack mount and to avoid my legs hiding them. Even on full the Danger Zone with its wider angle than the Turbo, can be seen, but isn't really saying 'Bike!' very loud in bright sunlight

http://imageshack.us/clip/my-videos/695/riy.mp4/

The angled headlight anounce the bike right into the 'intersection' straight in front of the camera. The angled taillights show well just before the bike exits left and then right. The spill area of the beam isn't as strong as the headlights (20%) so that isn't surprising. Drivers turn their heads so these taillights would be seen sooner and better. The right hook after the bike passes should be less likely as the driver turns throught the beam of the right one and into the beam of the left one.

Th rear aimed Turbo and Suprerflash add no daytime side lighting. Unsuprising. The Danger Zones sitting on the fender can be seen. But full sunlight isn't their strength. I think this says you could use 100-200 lumens in an oval beam on each side for the day. Having already seen the helmet light at night, I think the bubble lenses will provide more than enough side output at night.

So lights bordering on too powerful at night for street use are OK and help some in full daylight when angled out. They will offer a lot on overcast days or when riding in shadow. There is a period of time when the bike crosses in front of a car approaching the intersection, when the headlights can no longer be seen and the angled tail lights are not yet obvious. This broadside (or nearly so) time could use a day specific light. I suspect that the actual risk of a driver not seeing you in front of his grill is small, and for that small fraction of drivers who cant see big yellow school busses or emergency vehicles with flashing light arrays, maybe a bright light won't do diddly. All human endeavors have risk.

I will try to reride this on a cloudy day to see if clipping a pair of high power blinkys to fire sideways off a rear rack helps that much with less skylight to deal with.

If you have twin headlamps and tailights, this angled technique, seems to work to allow a good night system to be a good day system.

[1 what]

Under construction:

EDIT (Sunday)
Somehow I installed the wrong driver board and have just blown up both the red crees!
Must be getting old and careless.
Since the whole assembly was sealed with epoxy to make it waterproof it's now back to the very start so don't hold your breath.
Grrrrrrrrrr!!

[1 what]

Further to post 137:
Took delivery of some new Cree XLamp XPE’s in red and rebuilt the project!
Made sure I had the right driver this time so as not to vaporize another 2 led’s.
The driver is one of the 3W DX models with a strobe mode and memory.

As stated in an earlier post the aim was to have 2 taillights in 1. One of them to be very bright and focussed in a tight beam for maximum day time visibility a long way behind the bike and the other (for night time use) less bright so as not to blind drivers. The night one should also have a wide horizontal spread and minimum output on the vertical axis (no sense in wasting lumens). Much to my surprise I found the 35mm optics/light diffusers from DX (SKU1919) did this job well and only required a small modification to fit the Crees. These diffusers also give excellent side visibility and if mounted on a 25mm upright they have enough “hangover” at the sides to be very visible in front of the bike as well. I guess I’m trying to say that they give a good spread of output through almost 360 degrees in the H axis.

The photo does not capture how visible the light is in daylight. It is much brighter than car brake lights and I’ll try to get some better shots.

The concept evolves!

[BrianMc]

I like the one unit two modes so you always have the right light for the job. At dusk, is the switch 3 position so both can be on?

Cameras don't seem to do these justice compared to what the eye sees. I wonder if this is a relative brightness processing in the visual cortex. Like the relative color thing when seeing white as light blue if near a blue car, slightly pink near a red one. Too bad you didn't place a brick on the car's brake pedal. Then the relative brightness would have been confirmed. Using a camera as a measuring tool is trickier than you'd first think.

[1 what]

Hi BrianMc.
2 position only....But I'm working on a different idea that will solve this problem...might be a couple of weeks.
Almost as soon as I finished this light I realized that a different design might be much better.
Re the camera - I suspect part of the problem is the limited resolution of the camera especially when you try to capture a small object a long way away..
As for the brake lights - Since it's not my car and not my driveway I dont know what the owner would have thought when he saw me approachinh his car with a brick!!
Have a good weekend.

[BrianMc]

I what: Tell the owner that it is a bare bones budget documentary for a school project to see just how far away people can easily see car brakelights or a bike light. A white lab coat with clipboard, forms, and a pocket protector with pens and a laptop would complete the disguise. Heck you could likely find a kid who'd love to turn it in as their own work. Save them stealng it off the internet.

[1 what]

Beamshots for post # 138:

[BrianMc]

Gonna try that. But not tonight.

[Savvas]

I have been following this tail light thread with a great deal of interest - many thanks to all of those who have contributed such original ideas. I am particularly impressed by OneWatt's all plastic triple tail light! However I wonder if any of the insights gained in this thread could also be translated to use with dynamo (generator) front and rear running lights? I'm very much a 'dynamo man' myself although I'm having a flirtation with battery lights at the moment, seduced (I think) by the sheer power available!

I have built a 6-XR-E 'trial' dynamo head light in the past and have been a bit dismayed at the speeds required to get all 6 LEDs to fire up and produce a decent amount of light. Translating this experience to contemplation of separate front and rear triples - red at the rear and white at the front - I'm left wondering if my sedate commuting pace is really going to cut it and provide enough current for 6 leds! Can a dynamo-driven combination of front and rear lights (or even just a rear) provide really bright 'running lights' of the sort of brightness and conspicuity that the creative folks on this list (or on this thread anyway) aspire to?

Maybe my choice of XR-Es for my trial 2 X 3 was not a good one? Are XP-Es (for instance) more efficient? If I'd used XP-E leds in triple configurations would I have got decent light at lower speeds in my trial headlight? Incidentally my average 'best pace' riding to work is around 22kph. Not fast I know but it gets me there. I use modern hub dynamos. I also run an e-delux on my commuter bike so in reality I'm probably just aspiring to build the 'world's brightest rear dynamo light!'

Any input welcomed,

Sam.

[Savvas]

Thanks !-Watt.
You said: "The heat is managed by derving the LEDS at 50% of max forward (pulsed) current with a 10% duty cycle of a 2Hz square wave. The final result is a rapid strobe."
Can you tell me what driver you used please? Or maybe where I could get one with a similar flashing duty cycle?

Thanks,

Sam.

[BrianMc]

Thanks !-Watt. .. driving the LEDS at 50% of max forward (pulsed) current with a 10% duty cycle of a 2Hz square wave. ....Can you tell me what driver you used please? Or maybe where I could get one with a similar flashing duty cycle? Thanks, Sam.

+1 1What.

Sam: You may need Steve to chime in here as he sequentially lights up his multiple LEDs, or you can search for his threads. I use Martin's circuit for a three XP-G light which performs well. I haven't done my beam patter=n evaluation yet on it. A twin setup would let me splay them out for daytime riding, but with a helmet light on battery it seems to catch driver's attention and is easy to fire well at a moderate pace (it is on my errand/semi-cargo bike).

The output per watt input (efficacy) has increased from XR-E to XP-E to XP-G and now XM-L.
The Vf has fallen ands has the thermal resistance from the jucntion to the back of the LED allowing higher currents and more output.
Unfortunately you also get larger dice requiring bigger optics to tame into a road width beam.

The XR-E is about 3.7 Vf at 1 A, 3.5 at 500 mA, and about 2.9 at 0.1 A. So 6 in series are almost 21 volts at 500 mA, 11 Watts. and about 700 lumens OTF assuming you can pedal fast enough to generate 11 watts.

Going to the other currently available extreme, four XM-L's at 500 mA are 2.85 Vf each, for 11.4 Vf total, a bit under 700 lumens OTF, at 5.5 watts. I have two on battery at 1.4 A using 35 mm optics, so if size and weight are issues, you'd want either one of (or a mix of) the XP-E and XP-Gs. to tame them for the road. There are a decent 20 mm lens and a reflector, which at this lower current might work well reducing size and weight considerably.

I figure with three XP-G's and Martin's doubling circuit, I can fire up a tail light too, but I will need at least two and likely three smaller capacitors to replace the single larger one that barely fits in my headlamp housing plus the wiring to the rear light. This is likely a winter job coming up. The lights won't be flashing. I have flashing front and rear helmet lights, so that is fine.





I am also thinking of adding a

[1 what]

Hi Guys,
Savvas - I used to be a dyno man but slowly fell for 18650 battery packs because you could simply do soooo much more with them.
Also they did not clog up with mud in fowl weather ( I ride alot off track).

As for that driver....well it's show and tell time...It was an old 555 timer module I'd made for something else years earlier and I sat down one day and "tweaked" the components to give me what I wanted ( I have a small lab at home). I dont have a formal circuit but there are hundreds of 555 timer circuits on the web. I added an output transistor to drive the leds.

It would not win a beauty contest!
These days I use DX drivers for my taillights.

[panicmechanic]

DX are offering a "LightLane" clone.

[Savvas]

"Sam: You may need Steve to chime in here as he sequentially lights up his multiple LEDs, or you can search for his threads."

Yep - Steve K has been a considerable inspiration! His K.I.Simple approach is, I think, worth remembering! I'm slowly getting an arrangement like his together on my main commuter bike. The problem (for me) is that I keep getting rather inspired by the creativity of others as well!

Thanks very much for the succinct summary of the progression of efficiencies and lower Vf across the Cree range. The lowering of the Vf in particular is probably the single biggest bit of 'progress' I guess for those interested in running multiple LED dynamo lights.

"I figure with three XP-G's and Martin's doubling circuit, I can fire up a tail light too..."
I've built a few of Martin's simpler circuits with satisfactory results. You've reminded me that I actually have one of Martin's circuit boards that will drive (I think) up to 6 leds. I had put it away until I felt more confident in choosing the right led and optical configurations. I think that circuit (#10?) has the doubler built in doesn't it?

I've actually just discovered the Nischia Regal led which seems to be available in red from LEDSupply (though they seem to have some problem posting to Australia), Anyway, it seems to have a v/low Vf of 2.? volts. Three would seem to be a great combination for a rear dynamo 'running' tail light. Must investigate more.

Sam.

[Savvas]

Thanks 1 what (got the spelling right this time!) I've ordered some 55mm and 40mm acrylic tubing. I'm going to have a go at building a 'glow tube' like yours. I've got some alu flat in the shed so I can see a filing session coming up (I don't have a 50mm hole saw unfortunately!) I'm very taken with the large red glowing surface of your tail light, the fact that it has the optics for a bit of longer distance visibility and of course the prospect of some side vision as well. We'll see what can be achieved!

Sam.