The Effects of Reflector Design and Lamp Orientation

[18wheeler]

Quoted from this website

The Effects of Reflector Design and Lamp Orientation


The orientation of the lamp and the design of the reflector influence the performance characteristics of the light source in the following ways:

Optical Power

The obvious way to increase the optical output of a light source is to employ a bigger lamp. A higher power lamp, however, will not necessarily result in higher optical output from a light source. Since light is emitted by the lamp in all directions, the important thing is not the amount of light emitted, but the amount of light collected.

The A-1010B and the A-6000 have the highest optical power output because they have the highest collection efficiency. They collect and focus emitted light with a single optical element: an ellipsoidal reflector. Because the ellipsoid encloses a large portion of the solid angle about the arc lamp, over 65% of the lamp's emission is collected and focussed at the output.

Compare this to the typical performance of a traditional lamp housing design, such as that used in the A-500, where roughly 10% of the light is collected. Using identical lamps, the total optical output of the A-1010B is about seven times higher than the A-500. To get the same optical output as a A-1010B with a 100 watt lamp, a traditional housing would require a 1000 watt lamp!



An ellipsoid has two focal points. With the arc lamp precisely positioned at one focal point, the collected light is focussed at the second focal point. In your experiment, the sample, the monochromator slit or the fiber optic bundle would be positioned at the second focal point.

Power Density

For many experiments involving the illumination of a relatively small area, the power density of a light source is a critical specification. Optical power density is a function of the output of the lamp and the size of the area illuminated. Note, however, that increasing the lamp power does not necessarily increase the power density; in fact, it can actually reduce it. Assuming the optics are the same, the maximum practical power density is a function of arc size.

For example, a 75 watt lamp has a luminous area of 0.2 by 0.5 mm, or 0.1 sq. mm. A typical 450 watt lamp has a luminous area of 0.9 by 2.7 mm, or 2.43 sq. mm. The illuminated area from the 450 watt lamp is 24 times greater than the 75 watt lamp, but it has only six times more power. The power density of the 75 watt lamp, then, is four times higher than that of the 450 watt lamp.

Stability

The interior of an arc lamp housing in operation can exceed 150 degrees celcius. Thermal distortion of the reflector is of concern, since it is the only optical element in the system. Deformation of the ellipsoid shape can dramatically reduce the useable output of the illuminator. The proprietary, thermally-matched optics of the A-1010B and A-6000 guarantee that the shape of the ellipsoidal reflector remains accurate during operation.

Uniformity

Neither a vertical nor a horizontal arc lamp produces perfectly uniform illumination. The majority of the light is produced near the cathode. When mounted vertically, the illumination between the electrodes is quite uneven from top to bottom. In the horizontal position, convection currents within the lamp cause slight arc wander. We have found that the ellipsoidal reflector produces the most consistent intensity at the output when the lamp is operated in the horizontal position.

[Guy Grotke]

That 10% efficient spherical reflector design in the picture is nothing like DIYers are using! We use a much larger condensor lens almost touching the lamp outer envelope, and a much closer reflector sized to match the condensor lens. The result is the capture of around 65% of the light coming out of the lamp arc. Which just happens to be the amount they claimed for their elliptical reflector!

There are some very nice things about elliptical reflectors, so there is no need to create a "straw man" spherical example.

1) The reflections don't go back through the lamp, so these reflectors are inherently "cold".

2) The light all gets focussed at a point outside the reflector, so it is very easy to insert dichroic filtering there.

3) No need for a heat-resistant condensor lens.

4) You can increase the efficiency even higher by adding a large spherical reflector to recycle rays that missed the elliptical reflector. (Light output goes through a hole in the middle.) This design could capture >95% of the light, but some of that would be attenuated by absorbtion at the reflectors.

[ace3000_1]

Read abit closer guy lol

Because the ellipsoid encloses a large portion of the solid angle about the arc lamp, over 65% of the lamp's emission is collected and focussed at the output.

http://www.thestagecrew.com/Pages/Ch...eflectors.html

I would have to say, with mine its improved my lamp by 75% +, nothing like it was ever on the spherical, and ive got 5 diff types fyi. Not trying to start a war here, but hands down the ellipsidal wins the race by a long shot, (more then what i thought myself until i saw it).


Trev

[Guy Grotke]

I think we can all agree that you would get less light captured without a condensor lens (or else why bother using one?). So here is the calculation of the percentage of light captured by a 15" LCD with the correct spherical reflector, but without a condensor lens:

The light going from a point source lamp arc directly to half of a 220 mm fl condensor fresnel forms a triangle with a 220 mm base and a 7.5" height. By Pythagorean Theorum, the hypotenuse is 291 mm. Taking the arcsine of (7.5" / 291 mm), we get a half angle of 40.9 degrees. So the total arc of light falling on the 15" fresnel is 81.8 degrees. This is 81.8/360 = 22.7 % of the light coming out of the lamp. With a 90% reflective aluminum spherical reflector behind the lamp, the fresnel will also get another (22.7% * 0.9) = 20.45% of the lamp's light. So the combination of NO condensor lens and a run-of-the-mill aluminum spherical reflector will capture over 43% of the light coming out of the lamp.

A 65% efficient elliptical reflector IS better, but it's not 6.5 times better. I was just trying to point out that the original article created an unrealistically bad spherical reflector & condensor example, so they could hype their elliptical.

[ace3000_1]

This spherical will capture close to 42%, forget the condenser. Take a note of how much light you loose on the sides.

You can never get 50% from our bulbs Guy on this reflector, why? well unless the focal is set at the top of the spherical (severe abberations btw at that point) and the bulbs is arc is dead level with the top you will never get the 50%. Our bulbs are simply to big to place there, and no spherical is cut at a half sphere simply because of severe spherical abberations.

Bulb radiation patterns also play a huge role, and with our setups you loose alot of light from the spherical type system.

Trev

[ace3000_1]

This one will reflect even less light, but has less spherical abberations.

Trev

[ace3000_1]

Take a look at this, ive even included the beam angle for you. the ellipsidal that i have directs all of the intensity from my bulbs beam angle to where i want it to go, its not lost like in the spherical design.

If you wer here, i could clearly show you the difference on the screen for your self. The 150w bulb is way brighter then the 250w running on a spherical, even with a condenser. Its only until you condense the arc the 250w catches up and they run close to even, and even then the 150w is still brighter.

I dont use a condenser on my ellipsidal, why? cos all of the light is going to where i need it, its that simple. I can prove this if you want with a pic of my light on. Its clearly visible. From memory you even drew it out on paper just as i did before i bought it and it all matched up. Why condense the beam more if its spot on where it is? you will end up with a bright center on the projected image.

The only gain you would get with a condenser in this setup is the heh, roughly 5% of light you do loose, its that small its not worth the gain. Hell i hardly see any light on the table other then where it should go, and thats why its so bright, its going where i want it. And thats the key.

Trev

[TaskMaster]

So if you couldn't match the ellipsoidal and the fresnel you could incorporate a condensor to match the fresnel?

I've read posts ststing different places to place the condensor in that configuration. Does it go before the ellipsoidal's second focal point, at it, or after it? On the pic below would it go at A, B, or C?

Thanks

[ace3000_1]

C if you wanted it wider. You can control the beam with 2 pcx lenses, but the lcd realy has to be the same size as the reflector for that to work efectivly.

Take a look at stage lights and how they work with em to give you some ideas


Trev

[Me2!]

Yes Guy, there is no doubt they padded their marketing blurb for sales purposes. You can see the puny angle on the spherical reflector in comparison to the angle they display on the ellipsoidal.
I totally agree with the premise of your arguement.

I would argue however that your math is asking a physical home made system to produce theoretical efficiency. I dont like sphericals in that they send the reflection back through the arc. Remember that the bulb is 2 jackets of curved glass and going back into the arc will likely be deflected somewhat on its way in. Its unreasonable to assume the reflector will be equidistant at all points from the arc. The reflector was not aligned in a special jig - probably with adjustment screws, ruler and eyeball! The jackets are not perfect spheres either. Close enough for light going out but not for the return in. I expect light going straight to the fesnel to be slightly misaligned with the distorted reflected light from a sphere. Thats the real problem because some of those waves will cancel in an out of synch interference as you know. I would be worried about lighter/ darker patches on screen. Remember the rings when you tested the light tunnel to help someone out?

Ellipsoids have the advantage that the beams dont cross in unexpected ways so you dont need to worry about these things. I know your thinking that when the light opens up to the width of the LCD you dont need to worry about it. Considering the path its gone and how close to parallel the beams should be, I (personally) wouldn be sure they arnt meeting at unknown points.

regards,
Arthur