Thanks to all who helped me with my previous question on halogen bulbs!

Here's the next one: What is the voltage-limiting or power-limiting requirement, based on how voltage that is greater than 6.0 volts degrades the life of the bulb? Clearly 12 volts is too much.

Has anyone looked at the voltage limiters on some existing halogen headlamps? I would, if I had one.

I'm thinking about using a halogen lamp with a bottle generator.

Let me simplify: If I want to build a voltage limiter so I can use a 6 volt halogen bulb with a 6 v bottle or other dynamo, for what voltage should the limiter become active?

3.5V (and it should be driving an LED!)

I'm not sure we're on the same page. If I run a 6 volt 500ma halogen bulb direct off of AC, the peak value of voltage is 6 times the square root of 2, or 8.5 volts. If I want to limit the voltage the bulb is exposed to 110%, I would want a slicer circuit (say, using back to back zeners) that "slices" the peaks off the sine wave at positive 9.3 volts and negative 9.3 volts. My real question is, what is the allowable overvoltage percentage, in this context?

Where does your 3.5 volts come from?

Sorry if I'm not showing a sense of humor, but I am REALLY looking for some input from people who might understand how to protect a halogen bulb.

Or I just don't get it ...

But let me be a little clearer if I can: I'm not making an LED light right now. I am converting a bike lamp that uses a conventional hard-vacuum tungsten bulb to halogen, and I am concerned about the effect of overvoltage on the bulb. Based on my experience in automotive lighting, a little overvoltage shortens bulb life by a lot. An AC slicer would be the simplest circuit, and a rectified and regulated DC voltage driver would be one of the most complex. Less complex but more expensive is to run the light off a battery and use the dyno just to trickle charge the battery.

According to Wikipedia:
That being said, a 5% over voltage will decrease your bulb to halff life.

And according to this site, at 10% above 6v, your bulb will last only 50% of the rated life.

You mention a peak voltage of 8.5 volts, however the RMS of the AC voltage is still only 6 volts in which case that value is what is important to a 6v halogen bulb. Whatever above the 6v produce by the generator happen to be, that extra voltage will shorten the life span of the bulb. Let say the generator produce a 6.6v. That is 10%(if my math is correct) which means the bulb will last only half the rated life(according to Reflectalite) but will be brighter by probably 30%(according to Wikipedia).

Personally I never played around with generator for a bike but I have looked into it. So I cannot be 100% certain this info is correct but at least it looks good

If the bulb is rated at 6 volts, I'd limit the voltage to 6 volts.

The circuit is very simple. Take two zener diodes and wire them in series with a cathode-to-cathode or anode-to-anode configuration. Then place these zeners in parallel with the dynamo output, either at the dynamo or at the headlight. Most modern bike headlights already have a zener limiter. The the sum of the zener and forward voltages of the diodes should equal 6 volts, the limiting voltage. The beauty of this limiter is that it does not degrade low speed/voltage performance. The two zeners probably cost less than a single halogen bulb.

Many thanks to Colleen for finding that basic info on degradation rates with voltage. That's exactly what I hoped to learn.

For the 6 volt slicer advocates, please tell me: how do you plan to deliver 3 watts to a lightbulb using a 6 volt sine wave, if you are going to re-direct some of that power to heat a pair of zener diodes rather than a tungsten filament? That's what designing a slicer such as you suggest will result in, when the voltage is such that the bulb would NOT be over driven.

What is needed is for the limiter to absorb zero power for any effective generator voltage up to say 1.03*6.0 volts (3% overdrive), or 6.18 volts. This implies that we don't want a limiter to act until the instantaneous voltage exceeds 8.74 volts on the positive half cycle and -8.74 volts on the negative half-cycle. Ideally the concept needs to be an 8.74 volt slicer. Practically, well, it's time to head off to the ON Semi website and see what I can get in a power zener diode these days, and it a small power Schottky diode. I might not be able to get a zener that will so precisely protect the bulb.

Colleen, I think your math IS correct. Again thanks.

This link describes what diodes and zeners do.
https://www.reuk.co.uk/What-is-a-Zener-Diode.htm

The basic shunt regulator circuit is shown below the "Zener Diode Voltage Regulator Circuit" heading. Not shown is the load across the "stable voltage" terminals. The load has a resistance r.

The Zener conducts only when the the voltage at its terminals exceed its zener voltage. If the voltage at its terminals is less than this threshold, then it does not conduct and is out of the circuit. Thus, all the current passing through series resistor R, will pass through load resistor r.

If the voltage at the zener's terminals exceed its zener voltage threshold, then the zener will conduct without limit (up to its power ratings). This added current will pass through the zeries resistor, R, and increase the voltage drop across it. This will guarantee that the zener voltage appears at the load. The current flowing through the load will be the zener voltage divided by the load resistance, r.

The series resistance, R, is the winding in the dynamo. It also contains a fairly high inductive component. This helps voltage regulation because going faster not only increases the dynamo's voltage output but also increases its frequency. Thus, dynamo output will not increase linearly with speed when there is a resistive load.

The big complication is that the dynamo output is AC not DC. That's where the 2nd zener in series with the first (cathode-to-cathode or anode-to-anode) comes in. A zener acts like a normal diode in the forward direction. It will conduct without limit (up to its forward power rating) and present a nominal fixed voltage drop of approximately 0.7 volts. In one half of the AC cycle, one zener will conduct in the forward direction and the other will act as a shunt regulator. In the other half cycle, the zeners' roles will be reversed.

The peak voltage across the "stable voltage ouput" terminals will be limited to the forward voltage drop of the conducting diode plus the zener voltage of the other. This will not be a simple sinusoid; it's a clipped sinusoid.

The question of where to set the zener voltage depends on how one interprets the longevity specs for incandescent bulbs. I interpret them to be DC volts because that's what's on the spec sheet. That's also the instantaneous voltage. Therefore, I'd opt to set that threshold at 6 volts. I've used a 5.6 volt zener and assumed a nominal 0.7 volt forward voltage drop. That comes to a clipping voltage of 6.3 volts. I'm slightly over my desired threshold.

I used a 5 watt zener, like the 1N5339 because it comes in an axial package. I can solder the 2 zeners together and add leads to both ends. I can also put heat shrink around the diodes. The package looks like a wire. I attach it across my headlight because I use a bottom bracket generator and attachment there is not practical.

The setup is discussed here.


The 5.6V Zener is well too low, leading to routine underpowering of the lamp. Without overvoltage, i.e 6V rms, the AC swings between -8.5V and 8.5V. Clipping it at 6.3V will reduce power to well below nominal, see in particular the above site. I have been using 8.2V Zeners, clipping at 8.8-8.9V (the knee is soft anyway). If you are cautious, you can go with 7.5V. In any case, the diodes are supposed to prevent the burnout under unusual circumstances, such as riding down a hill and not do much under usual circumstances. Correspondingly, the discussions of bulb life shortening under overvoltage are not particularly relevant.


The diodes indeed should have 3-5W to prevent diode burnout when a bulb burns out and possibly to prevent tail lamp burn out if the front bulb burns out.

Thanks for the reminder on zener basics, but I think I understand them pretty well. My uncertainty is in the area I've highlighted.

I'm going to put my limiter inside a the headlight shell of a Soubitez front block headlamp/generator. It needs to be small, as well.

If you're lucky, your Soubitez generator will have an external terminal for powering a rear light. Some models did. Then you connect the diodes externally between that terminal and the frame.

Some Soubitez block generators have a propensity for popping open due to road vibration. You will want to make sure that internal diodes do not push the bulb/reflector outwards to make the generator more susceptible for popping open.

Good Luck.

Thanks.

Mine has a screw clamp to hold it closed - I don't think we're popping open. It also has a Fahnstock clip for the output lead, so I can attach the extra lead there.

OTOH, I've since found my front rack gets in the way of the front block, so I may have to use a different bottle dyno (another Soubitez, actually) on the rear wheel, ahead of the seatstay to ensure access while panniers are mounted. In that case I'll use a B&M front light. This generator also has a spring loaded output clip, so I should be able to solve wiring the same way.

I finally dove into the Easterbrook article, and it seems that any slicer is a compromise solution. There may be potential to build an active adaptive slicer that adjusts the clip level in response to the generator input voltage, but such a circuit is complicated and, I'm not sure it's even feasible. Looks like halogens should be abandoned in favor of LEDs when they get big enough ... Hey, wait, they have been!