Sturmey dynohub + Leds - Am I missing something ?

[FATTYTRIATHLETE]

I have wired up an old Dynohub to 2 Q5s ( I'm told they are Q5s - I nabbed them from Tescos AA torches ) in series - no other components.

At very low speed it flashes but at about 10 mph it flashes so fast it appears as solid light and just as bright as running each LED from the AAs

Is it as simple as this and I dont need Bridge Rectifier and Capacitor or am I missing something else ? Would they blow if I go any faster ?

When stationary I will have a battery operated LED flashing so junctions etc should be visible.

Could I then progress to 2 x Cree MCE in series or a combo of 1 MCE and 1 x Q5 ?

Any advice wecome please.

[Steve K]

The Sturmey Dynohub was rated at about 1.8 watts (I think), versus the typical 3 watts for a modern dynamo. My guess is that you'll be limited to lighting up two Cree XR-E's in series, but you can certainly try using a MC-E. The worst that can happen is you'll notice that you're not getting as much light... or that you spent a lot of money to get a tiny increase in light. :-)

I would suggest installing a bridge rectifier between the dynamo and LEDs. A Schmidt would have exceeded the LED's reverse voltage rating by now and killed them. Your LEDs have been spared because the Sturmey is not as powerful. However, the faster the dynamo turns, the higher the reverse voltage will be. Better to use a bridge rectifier than to risk killing the LEDs. Just use any basic 1 amp rectifier diode with a voltage rating greater than 20v or so. The trusty 1N4001 meets this spec, and costs a few pennies each.

good luck,
Steve K.

[FATTYTRIATHLETE]

Many thanks

- Looks like I've been lucky not to fry them. Will sort the Bridge rectifier as want to move on to a Shimano 80 Hub fairly soon.

[broadgage]

Use of a bridge rectifier would certainly avoid risk of killing the LEDs by reverse voltage, but remember that the bridge rectifier will drop about 1.5 volts, which is a large percentage of the very limited voltage available.
It would probably result in the LEDs not lighting at all at lower speeds.

A better approach might be to use 4 LEDs, and no other components.
Wire the LEDs in series pairs, and then connect the pairs in inverse parralel, then the AC from the generator will light first one pair and then the other.
Except at very low speeds the light will apear continuos, and the maximum reverse voltage to which the leds can be subjected, will be the same as the forward voltage.
The LEDs may be slightly more efficient wired thus, since each pair will only be lit half the time and will therefore run cooler.

[saltytri]

Use of a bridge rectifier would certainly avoid risk of killing the LEDs by reverse voltage, but remember that the bridge rectifier will drop about 1.5 volts, which is a large percentage of the very limited voltage available.
It would probably result in the LEDs not lighting at all at lower speeds.

A better approach might be to use 4 LEDs, and no other components.
Wire the LEDs in series pairs, and then connect the pairs in inverse parralel, then the AC from the generator will light first one pair and then the other.
Except at very low speeds the light will apear continuos, and the maximum reverse voltage to which the leds can be subjected, will be the same as the forward voltage.
The LEDs may be slightly more efficient wired thus, since each pair will only be lit half the time and will therefore run cooler.

Would you care to diagram this? Could you do it with a single MC-E?

[Steve K]

Use of a bridge rectifier would certainly avoid risk of killing the LEDs by reverse voltage, but remember that the bridge rectifier will drop about 1.5 volts, which is a large percentage of the very limited voltage available.
It would probably result in the LEDs not lighting at all at lower speeds.

A better approach might be to use 4 LEDs, and no other components.
Wire the LEDs in series pairs, and then connect the pairs in inverse parralel, then the AC from the generator will light first one pair and then the other.
Except at very low speeds the light will apear continuos, and the maximum reverse voltage to which the leds can be subjected, will be the same as the forward voltage.
The LEDs may be slightly more efficient wired thus, since each pair will only be lit half the time and will therefore run cooler.

The bridge rectifier does add to the voltage drop of the load that the dynamo sees. However, the dynamo does not behave like a simple voltage source. It acts like a speed dependant voltage source in series with a large impedance. For the least sensitivity to operating speed, the load is usually designed to operate near the short-circuit area of the dynamo's voltage-current curve.

i.e. adding the bridge rectifier will reduce the current out of the dynamo slightly, but the dynamo won't have a problem producing the needed voltage. Adding the bridge rectifier is the most cost effective solution and is unlikely to produce a noticeable reduction in light. The MC-E solution will certainly work, but is going to cost more. A bigger factor might be whether the user wants to just install a single LED package and optic.
.... And how easy it is to get optics with the desired characteristics. There's not nearly as many optics available for the MC-E.

Steve K.

[FATTYTRIATHLETE]

[quote=broadgage;2845291]Use of a bridge rectifier would certainly avoid risk of killing the LEDs by reverse voltage, but remember that the bridge rectifier will drop about 1.5 volts, which is a large percentage of the very limited voltage available.
It would probably result in the LEDs not lighting at all at lower speeds.

"A better approach might be to use 4 LEDs, and no other components.
Wire the LEDs in series pairs, and then connect the pairs in inverse parralel, then the AC from the generator will light first one pair and then the other.
Except at very low speeds the light will apear continuos, and the maximum reverse voltage to which the leds can be subjected, will be the same as the forward voltage.
The LEDs may be slightly more efficient wired thus, since each pair will only be lit half the time and will therefore run cooler. "

Hi, could anybody tell me / diagram how above would be done ?

Many thanks.

[elagil]

would be like this:

[Steve K]

a partial quote:

"A better approach might be to use 4 LEDs, and no other components.
Wire the LEDs in series pairs, and then connect the pairs in inverse parralel, then the AC from the generator will light first one pair and then the other.

Hi, could anybody tell me / diagram how above would be done ?

Many thanks.

I think the phrase "series pairs" was intended to mean this:


This should be appropriate for most dynamos. Running two Cree's in series works fine for my SON dynamo, and provides some light even at walking speeds.

I'd still recommend using a bridge rectifier, followed by two Cree's in series. Much cheaper, fewer LEDs and lenses, etc. I'd at least give this a shot before commiting to using four LEDs.

regards,
Steve K.

[FATTYTRIATHLETE]

Many thanks for all the advice.

[Bandgap]

I have run reverse pairs of LEDs and the output is pretty 'flashy' at jogging pace - the bridge way of driving a single pair is not so flashy.

I think they would be equally flashy if the leds were well matched in brightness and beam, but this is seldom the case.

I think LED will rescue many old Dynohubs from obscurity!

Steve

[Nemozoli]

I have run reverse pairs of LEDs and the output is pretty 'flashy' at jogging pace - the bridge way of driving a single pair is not so flashy.

I think they would be equally flashy if the leds were well matched in brightness and beam, but this is seldom the case.

I think LED will rescue many old Dynohubs from obscurity!

Steve

I think that one well-placed (and well-calculated) capacitor should eliminate the flashing...

Nemo

[Resqueline]

A cap would reduce the flashing at low speeds in the bridge rectifier circuit, but there's a risk involved. If the connection between the cap & the LED's were to break - & then reconnect itself, instant destruction of the LED's might result.
I can't think of any way a cap could reduce flashing in the anti-parallell circuit however.

[syc]

Has anyone tried the MC-E in anti-parallel? Does having all 4 dies on the same emitter diminish the flashing?

[Nemozoli]

Resqueline,

the anti-parallell circuit can not be fitted with a cap (or two, for that matter), because the cap's would not be able to discriminate between the two directions of inflowing current :-)
The risk you have described is real, I can imagine the massive buildup of power in the cap destroying the LED... whoooh! /shivers/

syc,

I think that having all the 4 dies in one emitter can not diminish the flashing effect, because it is caused by the rapid reaction of the LED's glowing matter. The emitter doesn't have the wolfram's kind of sluggish response to the abrupt cutoffs and onsteps of power, so it will just flash from one small place instead of two places :-D

Nemo

[unterhausen]

has anyone used the red Cree XR-E devices for a tail light with a dyno? I just ordered some. I'd like to be able to run them because I have forgotten to turn on my blinkies at least once. Might also be useful for brevets.

[Alex Wetmore]

You don't need something as fancy as the CREE for a tail light. You can build a blindingly bright tail light just using standard 50ma LEDs wired in parallel. I've built a few of these, this is from my first one:

http://blogs.phred.org/blogs/alex_we...taillight.aspx
and
http://blogs.phred.org/blogs/alex_we...ht-output.aspx

I use the LED-94 from http://allelectronics.com. 8 of them is enough and they are 54 cents each. This builds a light which is brighter than a Planet Bike Super Flash and close to the DiNotte tail light in apparent brightness.

alex

[znomit]

You don't need something as fancy as the CREE for a tail light.

No but it means just one LED, its incredibly bright, and you can select optics (and they're very cheap from DX, under 4$).
I'd just use a single cree, and not run at the full 500mA (though that would make a good daytime mode).

[sfCyclotourist]

A cap would reduce the flashing at low speeds in the bridge rectifier circuit, but there's a risk involved. If the connection between the cap & the LED's were to break - & then reconnect itself, instant destruction of the LED's might result.
I can't think of any way a cap could reduce flashing in the anti-parallell circuit however.

I've built just such a light recently:

http://www.flickr.com/photos/jimg/se...7614980513926/



Even though I'm using a cap, the light still flickers as can be seen in the video at the above link. In fact, I can't notice a different with or without the cap. Also, at most I only get a couple of seconds of (very dim) standlight function from the cap. Two factors come to mind here:

1. I used a common silicon diode bridge rectifier, not a Schottky bridge, so there are more voltage losses. This may prevent the cap from charging as high as it might at lower dyno-wheel speeds.

2. Since there is only one LED paired with the cap, there is not as much voltage present in the circuit as there might be with multiple LEDs in series. This may also be preventing the cap from charging as much as it might.

Thoughts?

Also: I did fry a LED during prototyping when a cap disconnected and then reconnected to the LED. There is no visible damage to the LED itself, but it measures zero resistance across the pos & neg terminals on my multimeter now!


Thanks!

-Jim G

[pe2er]

This is a taillight I built some time ago. It uses a single Luxeon LED and is powered by three AA type Alkaline batteries:

(Click on picture for my page with more information. (Page is in Dutch...)

The light is made from a small taillight. I Removed the four 3mm LEDs and replaced them with a single Luxeon and a 230mA driver. The reflector was drilled and a lens glued in place. A Switch protrudes from the reflector at the bottom.

Here is a comparison with the Planet Bike Superflash.
Top: My light.
Bottom: Planet Bike Superflash.

My lamp puts out a little more light than the Superflash. In fact, I kept increasing the current during the build until it did

[Steve K]

I've built just such a light recently:

....<pictures snipped>...

Even though I'm using a cap, the light still flickers as can be seen in the video at the above link. In fact, I can't notice a different with or without the cap. Also, at most I only get a couple of seconds of (very dim) standlight function from the cap. Two factors come to mind here:

1. I used a common silicon diode bridge rectifier, not a Schottky bridge, so there are more voltage losses. This may prevent the cap from charging as high as it might at lower dyno-wheel speeds.

2. Since there is only one LED paired with the cap, there is not as much voltage present in the circuit as there might be with multiple LEDs in series. This may also be preventing the cap from charging as much as it might.

Thoughts?

hi Jim,

Don't worry too much about the diode choice. This is mostly a factor in how slow you can go and still get light, but there's not much other effect. Once you are at a modest pace, the dynamo is putting out full current and you have full voltage across your led.

Standlight design is an interesting area of study, and it's one place where you can improve over commercially available lights. However.... just adding a big cap isn't going to produce the best results. There was a good standlight thread a few months ago (might have been all the way back to December??). We discussed quite a variety of designs and the pro's and con's of each. Alex Wetmore ended up using one of the better designs, I think, and wrote it up on his blog(?)... my memory isn't that good.

In any case, do a search in this subforum for standlights, or for Alex Wetmore, and you ought to find it.

regards,
Steve K.

[syc]

Standlight design is an interesting area of study, and it's one place where you can improve over commercially available lights. However.... just adding a big cap isn't going to produce the best results. There was a good standlight thread a few months ago (might have been all the way back to December??). We discussed quite a variety of designs and the pro's and con's of each. Alex Wetmore ended up using one of the better designs, I think, and wrote it up on his blog(?)... my memory isn't that good.

It was this thread. It looks like if you don't want to build the Zetex circuit that Alex used pulling power off the supercap, you can just pick up a Micropuck. It seems to be (not coincidentally) almost identical to what Alex built up.