DX-dynamo light setup

[drmaxx]

I am thinking of using this DX outdoor LED light for a dynamo operated bike light:
http://www.dealextreme.com/details.dx/sku.12958

My understanding is, that I simply need a rectifyer (AC -> DC) between my hub dynamo (Shimano NX71 or similar) and this specific light and I should have an nice water proof bike light.

I am planning to use circuit 7 from:
http://www.pilom.com/BicycleElectronics/DynamoCircuits.htm

I am really new at this - so any thoughts from you guys are highly appreciated? Would this work? Especially considering that I regularly ride faster then 40 km/h.

 

[rayman]

If I were you I would take the DX light. I would use a Tri-Q5 or a single P7/MC-E light.

rayman

 

[Resqueline]

No problems at high speeds. Dynamos are inherently current limiting (450mA is common) above a certain speed like the charts indicate, perfectly suited to drive LED's directly (via a rectifier of course).

Nice diagrams, should work great. Neat trick with that tuning/boost capacitor in series with the dynamo..

 

[Eamon]

I looked around for those emitters, and I could only find them in triple mountings. I wonder if the light is an adaptation of those triple-emitter incandescent replacements that are so common. If so, you might not have as much space for electronics as you thought, and upgrading could be problematic.

Eamon

 

[znomit]

I dunno about this. You aren't driving the LEDs off the dynamo, you're driving the LED driver. You might end up with a higher voltage then the driver can handle.
They are 1W LEDs too, so can only handle 300mA.
You'll need to hack the driver out and replace with a voltage doubler (250mA).

 

[Resqueline]

Yes, using a dynamo requires the LED's to be driven directly (a series resistor is ok) - but not via an electronic driver (it would most likely be damaged).
I didn't look at the DX page to begin with so I didn't see it's only rated at 320mA. There's insufficient info, but I guess it simply has a series resistor in it. If you use that light alone (with circuit 7) it will be overdriven anyway, and the life of it can only be guessed at. Two in parallell should be ok however, or the use of the voltage doubler (circuit 8) instead - as it halves the output current of the dynamo.
The electronics can be placed anywhere.

 

[drmaxx]

Thanks for all the replys and advice. I was hoping that I can use the waterproof light setup and just tinker with the electricity supply. But it seems that there is more to do and for all the work the output is only somewhat mediocre. Sounds like I really should go with a complete self build version.
You gotta love this knowledgable forum!

 

[nintendo64]

anyone know where to buy a RECTIFIER (ready made) so that I can use LEDs from DX?

 

[FrontRanger]

No problems at high speeds. Dynamos are inherently current limiting (450mA is common) above a certain speed like the charts indicate, perfectly suited to drive LED's directly (via a rectifier of course).

It's not explicitly stated on the DX page linked by the OP, but it is quite likely that the light contains a current-mode driver designed to operate from a low-impedance supply. For that case, a dynamo is not an appropriate power source, as it has a high output impedance.

 

[FrontRanger]

Yes, using a dynamo requires the LED's to be driven directly (a series resistor is ok)

This is only true for sufficiently small resistors. (The idea behind driving an LED with a voltage source and an explicit current-limiting resistor is that the explicit resistance is quite a bit larger than the incremental resistance of the LEDs themselves.) Except for sensing the LED current in a feedback system, as in Martin's circuit 13, I can't think of any reason you'd want a series resistance in dynamo-driven LED string.

 

[FrontRanger]

anyone know where to buy a RECTIFIER (ready made) so that I can use LEDs from DX?

Bridge rectifiers made from standard silicon diodes are readily available, like this one at Radio Shack. Schottky barrier diodes have lower voltage drop, so most DIY-ers use those for higher efficiency. Depending on the voltage rating you need, you can use 1N5817, 1N5818, or 1N5819. The voltage rating trades off with Vf, and therefore with efficiency. You can get them in standard through-hole type or surface-mount.

I believe I have seen some fully-integrated rectifiers but I can't remember where. However, the four individual diodes are so small and cheap that getting an integrated rectifier probably won't buy you much, unless space is really at a premium.

 

[yellow]

MOSFET

* anything between dyno and led decreases effciency (= no/less light at low speeds)
* normal rectifyer and diodes also waste power (also counts at low speeds)
* ideal number of led seems to be three
* Juergen H's MOSFET rectifier - for me - worked best of all circuits I have built so far. Useable light at walking speed.
Voltage doubler, switching circuits, ..., were large and more difficult to make (and did not give light at walking speed)

con:
- purely homemade
- very small parts (but doable, I used a dremel)

 

[Resqueline]

It seems to me that the OP understood most of what was stated here, is on the right track, and is happy about it. I'm sure he'll be back with more question should he be unsure about anything.

It's not explicitly stated on the DX page linked by the OP, but it is quite likely that the light contains a current-mode driver designed to operate from a low-impedance supply. For that case, a dynamo is not an appropriate power source, as it has a high output impedance.

Being that the DX light has 3 LED's, is cheap(?), and is specified for 12V exactly, it's very likely using only a series resistor. An electronic driver would have a much larger operating voltage span IMHO. I believe my statement did contain a caution and would inspire further investigation before applying dynamo power to a LED light.

This is only true for sufficiently small resistors. (The idea behind driving an LED with a voltage source and an explicit current-limiting resistor is that the explicit resistance is quite a bit larger than the incremental resistance of the LEDs themselves.) Except for sensing the LED current in a feedback system, as in Martin's circuit 13, I can't think of any reason you'd want a series resistance in dynamo-driven LED string.

Series resistors are usually quite small, dropping a couple of volts in this case.. Notice I didn't state that you'd want it, merely that it would be acceptable to leave it in place (at the cost of a slightly higher speed needed for max output, of course). Overdriving a resistor containing light have an increasingly detrimental effect, of course.

Please also notice that I did caution against leaving any electronic driver circuit in place.

yellow:
Very useful no-loss rectifier circuit you posted there. It's ideal for driving a LED directly, but I question myself if it would perform well together with a filter capacitor. Any experience?

 

[nintendo64]

Bridge rectifiers made from standard silicon diodes are readily available, like this one at Radio Shack. Schottky barrier diodes have lower voltage drop, so most DIY-ers use those for higher efficiency. Depending on the voltage rating you need, you can use 1N5817, 1N5818, or 1N5819. The voltage rating trades off with Vf, and therefore with efficiency. You can get them in standard through-hole type or surface-mount.

I believe I have seen some fully-integrated rectifiers but I can't remember where. However, the four individual diodes are so small and cheap that getting an integrated rectifier probably won't buy you much, unless space is really at a premium.

I managed to get 4x 1N4007, 4x 1N5819 (didnt have 1N5818) and a 2200uF 16v capacitor at the local electronics shop. I want to make a circuit 1 and circuit 2 and buy a R2 or Q5 drop in from DX.

 

[yellow]

with a filter capacitor. Any experience?

unfortunately the mosfet needs the opposing half curve of the AC to work, so bridgeing outputs does not work.
At least one led has to be used without resistor.

If doing so, using capacitors, Gold caps, ..., with the remaining led works
(to get a crude, short timed "stand-light". Technically not "nice", but rugged, easy, small and cheap)


circuit

good working light (all the 2nd driver shown here is skipped now, just the small Mosfet driver behind the aluminium plate and the Gold Cap remain)

also a way to use old Lux V

 

[Steve K]

that's a great use for a Lumotec! I've got a couple that might be used for this.

My only concern is that there's no heatsink outside of the package. Seems like the LEDs would got very hot with this design.

Steve K.