Pulsing halogen Headlight Shorten Life?

loudhvx wrote:

Just do one filament and a pulse-width modulation. Very simple. No heat dissipation losses in the controller since it will be doing full on-off cycling, but the frequency will be high enough that the filament doesn't fully cool in each cycle.

True. That's one of two designs I have but the one I am leaning toward can't use that because it runs on ONLY the ground wire coming from the headlight (N-FET in the ground line). When the light is OFF, that line pulls up to +12V through the headlight filament (abouut four ohms impedance) and it charges the cap the 555 runs on through a diode. Then when the 555 and the FET turns on the headlight, the 555 circuit runs on stored charge until the light goes off again. I only have to run two wires out of the headlight housing to the handlebars to hook to the flasher. To restore stock wiring, I just tie those together.

I like that design because I only have to cut one wire to install the flasher and it will pulse either the low or high beam since both share the same ground line.

If you use a PWM for the "low" state there is no +12V pull up interval so I would have to supply a separate +12V line to run the pulser.

I know you like your FET's

All good designers use whatever they can steal from work.....:laugh:

In that case, I guess you'll probably just have to use a big resistor in parallel with a 5hz pulser.

Why not tap into the power line feeding the light dimmer switch? I guess you're trying to hide the unit in the headlight bucket? (And it would still need a third wire to give a ground.)

loudhvx wrote:

In that case, I guess you'll probably just have to use a big resistor in parallel with a 5hz pulser.

Yeah, that looks like the only way.


loudhvx wrote:

Why not tap into the power line feeding the light dimmer switch? I guess you're trying to hide the unit in the headlight bucket? (And it would still need a third wire to give a ground.)

This way I only cut one wire (the ground line coming from the headlight) and put a female spade connector on the "light" end and a male connector on the ground end.

Now, I get an orange wire double end it with male connectors and a black wire double ended with female. Connect them inside the housing and run them out to the handlebars: the orange wire is the light filament which serves double duty as +12V source when the light is off, and the load wire when the FET is on (light on also).

The black wire is the ground all the time, for the pulser circuit, for the light current through the FET, or light current when the bypass toggle switch is ON to give steady light on.

Only need two wires if it works. If the pulser fails, you can restore the headlight to stock by unplugging the pulser and connecting the free wire ends together.

I'll try to post a schematic.

I have a schematic but I have no idea how to post it here. I don't know why this has to be such an ordeal. Photobucket claims they sent me my password and there is nothing in my email inbox or spam folder. They are basically useless. I give up. I don't know why we can't just post a JPEG.


Maybe this will work. After only 40 tries, I finally got photobucket to work by trying every password I could think of. They never did send me that email they claimed they did.

I have not built the circuit yet so it may need some tweaking. The FET acts as a switch to connect the headlight filament to ground (completes the ground lead). The switch allows you to manually bypass the pulser and restore stock wiring connection. The 555 is a free running oscillator set for about 4 Hz frequency, and it runs off about 12V derived from the diode and 220uF cap when the light is off the filament feeds 12V to it. When the FET is on, it pulls down and the 12V goes away so it runs off the 220uF cap. The two 100 Ohm resistors keep a little current going in the light when the FET is off so it won't go completely off. I put a red LED in the 555 output line that drives the FET on for a visual reminder when it is pulsing, since it does not have an automatic shotoff at dark. That LED is to make sure and turn it off when you don't want the pulsing action.

NOTE: this circuit does not fully comply with the DOT requirements listed above, I was going to try it because it is so much simpler. I can also post another I did a while back that does comply.

I thought you were going to do a 20% to 100% oscillation, no?

With a 50 ohm shunt, you are going a lot lower than 20%, or was this the almost-fully-off circuit?

I notice you are using basically a 50% duty cycle on the 555. I think those strobes typically use a much longer duty cycle. In other words the brightness is only going dim for a short pulse, then remains full bright for a longer pulse. Otherwise you are effectively dimming the bulb by 50%, no?

I realize both changes (longer duty cycle, and shallower oscillation) will reduce the power available on the capacitor requiring a bigger cap and perhaps smaller feeding resistor. A necessity only if you want to cut into the ground wire as an inline circuit. But why even cut any wires? Doesn't your headlight have a 3-prong connector sub-harness with 3 separate plugs? Every KZ I've seen has that. You could just put your circuit in between harness and the headlight sub-harness. Then your circuit would have full-time power and ground. Then you wouldn't need the high-power resistors cooking away 2 watts, and no cut wires. Since the power lines will be intact, the only change using my plan is one extra wire coming out of the headlight bucket (the two diodes should be easy to tuck in the bucket). You would still be able to restore normal operation by connecting the two ground wires (one is ground, the other is actually the output ground to the bulb).

As long as the diagram is not enormous, you don't need photobucket. Just use the "select image file to attach" option when posting.

Just use a piggyback connector or y-splitter connector to add a diode leg to each of the positive power connections and insert the ground side of the circuit into the ground connection of the bulb harness.

Pull down configured Diodes over Resistors, yup...

Pending silicon vs germanium diodes, .4 - .7Vdc drop and
no heat build up as the resistor would do...

Nice Loudhvx!

Old Man Rock wrote:

Pull down configured Diodes over Resistors, yup...

Pending silicon vs germanium diodes, .4 - .7Vdc drop and
no heat build up as the resistor would do...

Nice Loudhvx!

Thanks, OMR , but those are not pull down diodes. They are isolation diodes, and are very low power. They just feed the voltage to the timing circuit. They need to be there to isolate the high beam power from the low beam power otherwise both filaments would be on at the same time. The bulb is still being switched on and off on the ground side.

Seems to me they would be forward biased on an active low from the timer... would drop ~ .7Vdc on the head voltage... no?

Unless I'm missing your diagram, only one beam would be active off the handle bar control... The pulsing would be the diode drop "On/Off" action....

EDIT: Re-read your posting... Ok, see what your doing... Switching the ground On/Off... Interesting.... Me likey....

Old Man Rock wrote:

Seems to me they would be forward biased on an active low from the timer... would drop ~ .7Vdc on the head voltage... no?

No. The headlight voltage is still at battery voltage just like always. There is a .7v drop, but that only means the timer is operating on 11.3v instead of 12 (or 13.3 instead of 14, etc.).

It would take a massive diode to pull down the battery voltage enough to reduce the headlight brightness, and then you'd be blowing fuses and melting wires. Fun, yes! But only for a moment.

In a different configuration, I suppose you could use a diode to reduce the voltage to the bulb by putting it in series, but that would be no different than using a resistor to reduce the voltage by .7v. In that case, there is no difference between the resistor and the diode, both will reduce the voltage by .7v and both will have the same effective resistance, both will heat up by the exact same amount, and both will be dissipating the same power so both will have to be about the same size. Resistors are more reliable in that type of duty.

I wouldn't think the concept/need for dropping large voltage swings for the flickering effect... A 1 volt drop would be significant for adding "bike" visibility to other drivers...

Question for those that have this on their rides... What's it like for night driving or is the flickering switched off? Bypass switched ground to the filaments for night driving... Would work, yes indeed....

I just saw this yesterday morning, opposing direction guy on a HD had this... caught my eye immediately, even in the daylight...