700mA from 500mA

[tspoon]

I have a 6v 3watt bicycle hub generator (500 mA capability), from which I would like to run a luxeon III led. I have enough voltage OK, but the only switchmode type drivers I can find all need a higher or at least equal current capability to 700mA required by the LED (as I understand it). The generator will burn out if I connect it in this state.
AFAIK my options are:
1. Some other kind of regulator/Power supply circuit that can boost current, while reducing voltage. (I guess this would incorporate some sort of high freq. transformer)
2. Capacitors on the 6v Supply to act as some sort of current reservior.

Can anyone shed some light on either of these options, or even some other way round the problem. For the capactive reservior option, can anyone tell me how much capacitance would be suitable/practical. The LED needs about 3.7v at 700mA

 

[Doug Owen]

I don't think you have any real problem. Assuming it's a generator, not an alternator (DC, not AC) the most I can see your needing is a modest resistor, 2.3 divided by .5, about 5 ohms (two watts or more).

Hook it up and watch the current as you spin it up.

Doug Owen

 

[The_LED_Museum]

Most bicycle generators I know of produce AC, not DC. They're still called "generators", even though they produce AC.

 

[Pi_is_blue]

The LuxIII will run fine at levels less than 700mah.

 

[tspoon]

I would prefer to get as many lumens as possible from this setup, as this will mean greater confidence and safety/visibility during my daily commute. I think I'll probably do some experimenting with capacitors when my luxeon stuff turns up. May also use a 4.8v nimh pack as a current reservior - charging it when the light is off. Will let you know what occurs.

 

[gwbaltzell]

[ QUOTE ]
The_LED_Museum said:
Most bicycle generators I know of produce AC, not DC. They're still called "generators", even though they produce AC.

[/ QUOTE ]

I have never seen or do not know of any bicycle "generator" that produces DC. There a large number of issues here. The common bicycle "generator", the one that pushes against the sidewall of the tire (tyre), has a large amount of drag and should likely just be put in the trash can, it is a toy. Sanyo made, past tense, one that fit the lower bracket and pushed against the tread. This and the one that's part of the hub are likely the only ones worth trying to do anything with. Both types would have to be bought used. The two most effecient ways of converting the AC to DC would be a bridge made up of 4 Schottky diodes, or a syncronous rectifier made from 4 MOSFETs. The diodes are simplier but will have a higher voltage loss. Next a voltage regulator to feed either rechargeable batteries or a very large (multi farad) capacitor. See this thread on capacitors. An on/off switch. A current regulator to feed the Luxeon(s). I would use 2 or 3 1.2 W ones or a III instead of a 5 W for three reasons. 1. I would use some additional wide angle LEDs (cyan and amber) just to make the bike more visible to cars. 2. You can't make 5 W from 3.5 W (though you can underdrive the Luxeon). 3. The small die size makes the optics easier to calculate. Optics? The reflector or reflector and lens should be designed to produce a trapezodial beam on the pavement just like a car headlamp. There used to be an incandescant PAR36 4.7 V 0.5 A bulb (#4512) made just for this purpose. And incandescants are AC or DC.

 

[gwbaltzell]

A review of this thread made me realize a few people need a review of a couple of terms. Craig, of course, got it right.
Generator - 1. A. One that generates, especially a machine that converts mechanical energy into electrical energy.
Alternator - An electric generator that produces alternating current.
- American Heritage Dictionary

An alternator is a sub-class of generator. Typically the only time the usage of the term generator by itself implies DC is in cars. Even in cars the alternator's output of AC is rectified to produce DC.

Now, as to the original issue, the most readily available bridge rectifier (non-Schottky) is the ratshack 276-1152 a better solution is to make a bridge (4 diodes) from Schottky rectifiers. Available from digikey.com and others. Three main factors in selection are If forward current, 1 A would likely do, Vrrm reverse volts, now the typical bicycle generator is 6 V AC RMS but can produce higher at high speeds (6 V / 0.707 = 8.5 V) so likely 25 - 30 V is OK. The third factor is Vf forward voltage (loss). The lower the better but probably won't find below 0.3 V, so total for the bridge will be around 0.6 V. I would put a large capacitor on the unregulated side just to try to capture peaks, working voltage again 25 - 30 WVDC, capacitance as high as I could find within the physical limit I would tolerate.

 

[INRETECH]

Make sure you have something to limit the current and voltage, a couple cents worth of parts can protect the expensive LED(s)

 

[3rd_shift]

Luxeon3s seem to be quite versatile compared to some leds.
500ma ought to be fine with it.
It should still be brighter than a luxeon1 at 350ma and safely underdriven to boot.
At 1.5x the cost of a lux 1, It's still a bargain even at 500ma. /ubbthreads/images/graemlins/thumbsup.gif
Just remember to still give it some kind of a heatsink though.
500ma is still plenty of power to fry any emitter without a heatsink. /ubbthreads/images/graemlins/mecry.gif
Try to measure the voltage of your generator.
If it's really putting forth more than 6 volts, but less than 7;
You may have a good power source for a luxeon5.
I wish I knew how steady, or regulated your power source is.

 

[gwbaltzell]

I was still recommending from my prior post a voltage regulator to step down the voltage to feed the rechargeables or really big capacitor. And a current regulator to step up the voltage and drive the LED (s). Even with losses, if you have very good regulators and are riding at a fair clip, you might be able to nearly fully drive a III. But I think (2) 1.2W would give you more reserve for the battery and you could make due with simplier optics by having one throw near and the other far. And I'm still recommending cyan and orange wide angle standard LEDs, preferably flashing in some non-simple pattern, to get motorist attention.

BTW - I don't think you could burn out that generator, even with a dead short. More likely failure will be a broken wire or insulation failure. If either occur have it rewound! A hub or the Sanyo are the only ones even worth thinking about on a bike.

 

[jamaica]

Gee, please correct me if I'm wrong, but I'd think of the standard 3w
bike alterator at a given rotation rate as a *current source*. If true - and
I think it is - it won't make more than about 500mA no matter what you load it with.

The simplest configuration, then, would be a bridge rectifier. This would use
the full current capability, but not the full voltage output of the generator.

The brightest configuration would be to rectify and filter, then downconvert
with an efficient switching driver. This way you could harness and deliver
6v x 500mA = 3w to the LED versus ~4v x 500mA = 2w for the simple rectifier.

Best,
Jamaica

 

[gwbaltzell]

I don't think these are current regulated in any way. But it is true that at any given rotation there will be a max. current into a basic load. Now if we weren't worried about storing any power, and just running the warning lights from batteries, we might think about winding a transformer (this is AC) to step the voltage down and current up. I don't know the frequency. Its likely low and might take at lot of "iron". Then just rectify and smooth it and add a high wattage zener to limit the voltage (and therfore current).
I guess I wasn't clear that I was speaking originaly about a down converter to try to deliver full power to the LED (s). The second post was just a way to get DC to the regulator. Actually two. One to try to take as much power from the genarator as we can get at any given speed and produce a regulated voltage to feed rechargeables or a real big capacitor and a second to take that and regulate the current to the LED (s).
I'll try and run some numbers with my Sanyo. This may take a while.

Edit:
Having second thoughts about worring about current regulation to the LED. I was really thinking 2 NiMh cells and if we are fairly constantly charging them with a fixed voltage why not use a simplier current limited boost like just mentioned here. The added advantage is longer run time if for some reason we fail to get the charging and indication (eventually) of the failure without total loss of light.

Edit again:
I'll just add to this post with prelimanary info. if thats OK with everyone. I should always check real world devices before posting. The Sanyo can put out up to 20 VAC RMS with no load! (20 / 0.707 = 28.3 V peek) Must revise some values. I can see were the idea that these are current sources comes from. The current comes up at a fairly low speed. I got 300 mA at low rotation. (Got to move my speed sender to the rear wheel to get better estimates.) However, for a dead short I could get up to 660 mA. The rest of the current issue comes with the behaviour of incandescants. My 4.7 V bulb made for bike use draws 390 mA on a DC bench supply at 4.5 V. At 2.5 V it was still drawing 340 mA. The frequency from the generator went from around 25 Hz to over 80 Hz. A transformer is likely out. And a hub will be turning at a slower speed, so may behave differently!

 

[jamaica]

Hey, thanks for the real-world data! I know the bike alternator doesn't
have a built-in current-limiter, but it doesn't need one -- it produces a
certain max current at a given RPM, and not much more even into a heavier load.
As you said earlier, there's no danger of burning the alternator out. 3 watts
simply isn't enough energy to hurt the windings I've seen in these devices.

The large open-circuit voltage argues in favor of simplicity: a bridge
rectifier straight into the LED.

For constant light, feed the bridge into a battery charging circuit, then run
the light off the battery using a suitable driver. (but hmm..I think you said
this already too, eh? Bravo!)

Good stuff on that ol' alternator -- I always loved the things when I was a kid. Thanks.
Best,
Jamaica

 

[PeLu]

1. Almost all bicyle generators/alternators or however you call them are (pretty good) constant current sources. This current does NOT depend on the RPM (as long as you stay above a minimum RPM).
2. Of course, the maximum voltage they can supply, depends on the RPM.
3. There is nothing built into the generator/alternator to limit the current (it is just not necesssary). But very few of them have Zener diodes to limit the output voltage.
4. The open circuit voltage can easily be 60V or more and with some other oddities (Ferranti effect) you could get even higher voltage spikes.
5. By far the best solution are hub dynamos, nothing else comes close. The newest Shimano is a little bit better, but does not come in range of the SON. But for most people they are sufficient.
6. There are several pretty good side mounted dynamos, just in case a hub dynamo is not possible.
7. An almost perfect manual about all of this things is Olaf Schutz' 'prayer wheel', only available in German. He tested many dynamos in between zero and ~ 200 km/h and SON and Shimano change short circuit current only by 10% or less in this range. That also means that shorting the generator will not harm it.
8. That means, if you can use more than one LED, just take two of them anti-parallel. So, no losses in rectifiers. Or four of them, just two pairs. When you need more light, you could add even more pairs of LEDs in series.
9. If you want to use some switching power supply, be shure this device fits to you dynamos behaviour.
10. hat also means, ther is actually no 6V bicycle generator. But they have to state a voltage, so they do it for the usual load of 12 Ohm.
11. If people are intersted which dynamos do NOT fit into this pattern, I could look up the prayer wheel.

 

[gwbaltzell]

Thanks PeLu. Didn't know anyone was still making hub units. The U.S. has always been a little backward when it comes to bikes. I've never seen a good side mount in this country, or on any mail order site. While I like the idea of back to back LEDs for its simplicity and no outside losses, I think storing power would be better. Especially for stop and go city riding. And I think this is what the original poster had in mind. Still, for trail riding it might be the way to go. Have to have Lux. IIIs if we're only using two and they would be underdriven. Maybe steal a little power to charge a capacitor for a small LED for stops.

 

[PeLu]

Yes, there are lots of fine rearlights available which contain supercaps. At least mine is bright enough to light up the whole bicycle room (with about 30 bicycles) for several minutes.
Shimano just came with a new hub dynamo to the market. The came closer to the SON in efficeincy and price.
But what may be of interest for flashaholics: The strict German and Austrian laws about bicycle lights (limit to 2.4W and so on) forced manufacturers to make real good reflectors and lenses, sometimes this lights have more useable light than 10W halogen/battery setups.
I have a SON in my most used bicycle and did not even make a switch, light is on all the time, just changing the front bulb once a year. One reason more I'm looking forward to a LED setup.
And for side mounts: There is the Swiss LightSpin (one of the few exceptions mentioned before), read Peter DeLeuw's comment:
http://www.pdeleuw.de/fahrrad/lightspin-e.html

But the main thing about dealing with generators is that they behave so different from batteries.

 

[Bandgap]

Bridge rectifier, Shottky diodes if you are worried about cutting losses, and two Luxeons in series to make max use of the available power.

I have done this with two generators (dynamos in the UK where I am) and it works well.

Two Luxeons are better than one-and-a-power-converter, as lower current means more efficiency from the leds.

And the second Luxeon is usually smaller than a converter.

However, on a hub dynamo, low-speed flashing may annoy you.

Steve

 

[gwbaltzell]

Well, I was never really happy with the DC soulution. The losses at each step leave little power left. Let's look at the losses. If the generator is producing 6 V into the 500 mA load and we lose .6 V at the bridge we are down to 2.7 W already. And this is assuming we can still draw 500 mA. Rechargeables, at best, give us 70% of the energy back. Now 1.89 W. Add in the two converters and we're down to a 1 to 1.5 W And every step also adds a point of failure, including the on-off switch. Looking at the generator as a current limited source, PeLu soultion of two anti-parallel (or back to back as we say in the U.S.) LEDs gives us almost the same result as an incadescant bulb. And since each is only going to be on half the time, my statement about them having to be IIIs may be wrong. Downsides:
1. Same as incadescant, no light when you stop.
2. Flicker (flashing) at low speed may be more apparent.
3. If one LED fails open the other's PIV will be exceeded.
A fix for #s 1 and 2 would be a third battery operated LED whose output would be inverse of the generator output. This would also provide a backup if the generator fails.
What about tail and "running" lights. We have some cheap battery operated ones that will run a long time. This is likely the best solution for these areas. Though you could buy or build a "super cap" based unit and steal some power from the generator. But remember the max. voltage will be the forward voltage of the LEDs (3.42 V).

 

[Steve K]

If there is a good reason to use the Luxeon III, then just full-wave rectify the AC power and drive two of them in series (as Steve aka Bandgap suggested). He's done a version using ElektroLumen's 30mm optics that has worked well.

I'm currently using a 5W Luxeon with my Schmidt hub dynamo, and am fairly pleased with the results. The lack of good optics designed for the 5W Luxeon led me to mount it in an old Union reflector assembly. Works well enough, but I've obtained an old BiSY reflector & optics for tinkering this summer.

By the way, heatsinking the LEDs is critical!! For me, this was the most time consuming aspect of the design.

I also have a led taillight wired in series with the Luxeon. This results in a very, very bright taillight! The only price that I have to pay is that there is less light produced from the headlight at very low speeds (5mph or less).

Steve K.

 

[Steve K]

[ QUOTE ]
gwbaltzell said:
Well, I was never really happy with the DC soulution. The losses at each step leave little power left. Let's look at the losses. If the generator is producing 6 V into the 500 mA load and we lose .6 V at the bridge we are down to 2.7 W already. And the is assuming we can still draw 500 mA. Rechargeables, at best, give us 70% of the energy back. Now 1.89 W. Add in the two converters and we're down to a 1 to 1.5 W

[/ QUOTE ]

This isn't true, due to the dynamo's high source voltage and large internal impedance. In practice, when traveling at typical speeds, the dynamo output voltage will pretty much increase to whatever is required to shove 0.5A through the load.

A fellow named Nick Ray did a great job of characterizing the Schmidt dynamo at various speeds and load resistances. It shows exactly what the V-I relationship is, and the effective source voltages and impedances can be derived from it. I would be happy to send it via e-mail.

Steve K.