gnome

I thought I'd better wander in here and get some help with a problem I'm having with blowing bulbs with my dynohub setup. Disclaimer: I have very little knowledge of electricity and if everyone starts getting too techincal I will look blank and ask to to repeat it, simplified for dummies.

I am running a Shimano DH 3N20 dynohub built into a 20" (ERTO 406) rim on a fixed gear commuting bike. The light is wired to a 2.4w halogen headlight designed for a different hub dynamo and a B&M taillight. The head light has a sensor and comes on automatically but doesn't have a switch. I am constantly blowing headlight bulbs due, I am sure, to the higher speed of the small wheel.

I am sick of replacing bulbs and am looking at buying a new LED headlight, such as the B&M Lumotec IQ Fly Plus. I will have to import whatever light I buy. I am not skilled enough to build my own headlight.

Will I fry the LED headlight as well or will it be safe to use with my setup? Can I run a higher wattage/voltage bulb in my extisting headlight?

I did have the same problem with my recumbent which is fitted with a Sturmey Archer XFDD dynohub, but I solved that by adding a secondary headlight to it.

znomit

Is it the right bulb? I'm not 100% sure but I think the total light is 3w, some run 2.4w up front and 0.6w tail light. If you're not running a tail light you need a 3w bulb.

I can throw together a simple single LED light for 20-30 kiwi$ if you're interested PM me.

bobdell

Use a 3.ow bulb if you are not running a taillight. Go here https://www.peterwhitecycles.com/index.html for lots of info.

Unknown Cyclist

Do you have a rear light wired up to the dynohub ?

Generally dynohubs are 3 Watts.

A 2.4W front + a 0.6W rear = 3W

A 2.4W front + no rear = blown bulbs.

You should be able to buy 3W halogen bulbs to suit your headlight.

gnome

I'm running a taillight as well as the headlight and running a 3w on the front as well. It is just using a standard hub-dyno with a small wheel instead of an expensive SON-20 (designed specially for small wheele d bikes).

Thanks for the offer of the LED light. PM will be sent soon.

n4zou

An LED headlight will not burn out like your halogen lights are doing. Your running a smaller wheel so burning out halogen head and taillights would be a problem as you've found out. LED's produce light from current and not voltage like a filament bulb. Generally a standard bicycle dynamo is limited to 500mA but my cheap Schwinn dynamo runs 700mA. My original filament bulbs burned out after a couple of hours due to that problem. I built an LED headlight with LED's rated for up to 1 amp. I have over 100 hours on it with no problems and it's bright. Having a dynamo running 700mA is now an advantage. You should let Znomit build you an LED headlight.

Michel Gagnon

You have a wired taillight? So the problem isn't there. Is it one of the Lumotec or Schmidt headlights? If so, there is some overvoltage protection in them. But if you have a basic Union – or any no-name – headlight, there is no zeener diode to protect it from overvoltage.

So if you don't have overvoltage protection, you will burn the headlight almost as soon as you ride faster than 8-10 km/h. If you have overvoltage protection, you should be OK, except that you are using a 700c generator in a 20" wheelset. In practice, this means the wheel is turning faster than the designer thought it would, and for the headlight, riding at 30 km/h (real speed) is the equivalent of riding at 55 km/h. What happens? The zeener diode overheats and shunts off; it doesn't protect the bulb anymore, so the bulb burns.

As for the rear taillight, it won't burn immediately, but since it receives all current, it will die sooner if you ride without a headlight. Likewise, using a LED-based headlight will work, but the headlight won't last as long as if it were not overvolted. How long? I don't know.

Solutions
– Get a proper 20" generator.
– Install a secondary headlight. You could use the Schmidt E-6-S. Any other standard headlight would work, but bear in mind that since they are connected in series, you won't be able to shut one off.
- Add some kind of additional overvoltage protection on the wire between the generator and the headlight. This would be some other zeener diode, but don't ask me for any technical information on its particulars. The diode would clip the voltage at a comfortable 6,5 V, and since it would sit on its own, it would not overheat the generator or headlight. I assume this latter solution would cost less than 10 $.

znomit

Yep, doing that

znomit

Coming together nicely.
Nice tight beam and 140lm.
Lucky to find an end cap that fitted my al bar and Jaycar had splashproof toggle switches.
The cree optics from Kai are very nice, and only a buck each

n4zou

Looking good!

trekker pete

got a question about something I have heard here multiple times concerning LED/halogen bulbs.

I keep hearing that voltage determines brightness in a halogen and current in an LED.

Way back when, in basic electricity and electronics school in the navy, I was taught Ohm's law. Ohm's law states in a DC circuit, current is a function of voltage and resistance. I=E/R I being current, E, voltage and R resistance.

A halogen lamp has a constant resistance, as far as I know. Therefore, as the voltage increases, so will the current. So, you could say it is a voltage controlled device, but, it is the increase in current that actually causes the increased brightness.

I'm a bit more fuzzy on LEDs. Do they vary their resistance as the voltage changes? Just curious. Thanks.

supcom

An LED, being a diode PN junction, will typically have a 1.7V maximum voltage drop when biased in the forward direction with enough current. So, yes, the LED does act as a variable resistor once the current level reaches a certain value.

This is why it is critical to have some form of current limiting circuit when driving an LED from a battery. In the case of a dynohub, the generator, by design, can produce no more than 500 mA of current, so modern power LEDs, capable of taking 1 amp or more, do not require additional circuitry beyond a simple rectifier.

snowranger

You are fast, and I like how your set-up is so compact. Bidirectional capacitor I guess?

n4zou

A dynamo and halogen bulb is AC current/voltage. You need an advanced electronics class to deal with the math involved. AC math is where most students drop out or flunk.

trekker pete

I used to be quite the impedance calculating mofo. Let's see.... uhhhh

Xc=2pifc

Xl=1/2pifl

Xl squared + Xc Squared = Z squared

yeah, I still got it!

Just don't ask me to figure out Q or any of that other stuff.

znomit

I got lucky finding the end cap and the switch, sealing the end and mounting a different switch would have been a lot of work.

Cap is just for smoothing, so polarized. The small wheel means that the dyno will be outputting 500mA nearly all the time for commuting speeds, no need to worry about tuning the circuit.

gnome

Thanks for all the useful information and suggestions. And a very big thanks for Znomit for building me a light.

Unknown Cyclist

Is the front light assembled yet ?

ellerbro

LEDs do have a variable resistance. Looking at the current-voltage graph can help understand their behavior. The slope of the curve (well, technically, the reciprocal of the slope if current is on the y-axis) gives the dynamic resistance (R = V/I). Looking at the attached pic, as the voltage increases from 0, the current does not increase = infinite resistance. As the voltage approaches the "knee" of the curve, some current starts to flow = less resistance. Once the voltage exceeds the knee of the curve, small increases in voltage cause large increases in current = small resistance. You can calculate this dynamic resistance by calculating the slope of the curve at the section you're interested in. Between 3 and 3.5 Volts (0.5 Volt change) the current increases by 0.4 A. Therefore, R = 0.5 V/ 0.4 A = 1.25 Ohms. So along the steep part the LED behaves as if it were about 1.25 Ohms (a 1.25 Volt increase results in a 1 A increase). This type of calculation, however, is only valid on a linear section of the curve where the dynamic resistance is constant.

The static resistance can be calculated by taking V/I at various points on the graph:
At V = 2, I = 0 so R = infinite
At V = 3, I = 0.2 so R = 15
At V = 3.5, I = 0.6 so R = 6
It is seen that R continues to decrease as V increases.

znomit

Yep, all done.

Should handle the rain ok.


Tidy thing. The plastic end cap made it very easy.


The dinotte mount. Geeze this was easy. I've spent days doing mounts that aren't nearly so nice. Mixed up some dental putty and plonked it on the bars, set the light lightly onto the top. 5 min later it was set.



Quick up and down the street to test it. All good.

ellerbro

Very cool light. I've been looking for a waterproof switch for a DIY tail light. Which one are you using there?

znomit

Its from Jaycar. Its only splashproof but on the back of the light that should be fine.
There are a couple of other ways you can go, most toggle switches have rubber boots you can fit. Or you can get a switch that has a built in rubber cover.