Incandescent Bulb

i can't understand the flickering in incandescent bulbs.

In folouro tube is this effect but not in element bulbs

Look in, may help

http://cr4.globalspec.com/thread/51906?frmtrk=cr4sd#newcomments

If the "reduced light" is because of the varying arc light intensity added to the [nearly] constant incandescent, there might be another approach. If you could use arc lighting yourself, but somehow have it 90 degrees out-of-phase (assuming that both positive- and negative-going portions of the cycle light the arcs, you want to be halfway between those 180-degree-apart peaks), you'd be getting far more uniform illumination on a time basis. I'm no EE, but perhaps someone can suggest a way to do this. It shouldn't be all that different from the three-phase converters made to run electric motors off normal home power. There's a way to use a second motor to generate that output for the actual drive motor, which suggests that a similar method might be possible on a reduced budget.

Run a test: illuminate an area ONLY with incandescent bulbs, and see how much variation you have, as seen by the camera. If negligible, it points to the already-present arcs being the source of your problem, and suggests that you want to add light only during the low points.

GA. An audio amplifier using a photo-transistor as input should do the job. It would automatically compensate for changes in the light due to anything. It would require a fast-acting light source.

In this connection, more light will drive the amplifier down. It might take some experimentation to determine the optimum values for a given Amplifier and photo-transistor. R2 and C1 filter out high frequencies and reduce the possibility of oscillation. C2 keeps DC out of the amp input. R3 is a gain control. R4 and R5 might be needed to set a minimum light output level.

A set of high-intensity LEDs would have the fast response required of the light source. Again some experimentation would be required; expect to blow out a few LEDs in the process.

As I think about it, a pen laser illuminating a photo-transistor to trigger a camera and flash is probably easier... In this case the photo-transistor and R1 would swapped (R1 to ground), and the only other component would be C2. It would of course require a camera with an electronic flash trigger input.

If your problem is obtaining constant lighting for a photo-finish, you might try a different approach. Instead of trying to obtain continuous illumination, you might employ a sensor to determine when the first wheel crosses the finish line. This sensor could be a line switch, a broken light beam sensor, etc. Then you only need one bright flash to photograph the winner at the instant his/her tire trips the winner switch. You can use a simpler camera and a conventional flash. It would also be less annoying to competitors and the audience.

I have to record complete results, sometimes more than 20 minutes of capture. Often my files are over 10 gigs. I'm looking for a capacitor, sized, to "damp" a 1000 watt load at rectified 120v, so I can mount them on each lamp.

If you are using full-wave rectified DC, the ripple is mainly 120HZ. In a 1000W lamp, the thermal mass of the filament should make that frequency essentially invisible. If so, your capacitor would do no good, no matter how large. If not, another approach would be to use low voltage lamps, such as aircraft landing lights. You would need fairly large transformers to power them, but they have such thick filaments that it takes several seconds for them to cool down - you would not need DC -60 Hz AC would be fine.

On the other hand, if your lights are supplementing existing sodium-vapor or similar lights, The ripple of those lights will always be there, even if your lights have none whatsoever. If that is the case, then my earlier post gives a concept that should 'fill in' the dim spaces between peaks of the existing light, and do so with considerably less energy.

good luck on your project!

As has been pointed out, incandescent lamps with thick filaments produce fairly constant illumination. If flicker from arc lamps is your problem, it may be easier and more energy efficient to synchronize your camera shutter with the flicker, than to attempt to condition the lamp power.

Good point!

Yosemit3. I can't imagine needing better than 1ms resolution in timing. Anything closer than that is a tie!

You might try using 600 or 1200 lines per second, or other multiples of the line or ripple frequencies.

I'm having great difficulty imagining what the output of a linescan camera might look like with a fixed background and moving foreground. Could you post a sample? Surely you must supplement the line camera with images from a frame camera?

Another possibility: I assume the camera is aimed along the finish line, taking repeated images of nothing but that line. If so, a laser with a cylindrical lens could illuminate nothing but that line. Use a red laser, with a red filter over the camera lens (ideally a dichroic filter set for the exact color of the laser). The filter will remove virtually all the light from sodium- or mercury-vapor lights, leaving only the laser light for the camera to see. (ever notice how women's lipstick appears black under such lighting?)

Now that I think about it, that red filter just might do the job with your existing incandescent lamps.

that would be some laser!, I'm using 5000 watts of very narrow spots.

1ms is approx 0.6 inches in a sprint(600 inches/sec), I can interpolate to 0.04 inches in normal conditions.(at my max line rate 0.006 inch, I'm sure my optics aren't that linear, but I can show it on the screen)

(if the leading edge of both front wheels are in the same row of pixels(blow up to 4000% in photoshop) then you count the number of pixels each wheel "illuminates")

Pics: 1 daylight, 2 track under arc lighting, 3 image of 1000w incandescent bulb on the wall @3000 lps

Thanks for the pics. That helps a lot!

It is no surprise that, at least at this image size, I can't see the tires clearly, only the rims. I presume you can lighten the image as required.

A well-focused laser with a quality cylindrical lens would only illuminate a band around 1/8" wide or less at 30 feet, so the energy is well concentrated. The lasers that builders use to draw lines on walls etc. are only a few milliWatts. Some of them use a spinning mirror, which would not work for you (I presume), but some do use the cylindrical lens. The laser would have to be mounted on the same tripod/assembly as the camera, and the two would have to be correctly aligned, or you would get nothing.

Do try putting a red filter over the lens, with your current setup. It just might work, and is certainly cheaper than a laser.