Single LED Dyno Light w/ Standlight

I have a 1972 Schwinn Twin that I've been fixing up to make the ultimate around town bike. I've updated the equipment where it makes the bike more enjoyable to use, but have tried to keep most of the original look.

To that end I've installed the Sturmey Archer XL-FDD dyno/drum break hub in the front wheel.

I've acquired an vintage bullet style light to run off the dyno hub. The optics of the bullet light where designed for an incandescent bulb so will be sub optimal for an LED emitter - furthermore they will only accommodate a single light source - so I will only use one LED as opposed to most other designs that utilize multiple LED's. The usage of a single LED simplifies the needs of the driver circuit since it will no longer require a voltage doubler which is usually necessary to get good low speed performance.

I have (very) little electronic experience but have reviewed a number of the circuits for powering LED lights off dynamo's that have been posted here. In particular Martin's (no. 7) and Steven K's designs have been of interest.

http://www.flickr.com/photos/kurtsj00/5420960597/in/set-72157621965148305

and

http://pilom.com/BicycleElectronics/DynamoCircuits.htm

While circuit 7 meets my needs for driving a single LED efficiently it has no stand light feature. I've thought about increasing the value of C1 with a super capacitor (maybe 1F?) to provide the circuit with the ability to store energy, but I'm unable to conclude if that would indeed function as a stand light or if it would just discharge so rapidly as to be essentially useless. Additionally I am unable to break apart Steven K's circuit design to utilize just the NiCad charging/discharging portion.

If I could get some help with a circuit design that would meet my needs (Good output with a single LED (Cree XM-L?) with a stand light.) It would be greatly appreciated!

-Will

Will,

If you want it as simple as possible, just get yourself a 1F button supercap (1f/5.5 v) and bung it across the led after the rectifier - observing appropriate polarity of course. This should work fairly well. Further standlight efficiency - while worthwhile and (sort of) fun - does involve further complexity and is eventually and inevitably an exercise in diminishing returns. Given what you are after, I'd just keep it very simple and put your efforts into getting the housing and reflector set up right. Given you are using an original 'bullet style' light unit - probably with plenty of room inside - I would consider installing an LED-specific reflector or optic behind the front 'glass' and infront of the original reflector if you can. I can think of a couple of ways it could be quite easily done without too much 'destruction'. You don't have to fret too much about a heat sink using a dynamo circuit - maybe just a light alloy bracket or even a large aluminium bolt through the place the old bulb would have sat! Some large 50mm reflectors (from DX) are metal and will act as a heatsink anyway!

My 2 cents worth anyway.

Savvas

Will, old chap. Just the ticket those bullet lights.







Maplin do a series of lenses for these high power LEDs. I've found the 8 or 10 degree ones the best as it really makes the light glint from a distance.

With only space for 1 housing, I'd be tempted to run a 3W white LED.

PM me and I'll send over the circuit diagram. It's dead simple and lasts for ages, just not at full brightness.

Thank both of you for your help.

A couple of questions:

1.) With a 1F capacitor in parallel to the LED does that affect the tuning capacitors used in circuit 7? (pilom.com)

2.) When referring to a single 3W Led do you mean some thing other than the Cree XM-L led I referenced? Any thoughts about selecting between the cool/neutral/warm colors available?

3.) How did you affix the LED/lense in relationship to the reflector? I just got my light on ebay (http://www.ebay.com/itm/11081367024...AX:IT&_trksid=p3984.m1423.l2649#ht_500wt_1415), so I haven't been able to look at how I might attach it. As soon as I figured out what parts I need I'll order them so that way I can get it together ASAP - I'm very excited about this project.

jdp298, I'll PM you for the circuit diagram as soon as my 3 post requirement has been fulfilled.

-Will

spinningcog said:

.......I have (very) little electronic experience but have reviewed a number of the circuits for powering LED lights off dynamo's that have been posted here. In particular Martin's (no. 7) and Steven K's designs have been of interest.

http://www.flickr.com/photos/kurtsj00/5420960597/in/set-72157621965148305

and

http://pilom.com/BicycleElectronics/DynamoCircuits.htm

While circuit 7 meets my needs for driving a single LED efficiently it has no stand light feature. I've thought about increasing the value of C1 with a super capacitor (maybe 1F?) to provide the circuit with the ability to store energy, but I'm unable to conclude if that would indeed function as a stand light or if it would just discharge so rapidly as to be essentially useless. Additionally I am unable to break apart Steven K's circuit design to utilize just the NiCad charging/discharging portion.

..<snip>...
-Will

Click to expand...

It would be worthwhile to try a basic supercap, in the interest of simplicity.

If you felt compelled to do more, you might be able to adapt the standlight circuit that I usually use for my taillights. Well, it might be an adaptation of my headlight circuit too... Basically, the headlight circuit would be fine, except for the problem of trying to feed power back to the 3W led itself. The simple solution is to feed the power to a smaller led (or string of leds). This is what I do for my taillight:

http://www.flickr.com/photos/kurtsj00/4803579536/in/set-72157617009273346


regards,

Steve K.

spinningcog said:

3.) How did you affix the LED/lense in relationship to the reflector? I just got my light on ebay (http://www.ebay.com/itm/11081367024...AX:IT&_trksid=p3984.m1423.l2649#ht_500wt_1415), so I haven't been able to look at how I might attach it. As soon as I figured out what parts I need I'll order them so that way I can get it together ASAP - I'm very excited about this project.



-Will

Click to expand...

That's a beautiful looking lamp. Nice find!

I've devoted an unhealthy amount of time to figuring out how to retrofit LEDs into vintage lamps. I've just written up a post about putting a Cree XM-L into a Sturmey Archer bullet lamp: http://minisystem.blogspot.com/2012/01/sturmey-archer-headlight-cree-xm-l.html

I'ts a rather elaborate solution, but does result in the LED having adequate heat sinking and being securely mounted in the reflector. I will say, though, that I was disappointed with the reflector's performance. Putting an LED into a reflector designed to focus the light of an incandescent bulb resulted in a very big floody beam with not much of a hot spot. Great for being seen in the city, but maybe not so great for riding around country roads in the pitch black!

minisystem said:

That's a beautiful looking lamp. Nice find!

I've devoted an unhealthy amount of time to figuring out how to retrofit LEDs into vintage lamps. I've just written up a post about putting a Cree XM-L into a Sturmey Archer bullet lamp: http://minisystem.blogspot.com/2012/01/sturmey-archer-headlight-cree-xm-l.html

I'ts a rather elaborate solution, but does result in the LED having adequate heat sinking and being securely mounted in the reflector. I will say, though, that I was disappointed with the reflector's performance. Putting an LED into a reflector designed to focus the light of an incandescent bulb resulted in a very big floody beam with not much of a hot spot. Great for being seen in the city, but maybe not so great for riding around country roads in the pitch black!

Click to expand...

Many years ago, I had a modest amount of sucess putting a Luxeon V side-emitting LED into the reflector of one of these big vintage headlights. Finding a side-emitting LED is pretty tough nowadays. Have you considered just glueing on a small molded plastic optic to your XM-L? It might still fit under the headlight's original lens.

Steve K.

Steve K said:

Many years ago, I had a modest amount of sucess putting a Luxeon V side-emitting LED into the reflector of one of these big vintage headlights. Finding a side-emitting LED is pretty tough nowadays. Have you considered just glueing on a small molded plastic optic to your XM-L? It might still fit under the headlight's original lens.

Steve K.

Click to expand...

The opening is only about 3/4 of an inch, or 19 mm. I haven't found any practical way of fitting a reflector in there without removing or modifying the existing reflector. Changing the position of the LED within the reflector doesn't seem to help much. The light is still plentiful, just not well focused. For my purposes, I'm not bothered by it.

minisystem said:

The opening is only about 3/4 of an inch, or 19 mm. I haven't found any practical way of fitting a reflector in there without removing or modifying the existing reflector. Changing the position of the LED within the reflector doesn't seem to help much. The light is still plentiful, just not well focused. For my purposes, I'm not bothered by it.

Click to expand...

I wasn't thinking of a reflector... just a molded optic. The Ledil Iris looks like it has a good, narrow beam pattern suitable for a bike light, but it's 28mm tall and 37mm diameter. Maybe a little large for your housing? I've seen much smaller optics mentioned on CPF, but ... maybe the Carclo 20mm optics would be good? The max diameter is 20mm and is only 10mm tall. Seems like it could fit almost anywhere.

Steve K.

Steve K said:

I wasn't thinking of a reflector... just a molded optic. The Ledil Iris looks like it has a good, narrow beam pattern suitable for a bike light, but it's 28mm tall and 37mm diameter. Maybe a little large for your housing? I've seen much smaller optics mentioned on CPF, but ... maybe the Carclo 20mm optics would be good? The max diameter is 20mm and is only 10mm tall. Seems like it could fit almost anywhere.

Steve K.

Click to expand...

hmmm. good point. I'd forgotten about those 20mm carlco optics!

These are what I use, 8 or 10 degree makes it so much better than without, especially in those flat reflectors.

http://www.maplin.co.uk/high-power-led-lens-511377

Ledil has some 16.1mm diameter optics that might also work. Maybe this one. Just when I'd given up hope, CPF comes to the rescue!

Minsystems, your website is great! Thanks so much for taking the time to document your work for others to follow. Steven K. thanks for your input!


Circuit design:
Steven K. I reviewed the circuit diagram you posted (as best as I can) a couple of questions
1.) It doesn't appear that you rectify the power from the dyno before it gets put into the battery. Certainly there is a diode there to prevent the current from flowing the wrong way, but current will only be flowing to the battery half the time, the other half of the time current will be flowing out of it? Are we using both the positive and negative swing of the dyno's AC in this circuit? How does this impact the life of the cell? In another thread you discuss the poor performance of a nicad in cold weather, how does this design fair? Would a pair of nicads in series help in this situation (Assuming there is room in the case)


2.) I tried to read the spec sheet for the ZXSC310 LED driver, but I'm still a little unsure about it's function - it drives a transistor to regulate the voltage to produce the maximum output on the LED? Could you clarify?


3.) This circuit design doesn't seem to have a boost converter, could I add a tuning capacitor from pilom.net's design to improve low speed performance? minisystem, how is your microcontroller boost converter coming along? I have an avr programmer though I haven't put it to much use as of yet I'd be interested in giving it a try - the idea of being able to adjust the load on dynamo to match whatever desired level of light output is pretty cool.


Optics:


minsystem, the lens you posted has a 37° viewing angle isn't this pretty high? jdp298 suggested an 8 or 10 degree lens.
jdp298 the lens you posted doesn't seem to have any kind of spec sheet, and I've read it's important to match the lens with the LED, how does it perform? Is it designed for CREE XM series leds?
What about this lens? http://www.mouser.com/ProductDetail/Ledil/C10684_Eva-XM-D/?qs=QsMJd9DV0TW5PckpKZBcYg==it has a 16 degree output. it's 35mm in width, perhaps it could be fitted farther in past the reflector.


Housing:
Minisystem, how did you get your copper sinks fabricated?


-Will

My dynamo buck driver circuit is on hiatus at the moment. It seems overly complicated just to squeeze a bit more power out of the hub, and only seems to offer benefit at higher speeds. swhs has been documenting what sounds like a very impressive MCU-based LED driver that can pull a surprising amount of power out of a hub dynamo at low speeds. Hopefully someday the details will be made public.

I haven't tried the reflector I posted. 37° is the narrowest viewing angle available in that format. I expect it will be fine, but I'm not nearly as picky about beam shape, throw, etc. as others are around these parts. It will almost certainly improve upon the floodlight pattern that I'm getting now.

I get my copper heat sinks fabricated by a prototype maker in Shenzhen. $20-$30 a piece for the size and complexity that I've been getting made. Turn around is fast and quality has been good.

spinningcog said:

Circuit design:
Steven K. I reviewed the circuit diagram you posted (as best as I can) a couple of questions
1.) It doesn't appear that you rectify the power from the dyno before it gets put into the battery. Certainly there is a diode there to prevent the current from flowing the wrong way, but current will only be flowing to the battery half the time, the other half of the time current will be flowing out of it? Are we using both the positive and negative swing of the dyno's AC in this circuit? How does this impact the life of the cell? In another thread you discuss the poor performance of a nicad in cold weather, how does this design fair? Would a pair of nicads in series help in this situation (Assuming there is room in the case)

Click to expand...

in my headlight design, I use a bridge rectifier at the input to the circuitry. In my taillight, I have a small array of leds wired in parallel, with another diode (or diodes) wired in the opposite polarity. The voltage across the array then goes through a series diode which does block the negative voltages. As a result, both designs only deliver positive voltage to the nicad. The taillight does have a disadvantage regarding battery charging, as it only has power over half the AC waveform. Hasn't proven to be a problem in my use.

Nicads do lose some capacity in the cold, and the internal resistance increases. Basically, the chemical reactions involved in both charging and discharging are slowed down. All battery chemistries have this problem, but to different degrees. From what I've read, Nicad does better than Nimh, and is quite a bit more tolerant of sloppy charge regulation than lithium.

spinningcog said:

2.) I tried to read the spec sheet for the ZXSC310 LED driver, but I'm still a little unsure about it's function - it drives a transistor to regulate the voltage to produce the maximum output on the LED? Could you clarify?

Click to expand...

The ZXSC310 is a boost converter, and provides a bit of regulation. Like any boost converter, it uses a transistor to connect the low side of the inductor to ground, which causes current to flow through the inductor and store energy in the inductor's magnetic field. When the current reaches the desired level, the ZXSC310 turns off the transistor, which allows the stored energy to now be directed into the load (i.e. LEDs). After some time, the ZXSC310 turns the transistor back on and the cycle starts over.
Honestly, it's not well regulated at all. I just experimented a bit with the number of LEDs on the output and the current sense resistor to see how bright I could get the LEDs.

spinningcog said:

3.) This circuit design doesn't seem to have a boost converter, could I add a tuning capacitor from pilom.net's design to improve low speed performance? minisystem, how is your microcontroller boost converter coming along? I have an avr programmer though I haven't put it to much use as of yet I'd be interested in giving it a try - the idea of being able to adjust the load on dynamo to match whatever desired level of light output is pretty cool.

Click to expand...

With a single LED in the headlight, the dynamo should be able to provide adequate light at low speeds... or are you using a Sturmey-Archer Dynohub? Let me re-phrase: a modern hub dynamo should be able to drive a single 3W LED just fine at slow speeds. I use two Cree XR-E's in series, with an array of parallel 5mm red LEDS in series, driven by a Schmidt dynamo. It works down to 3.5mph, although it is a bit flashy, and is adequate for me to slowly grind up the local river bluffs (with lots of grunting and groaning involved).

regards,
Steve K.

Quick hijack: SteveK, as usual your advice is spot on. The Ledil optic improves the beam significantly.

minisystem said:

Quick hijack: SteveK, as usual your advice is spot on. The Ledil optic improves the beam significantly.

Click to expand...

Glad that I could help! Appropriate optics are the key to good bike lights, and bikes usually require tighter beams than the average flashlight, so finding a good optic can be tough. Do you have pics of the light or beam?

edit: I followed the link to your blog and saw the results. The 37 degree optic is quite wide for a bike light. You'll get much better results once you can get a narrow optic.

Steve K.

Steve K said:

Glad that I could help! Appropriate optics are the key to good bike lights, and bikes usually require tighter beams than the average flashlight, so finding a good optic can be tough. Do you have pics of the light or beam?

edit: I followed the link to your blog and saw the results. The 37 degree optic is quite wide for a bike light. You'll get much better results once you can get a narrow optic.

Steve K.

Click to expand...

Yes it is very wide, but it is big improvement on just the reflector. Apparently I attracted the attention of someone from Ledil who says they'll work on getting a narrower optic available in North America. The choice for the XM-L was pretty limited in anything that was less than 20 mm. The choices for the XP-G are better. So, I guess I should've made these heat sinks for XP-Gs instead of XM-L's. Oh hindsight...

Spinningcog,

If you want to experiment with your headlight as a host for leds, I can recommend these TIR lenses from DX:
http://www.dealextreme.com/p/23-36mm-optics-light-diffusers-for-cree-emitters-glass-5-pack-1920
or
http:// http://www.dealextreme.com/p/35-6mm-secondary-optics-plastic-5-pack-4544

Although made from the XRE rather than he XPG or XML I have found them to be pretty good. Their advantage for your purpose may be that they are a 'jam fit' (or the first ones are anyway) onto the metal ring around the XRE's die. And they are cheap enough to prove your concept before you invest in brighter/more expensive stuff. Personally I find the combination of XRE and a dynamo source perfectly adequate.

Savvas.

Hi to everybody, this is my first post here.

I have a traditional dynamo light system on my bikes (cheap bottle dynamo and classic front and rear incandescent bulb lights) and I too would like to retrofit a led system with standlight into the old light enclosures, keeping the original look.
I use those bikes basically in town and only rarely on unlit roads at low speed, so I don't need ultra powerful lights. Actually the main reason to switch to a led system would be to have the standlight function at intersections and a more reliable system as the incandescent bulbs don't last very long. I have limited electronic experience and therefore I'm looking for a simple circuit which is easy to build, cheap and small enough to be installed inside the existing lamps, like those described in this thread. I still haven't decided which one to try first, but I'm a bit concerned about the generator:

I will use the existing cheap bottle dynamos installed on the bikes for now, but I plan, if I'm satisfied with the results, to upgrade to better generators, like the Nordlicht or the Axa HR-traction
Those dynamos have a voltage limiter (I think that it consists of two zener diodes connected back-to-back between the "live" and "earth" poles) to prevent bulb failures at high speed. Do you think that those protection diodes may decrease efficiency when switching to led lights, and if so, anyone knows if they can be easily removed from those dynamos?

Alessandro.

P.S. excuse my bad english, I'm from Italy.