Designing good daytime rear commuter lights

[BrianMc]

Funny, it looks like a film cannister with two or three LEDs inside to me. Aluminium lid to shed some heat...

Might give it a try and clone it :devil:

It appears to be longer with the LEDs arranged linearly down its center (pictures from their web site copied to a service to avoid hot linking, but as this is not a commercial use, and having the cycling community familiar with the aftablasta, I think it fair use):






It looks like a linear array of red LEDs down the center line, and a cover a bit longer and wider than a film cannister.

That said, a film cannister bottom fits the bezel of the Bullet Marwi light body perfectly. And it just so happens two of those NOS lights should be delivered today one for spares, the other as a possible host for a driver and red LED board. :naughty:

I could stick with a triple MCPCB and aim the light straight back with the frosted translucent 'lens' like the earlier red flashlight extension difusing the light to the sides. Stick in a bFlex to match the front lights to keep programming the same (easier for me to keep straight) and for reliability, as it is a safety device. That would be straight forward and I would likely go the aluminum slug route as a firewall/thermal path to the aluminum body.

OR I could mimic their design more closely. The LEDs are cheaper on three 10 mm square or round MCPCB's and I could mount them on a spar fastened at one end and side to the body of the light with the three 10 mm PCPCB's side by next on the other side and end of the spar facing what would have been the top of the light. The cannister is 50 mm long, so the first LED could be 15 mm from the edge of the bezel and the last 15 mm from the end of the film cannister.

Full on, I would be moving 3 watts of heat or so down the spar to be radiated by the body. The spar needs to be 10 mm wide for the MCPCBs, and at least 5 mm thick at the LEDs. I think it will be "L" shaped so it connects top and bottom of the light body and side struts at 90 degree to place it laterally in the body, increase its rigidity, and improve its thermal path to the body. Does it need to be thicker than 5 mm ( 50 mm cross section) to be adequate to handle 3 watts?

 

[BrianMc]

Got a sunny day on the weekend so I could check PBSF's with the sun at a higher angle.

First, yes, the low sun did wash them out relative to about 2:00 solar time. Shadows still short. Confirming the 'worst case' previous setup.

It is EXTREMELY critical that the PBSF's aim STRAIGHT back, as I have done, but the helmet shifted on me after I walked away from the test stand on the car trunk and it may as well have been shut off for all the good it did. I thought the pivioting mounts some offered for helmets was excessive. Not any more.

You could see the lights well at 500 feet. They were brighter but much smaller than the cars 4-way flashers. Easily dismissed or overlooked. Having both flashing rather than just one increased the attention getting aspect noticeably.

So can you see them a quarter mile in the early post dawn? If the cyclist was riding in anmy angle except directly away from the sun, most definitely, in fact they shoud show well in the rider's shadow. Riding away from the sun, though they are pathetic. Eevn in the mid afternoon, though they might be seen at more than 500', I doubt that the average bike unaware motorist would notice. Too little. Too far.

I forgot fresh batteries for the Xenon strobe. With the 3-4 hours of use batteries in it, the flash wasn't overly slow yet indicating full drainage, but the distance was pathetic (50' straight back, <20 in side view).

Awaiting the Radbot 1000's for a redo. I'll try again for a light meter, but we may have to settle for car tailight and 'stick on brake pedal' brake lights for comparison.

 

[pe2er]

Might give it a try and clone it :devil:

Not cloning it today, but adding a brake-light to my taillight. So I took the opportunity to make a daylight comparison. VM Taillight, PBSF and a 1 Watt RED LED in a film cannister. Jut for fun and educational value

Straight from the rear of the PBSF.

And from the side

 

[BrianMc]

...I took the opportunity to make a daylight comparison. VM Taillight, PBSF and a 1 Watt RED LED in a film cannister. Jut for fun and educational value ....

Film cannister isn't transluscent enough. If you are messing about again photgraphing the new brake light, could you do some distance shots/ride bys? In another thread it appears, that the 1/2 watt and 1 watt LED's are not being driven anywhere near their max current and nominal wattage in self contained lights (heat & power/runtime issues, likely). Wheras your 1 watt rear light is, right?

 

[pe2er]

Film cannister isn't transluscent enough. If you are messing about again photgraphing the new brake light, could you do some distance shots/ride bys? In another thread it appears, that the 1/2 watt and 1 watt LED's are not being driven anywhere near their max current and nominal wattage in self contained lights (heat & power/runtime issues, likely). Wheras your 1 watt rear light is, right?

Yes, I came to the same conclusion - the canister absorbs way too much light.

The LED I use is a Dx 12V 1W LED Spotlight (Red, SKU 5325). It is a Luxeon-like emitter on star, driven at 1 Watt. I Stumbled upon the 1 Watt/ 1/2 Watt rear blinkies not driving their LED at full power phenomena before. It is simply not practical to do that on two AAA batteries.

I Could make some pictures for you, but at the moment that would not serve a purpose. I Installed the second LED in the same housing, without a lens/ diffuser.

This is the unmodied tail light:

And the modified tail light with the brake-light ON (I mention it, since you could oversee that in the picture ;p):

And this is what's inside. Two drivers and two LEDs with associated wiring:

The light is connected to the battery via a micro switch on the brake lever:

As you can see, the brake light LED light output is disappointing. I am thinking of installing a small 10° lens below the diffuser.

Edit: Here is some video:
http://vimeo.com/11627107 Short close-up of the light
http://www.vimeo.com/11627948 Short video in the street.

 

[BrianMc]

Erwin, I think the lens will do it. It is swamped by the tail light.

I didn't find that on DX before when I searched from your previous mention. I was looking for a star. A 12 volt driver, a case (of sorts), and the LED? Wow! Thanks for the SKU. Since I have a nominal 12 volt system, it looks very good to me. In theory, in your application, and as my solution. :twothumbs

One reviewer said almost 1.8 watts. Did you play with it at all before extracting the stars? If so, what did you think of the beam? Is the star's thermal path OK as is? I can let a couple of SPSF's and Radbot 1000's flash away and leave the bright red 'eye' on steady. Two are less than half the cost of one Radbot 1000. Of course no modes and not self contained with batteries. I guess I'lll see how slow the order is maybe I'll get them before you get your aspheric lens.
:nana:

 

[pe2er]

I Did measure the current when I first used one. Power consumption was less than 1.8Watt, but I do not recall the exact figure.

The Dx spot light has a plastic reflector. Beam pattern is tight like that of a flashlight. The reason for the diffuser (Dx SKU 1918) is spreading out the beam horizontally and improve side visibility, as the reflector does not leak light to the sides like the lens.

The driver is super-glued inside the plastic foot of the lamp. Do not try to pull it out as you will damage it. Instead, gently cut away the plastic using a sharp pair of pliers.

You can use the spotlight as is. The star is attached to the metal cup of the light with some thermal glue. The driver is not cooled, but that should not be required for a 1 Watt LED. Voltage can be anywhere from 6 to 16 Volt DC or even 5 to 13 VAC. You could bypass the bridge rectifier on the driver board and improve efficiency some (with DC only of course ), but I did not do that.

I guess I'lll see how slow the order is maybe I'll get them before you get your aspheric lens.
:nana:

I Installed a small 10° lens over the brake emitter.

The next picture is underexposed -2.1eV so you can see the shape of the light.

Now I am happy

Just shows what difference a clear plastic lens makes - as compared to a translucent cover.

Here is a video similar to the previous one. It is a sunny day, 9:00 AM, and the VM is in the shade. Video is underexposed -1.0 eV.
http://www.vimeo.com/11646045

Similar video of the VM in the street on a sunny day, around 10:00 AM.
http://www.vimeo.com/11647553

 

[BrianMc]

Excellent! You SHOULD be pleased. :twothumbs Nice video. I can clearly see the roundabout maybe because the grass grew? I was looked before like you were ducking into the side street. Keystone Kops-like.

Thanks for the feedback. I'll order two of the red spots today. That gives your order a bit more lead time on mine. I'll have a look at my new brake levers when they arrive to see if I can fit a switch, if not, the front caliper or cable might be a good activation point. I may drag the back brake at times to keep from accellerating down grade, but I don't really use it to stop except with the front.

I'll add a set of the lenses from the front beam thread too, since I am not sure how well the ones I have will work. Not having to change out bezels for day versus night will be nice. :thumbsup:

 

[BrianMc]

"Dx 12V 1W LED Spotlight (Red, SKU 5325). It is a Luxeon-like emitter on star, driven at 1 Watt . Voltage can be anywhere from 6 to 16 Volt DC or even 5 to 13 VAC." (from Pe2er)

DX doesn't list the voltage range.

I assume there were specs with the bulbs? If they can handle 16.7 volts, then this battery may be of interest too. I got one of these ($20 for 4.5 Ahr, clearance from another source a couple of months ago):

http://www.amazon.com/dp/B000HWZK60/?tag=cpf0b6-20

Originally listed for $99. When mine arrived it would not hold a charge after two attempts. So I set it aside to be examined later, the issue with my old bike lights battery was solved.

The package it came in indicated some rough treatment. So opened the case. No obvious damage or leaking. The cells were cool on charging. So I gave it another go before taking the case completely apart to track down the problem. It is up and running. I am guessing that the protection circuit board had become dislodged and reseating it to have thermal contact with the top cell and no contact with the I/O board fixed the problem. It is 16.7 volts fresh off the charger and still is 24 hours later. It is supposedly 45 Ahr.
It is lighter than the once (no longer) similar capacity NiMH. But not as much as I thought it might be.

I ran the helmet light at 500 mA for about about 1.3 hours on a ride and used it another 45 minutes of so at half (5.2 watts) and full power (11 watts) it recharged in a very short time, but I forgot to measure, maybe 30 minutes.

The reviews make me think that many users of laptops have run it down to its limits repeatedly which is not the way to get a long life from a Li-ion battery. Or they have dropped the unit as some experienced the complete discharge over a short time span problem I had at first.

It doesn't appear to be well suited to its intended purpose. For those comfortable with possibly reworking the electonics, repackaging, or just using it as an inexpensive source of 8 Li-ion cells, it might be of interest.

 

[pe2er]

My Li-Ion cell came from the discarded battery-bin in the office

I assume there were specs with the bulbs? If they can handle 16.7 volts, ...

No, there are no specs. In the review on Dx you'll find: "Works from 7-14 volts DC. Reduced output from 7 to 4.5 volts."
I Recall measuring the light output on my own light. It was near max brightness down to 6Volt DC. Otherwise I would have bypassed the diodes in the driver (6V is the minimum voltage for the 2s3p Li-Ion pack in the VM). I left them in place because it is easier when wiring - no need to observe polarity.

They do say it is a 12VAC light. After rectifying, this will give a peak voltage of 12*SQRT(2) = 16.97VDC. So I think it will survive being connected to your battery (leave the diodes in place...no warranty ). If you have an adjustable voltage power supply, try increasing the voltage gently and observe the current draw. Current should decrease if voltage is increased. If current starts to increase with voltage before you reach 17V, don't use it at that voltage.

 

[BrianMc]

My Li-Ion cell came from the discarded battery-bin in the office

Very green.

No, there are no specs. In the review on Dx you'll find: "Works from 7-14 volts DC. Reduced output from 7 to 4.5 volts."

Read that too. Doesn't indicate he tried it above 14 and found it did not work. Your 16 volt figure made me wonder. Now I see it likely comes from the AC sin wave peak.

They do say it is a 12VAC light. After rectifying, this will give a peak voltage of 12*SQRT(2) = 16.97VDC. ... If you have an adjustable voltage power supply, try increasing the voltage gently and observe the current draw. ...

Used to have access to such a power supply.

Haven't ordered from DX yet. Can't login. Maybe a time/processing thing. Will try tomorrow (Monday their time). Live help will be up, anyway.

 

[Offroad'Bent]

OK, ordered the drop-in, host and rechargeable CR123As.
I'll report back how it compares with the Superflash and Mars 4.0.

OK, reporting back. The DX is *way* brighter than the Mars 4.0, which is brighter than the Superflash.

Heavier, needs 4 clicks to get to flashing mode, no side visibility and has no specific bike mount, but it's a lot of rear daylight visibility for the $$.

Now someone throw this P60 emitter in a lighter plastic waterproof case with a bike clip and a translucent cover for some side throw and a big clunky High Flash/Low flash/Off switch, and we're done.
I don't even care if it uses AAA or CR123A.

 

[Offroad'Bent]

OK, reporting back. The DX is *way* brighter than the Mars 4.0, which is brighter than the Superflash.

Heavier, needs 4 clicks to get to flashing mode, no side visibility and has no specific bike mount, but it's a lot of rear daylight visibility for the $$.

Now someone throw this P60 emitter in a lighter plastic waterproof case with a bike clip and a translucent cover for some side throw and a big clunky High Flash/Low flash/Off switch, and we're done.
I don't even care if it uses AAA or CR123A.

Verdict from road test- I mounted the light flat on my helmet easily with some stretch velcro (facing backwards), and we rode in bright sunshine. Definitely good daylight visibility on rapid flash mode from a considerable distance. The weight wasn't too bad on the helmet, but the light may need to run on low power in lower light or night situations to avoid blinding other cyclists.

I just wish it was lighter.

 

[BrianMc]

An update.

The red DX spotlights are ordered. I ordered the same POS light meter as Pe2er is using, while I was at it. When they will get here? Well its DX.

The Radbot 1000's are in a larger order to save me shipping costs, but the rims for the wheels are on backorder. Not sure how long Deep Vees take when they have to come from Australia as the warehouse is out.

So I am making sure my NiMH are fresh, the SPSF's well aimed, and trust mostly to the ANSI vest. Most of my issues with drivers lately is no that they did not see or register my presence but being impatient, passed me on a blind hill within a car lenght or two of the stop sign. No brighter red light will fix that sort of stupid impatience. Oh and they were speeding by 15-20 mph to catch me there, so could have slowed to let me crest the hill so they could pass where they could see. Probably disturbed their phone calls. Sorry, I feel better now.

 

[pe2er]

What is the status? Did the goods from HK arrive?

For those who doubt the need for a good daylight-visible rear light, here is an example of why you need a good tail light on a French forum

 

[panicmechanic]

For those who doubt the need for a good daylight-visible rear light, here is an example of why you need a good tail light on a French forum

I doubt it would have helped - the driver, aged 84, was in a hurry, visibility was good. He would have crashed into anything that was in the way.
The driver might have to part with his license, that may be the only good outcome.

 

[EFMax]

Now I have been messing with getting a decent rear light for ages.. something that is very powerful in daylight and thus really bright at night.
I have tried various bike lights or all sorts of prices and brightness.. and for a while I thought this was bright all be it a bit ugly.. but very bright all the same..

It has 10 leds and cost around £30 GBP..

But then I came upon this LIGHT .. by chance. It only has 4 leds and cost about £10.GBP… Now if you have a decent front light these days, most run off a Li ion 12v – 18v system. Well this light can be fitted to a simple box and bracket and plugged into your existing front light system (mine is 14.8v) and away you go. It will work with any power between 12v-24v with no messing on your side of things… just wire and go.

In daylight this light is good for about 250+mts in straight line. At night it offers some side on vision and is just about impossible to miss even by drivers who are not looking.. and it totally puts the above light in to the shadows as if it was not there.. the same company offer bigger options but this is a bargain and its neat and HIGHLY VISABLE in daylight, especially if its foggy.

 

[BrianMc]

What is the status? Did the goods from HK arrive?

Sorry. Out-of-town. Shut the PC down and left early on the 10th for Canada. Just back. No PC with me.

I am still alive. I think. My allergies have defeated the third med. So I need to fix that tomorrow so I can ride (breathe).

Mail held, so maybe tomorrow there will be a DX delivery in the backlog of light meter and the red spot lights.

---Update they arrived and were waiting. Need 9 volt battery for meter. So 2 x 1 Watt Red spot.

Rims came, wheels built, order boxed, and ships tomorrow AM. So the PDX Radbot 1000's will arrive likely on Friday. Maybe Saturday.

---Update scheduled for Friday delivery.

I will be overhauling the bike, so no riding video of lights in action for a while, assuming I can breathe again.

I found the trunk lid of my car is a good test platform for distance observations and bike seat stay substitute. I'll shoot some video and some stills and get some light readings.

A Bientot

Brian

PS---Update Pe2er confirmed my guess on dissassembly of DX red spot lights. Will see if I can mod to fit Marwi MR11 housing to make a pretty paired tailight unit.

 

[BrianMc]

Two 1 watt MR16 DX red spot lights modded to fit twin Marwi housings:



These bulbs modded:

http://www.dealextreme.com/details.dx/sku.5325

Newer style Marwi in paired set :$20 plus shipping.

http://www.hoffmanamps.com/MyStore/catalog/partsled.htm (Link replaced, thnaks, EL34)

The front bezel/slotted faceplate of the DX LED bulb (friction fit) was removed. The bell shaped aluminum shell of the bulb was cut to just the plate and 1/8" or so and ground/cupped edge to friction fit in Marwi housing. The reflector was reused but I liberated the glass lens from the halogen MR11 bulbs, and made a spacer of 12 gauge copper wire to hold the original reflector forward enough so the bezel would tighten the lens against the reflector. The socket in the light took too much room so each was removed and the leads were soldered to the MR16 pins. The driver is polarity neutral which helps. I removed all switches as this will either be plugged in and on or not plugged in.

Testing for continuity on assembly led to retina after images. These are BRIGHT up close. Shining across a 15 foot room the image was about 6' wide, so a lane wide at 30 feet. Maybe not Dinotte 200, but they will make an interesting comparison to the Radbot 1000's and PBSF's.

 

[pe2er]

Nice build! Looks like that wil be BRIGHT

Looking forward to see action-pictures