Dynamo tail light input needed

[mschol17]

I recently purchased a very cool chainstay-mounted battery powered taillight from Jitensha, and I'd like to modify it to be dynamo-powered. The LED in it isn't very powerful at all, so I've ordered a 1W Red Luxeon LED to replace it. I think it will fit inside the housing-- there is only room for 1 LED so I can't use multiple LEDs.
I'm going to run the light in series with a Schmidt eDeluxe headlight, so the main priority is ensuring I don't damage the headlight. I was planning on a rectifier bridge with a current-limiting resistor (the LED takes 350 mA max), and maybe a standlight using a super cap.
Does anyone have any circuit diagrams that would suit my needs? I've searched the forum and come up empty.

Thanks.

 

[Steve K]

This should be simple. Get a basic 1 amp diode, like a 1N4002, and wire it reverse polarity with the LED. i.e. anode to cathode, and cathode to anode. This will allow the LED to conduct for half of the AC wave, and the 1N4002 will conduct for the other half. The LED will effectively see only 250mA, and everything will be fine.

I use a similar arrangement for my taillights, but I use 10 of the small 5mm red LEDs wired in parallel.

Steve K.

p.s. the diode only needs to be rated for 1 amp and a few volts. Anything rated for 1 amp of current should work fine. I only mention the 1N4002 because it is part of the 1N4001 family, which may be the most common diode in the world (tied closely with the 1N914). You can find it at your nearby Radio Shack, I believe.

 

[mschol17]

Sweet, thanks. I was just wondering if the half-cycle-on half-cycle-off would reduce the lifespan of the headlight in any way.

I suppose I need to get a little more sophisticated for a standlight though, right?

 

[Steve K]

The taillight won't affect the headlight. Current still flows both ways to the headlight.

A standlight gets a bit more complicated. I use a taillight like this, and use the voltage drop across the leds to charge a nicad AA cell. When the dynamo stops, a Zetex boost circuit is used to drive four 5mm leds from the nicad cell. It's only modestly complicated.

Steve

 

[hamilgs]

mschol17: Why not use the "tail light contacts on the eDelux? The current available there is already converted to DC. While I do not have an eDelux (yet!) I have an IQ Fly, with a wired dyno tail light 4D Toplight), and I like the combo quite well.--george

 

[mschol17]

mschol17: Why not use the "tail light contacts on the eDelux? The current available there is already converted to DC. While I do not have an eDelux (yet!) I have an IQ Fly, with a wired dyno tail light 4D Toplight), and I like the combo quite well.--george

Is this true? Do the standard wired dyno taillights not have rectifiers in them? I was planning on running a Supernova taillight with a eDelux, but not if it would be easy to build the equivalent of the Supernova.

 

[RadfahrerWannabe]

Is this true? Do the standard wired dyno taillights not have rectifiers in them? I was planning on running a Supernova taillight with a eDelux, but not if it would be easy to build the equivalent of the Supernova.

Hello,

I would like to know how this turned out...

:thinking:

I also would like to use a very nice old housing updating it with LEDs and connect it directly to the tail light contacts on the eDelux, so...

* I guess it is already rectified, right???
* any Idea how many mAmps & volts are delivered???
* may I use 5mm leds in parallel (or should I use them in series)??
* how many may I use???
* will the edelux stand light function also feed the rear light or a capacitor would be needed??

 

[Savvas]

The tail light output from the e-delux is AC!!! An LED tail light connected to it will need either a rectifier or need a second diode (or LED) connected in anti-parallel.

Savvas.

 

[jdp298]

LEDs run off current running in only 1 direction, so they can run off the half-wave of AC. Steve K's right that you probably want to protect it from a high reverse voltage, hence the anti-parallel diode.

But you can cram in a bridge rectifier, zener, Low Dropout, couple of supercaps into a small space. My circuit and build pictures are here. http://www.flickr.com/photos/jdp298/sets/72157626884852454/

I've also found that any 5mm superbright LED is pretty good behind the optics of a donor rear light. In fact the current setup outclasses the B&M rear standlight for length of time on and outdoes the B&M halogen front light for road illumination. Don't have the optics yet for the narrow spot from my own build but quite happy without it.

I think the electronics-elite thought-police are quietly ignoring me but this circuit really works, been commuting with it for several months now.

 

[BrianMc]

Pretty slick how it fits in that tail light. Hmmmm.

 

[Savvas]

Hi jdp298,

Can you explain please the purpose of the zener and the LDO in your circuit? I thought that the dynamo would simply act as a current source and that voltage would rise only to the level required by the LED. If this is the case, why the need to regulate the voltage? I have built a simple rectified circuit now quite a few times that looks just like yours but minus both Zener and LDO. Seems to work fine.

Sam.

 

[Steve K]

...

But you can cram in a bridge rectifier, zener, Low Dropout, couple of supercaps into a small space. My circuit and build pictures are here. http://www.flickr.com/photos/jdp298/sets/72157626884852454/

at the risk of hijacking the thread... the schematic shows a current regulator feeding a 1 watt white LED. Can I ask why this was chosen instead of just using a 3W white LED that could just be fed directly by the dynamo without a regulator? Or two white LEDs, and use the supercaps across just one of the LEDs? (or did limited space preclude the use of two LEDs?)

If it was a case of "these were the parts I had in my box", I can totally respect that!

regards,
Steve K.

 

[jdp298]

The zener is a bit of paranoia, I don't have one at the back of the bike. It's only there to protect the supercap which is rated to 5.5V because I reckon at the front of the bike the ohmic resistance from the wire is that much less. I don't know what the peak voltage you can get from a dynamo is, but it might get above 5.5V when you go fast enough.

The LDO is to squeeze the supercaps and get longer life when the bike is stopped. Energy is related to voltage squared for caps, so the higher voltage I can get on the supercap, the more energy squared I can store. Thing is, 5.5V is too much for an LED and if I just put a resistor in the way the run-time isn't that good, about 2 mins until too dim to be seen by. The LDO acts as a buck/boost regulator in this instance, always trying to give me 3.3V from whatever I have available. This extends the run-time out closer to 10 mins. Only need 1 at the back, LED current inside specs. Need 3 in parallel at the front to cover the LED fwd current and still use the LDO with the lowest voltage and current drop-out.

I didn't use 2 white LEDs as I wanted to keep the reflector of the donor light, putting the LED where the bulb used to go. I didn't go for a 3W LED as the 1W was cheaper and is enough to see with just fine. The backup plan was to go to 3W if 1 wasn't enough. But obviously with 3W, run-time on caps is that much lower. That was the real engineering trade-off and I think I'm happy with the balance. The 10-degree lens inside the reflector of the donor light and on top of the 1W LED makes the difference for the front. For the back I just let the moulded plastic cover do its thing like it would do if there was a bulb in there.