Ideal incan setup?

[zamboniman]

One of the nicer things about incandescents is that you can run them on a square wave without a visible flicker -- the thermal mass of the filament acts as a better and more efficient filter than a capacitor would be across the leads of an LED.

So if you wanted to create a regulated incandescent light, could you get, say, a 12 volt battery, and feed a 6 volt bulb with a square wave with a 25% duty cycle (since doubling the voltage to a bulb quadruples the power)? As the voltage of the battery dropped, you could increase the duty cycle of the square wave -- this system would eliminate the problems of in-line inductors.

All you would need is an in-line MOSFET or IGBT and a microcontroller with a set of lookup tables (what duty cycle to output for a certain voltage input). Is there a discrete circuit that anyone knows of that could solve this problem? It'd have to be custom-tailored to a certain type of battery, though.

Lastly, is there a way to use some elastic material to isolate the incandescent bulb from shock somewhat?

 

[Outdoors Fanatic]

Lastly, is there a way to use some elastic material to isolate the incandescent bulb from shock somewhat?

Weapon light bulbs are already shock-isolated.

 

[sysadmn]

One of the nicer things about incandescents is that you can run them on a square wave without a visible flicker -- the thermal mass of the filament acts as a better and more efficient filter than a capacitor would be across the leads of an LED.

So if you wanted to create a regulated incandescent light, could you get, say, a 12 volt battery, and feed a 6 volt bulb with a square wave with a 25% duty cycle (since doubling the voltage to a bulb quadruples the power)? As the voltage of the battery dropped, you could increase the duty cycle of the square wave -- this system would eliminate the problems of in-line inductors.

I don't think you can count on feeding 12v to a 6v bulb even for a short time/low duty cycle. At a minimum, you'll significantly decrease the life of the bulb. ISRT a driver for incans that is rated at 100% voltage; I think it uses current regulation, not PWM (but could be wrong).

EDIT: Found the link http://www.cs.indiana.edu/~willie/lvr.html
It is PWM, I was wrong. But it does match output voltage.

 

[2xTrinity]

I think it uses current regulation, not PWM (but could be wrong).

Essentially the only difference between a current-regulated buck converter, and PWM, is the extend to which the output is filtered (usually using an inductor that gets "charged" during the on-cycle, and discharges over the off-cycle) IMO even with the thermal inertia of the filament, filtering the output at least somewhat is definitely a good idea, it doesn't need to be as "paper flat" as one would need to drive an LED, but certainly feeding the filament "spikes" at double the voltage isn't good, either.

Also, input-filtering is probably a good idea as well, as form what I understand, a more gradual current will cause the batteries to sag less than a "pulsed" current.

 

[mdocod]

I think the idea would work fine, just need to have a fast cycle rate. proof that it works is all over your house. AC.

 

[defusion]

something just popped into my head, and i have no clue if it's possible, so here goes:
could one also use this method to have multiple output modes on a incan without having a lower efficiency?
just adjust the frequency of the square wave to make output higher or lower, and because the voltage isn't lower the efficiency isn't (much) worse then in normal operation either.

or is the average voltage trough the bulb lower (because it's not sustained), so efficiency still sucks?

 

[2xTrinity]

something just popped into my head, and i have no clue if it's possible, so here goes:
could one also use this method to have multiple output modes on a incan without having a lower efficiency?
just adjust the frequency of the square wave to make output higher or lower, and because the voltage isn't lower the efficiency isn't (much) worse then in normal operation either.

or is the average voltage trough the bulb lower (because it's not sustained), so efficiency still sucks?

This is how a household dimmer actually works -- it's typically just a big transistor that chops off portions of the waveform with a variable duty cycle. When it comes to driving an incan bulb, efficiency really only depends on filament temperature -- the hotter the filament, the more efficient (and shorter the life). The heating in a resistor like a filament is a function of the RMS current, I believe the only time the waveform might practically matter is if the bulb is overdriven so that it is right on the border of instaflashing.

 

[mdocod]

could one also use this method to have multiple output modes on a incan without having a lower efficiency?

nope... keep in mind, that to keep the lamp from burning out, you have to use a reasonably high cycle rate (like 60hz or better, but with a transistor, you could easily pull off many thousand cycles per second if you wanted, distortion in this situation doesn't matter, raw efficiency is paramount, and transistors operate most efficient when used "digitally.").... anyways, the whole reason you have to use a fast cycle rate, is to cause the filament to act as the "buffer" in the circuit, it doesn't react fast enough to the voltage change, but rather averages a temperature and output (much the way a capacitor would help buffer out the peaks in a PWM driving an LED)... since the filament averages the the spikes out, you get the same efficiency you would have gotten with DC at that particular power consumption.

 

[sysadmn]

Found the thread for AW's hotdriver:
http://www.candlepowerforums.com/vb/showthread.php?t=166223

There's a link to purchase it in the first post.