Advice to even out illumination on DIY diffuser head

[madumi]

I am working on a diffuser head to capture film negatives using a DSLR & wondered if I could obtain suggestions for improving the head design...

My limitations are these: Presently it needs to be lit with a xenon flash tube (doughnut shaped)... I'll correct for the orange mask using Rosco Calcolor color correction gels

So far, I've tried a 45 degree styrofoam reflecting head design (similar to http://www.largeformatphotography.in...-Light-Sources)... Sadly the illumination is 1-2 out of 255 RGB values brighter in the centre. The edges are about 10 RGB values darker.

I tried stacking clear prismatic acrylic sheets to even out the illumination (stacked simply, but with no sides to reflect the light back). This turned out predictably worse.

Most recently, I'm in the process of obtaining 98% reflectivity aluminum to line the sides of a square tube & I'll play around with opaque acrylic & clear prismatic acrylic to see if it will even out the illumination.

I ruled out using a collimating head design because of the doughnut shaped flash tube in the strobe, & I'd prefer not to resort to LED's as I'm leery of obtaining an even spectral output (I just don't know enough).

Any other suggestions for designs to try/include to obtain evenness in illumination?

Thanks so much!

[Michael Mutmansky]

I am working on a diffuser head to capture film negatives using a DSLR & wondered if I could obtain suggestions for improving the head design...

My limitations are these: Presently it needs to be lit with a xenon flash tube (doughnut shaped)... I'll correct for the orange mask using Rosco Calcolor color correction gels

So far, I've tried a 45 degree styrofoam reflecting head design (similar to http://www.largeformatphotography.in...-Light-Sources)... Sadly the illumination is 1-2 out of 255 RGB values brighter in the centre. The edges are about 10 RGB values darker.

I tried stacking clear prismatic acrylic sheets to even out the illumination (stacked simply, but with no sides to reflect the light back). This turned out predictably worse.

Most recently, I'm in the process of obtaining 98% reflectivity aluminum to line the sides of a square tube & I'll play around with opaque acrylic & clear prismatic acrylic to see if it will even out the illumination.

I ruled out using a collimating head design because of the doughnut shaped flash tube in the strobe, & I'd prefer not to resort to LED's as I'm leery of obtaining an even spectral output (I just don't know enough).

Any other suggestions for designs to try/include to obtain evenness in illumination?

Thanks so much!

The easiest way to do this is to make the integration chamber and the light source surface (acrylic sheet) considerably larger than the minimum needed to cover the film. The flaw I see with the approach in the photos is that the emitting surface is not very large, and the chamber behind it is much, much too small.

If the integration chamber were 2x the size, the light that would get out of the emitting surface (at the middle!) will be much more uniform.

Note also that you don't need 98% reflective materials. What you need are materials that are as lambertian as possible.

Here's what happens (even in a perfectly uniformly reflective chamber). A light source is put in, and the light bounces around inside. but, the corners of the sides "see" less of the rest of the inside of the box because they are 'crammed into the corner', so to speak. The place where two sides come together effectively only 'see' 1/2 of the box well. The center of each side sees more of the box, and therefore has higher illuminance. This goes doubly bad for the very corners of the box, because they 'see' only about 1/4 of the box well.

So, to reduce that, make the light emitting surface a much smaller portion of one of the sides. In the diagram that you linked to, I wouldn't use more than about the center 1/3 of the bright spot of the light emitting surface.Mas off the rest.

This is unlike a lot of the traditional enlarger heads because they rarely had the room to make a huge integration chamber, so they made them as small as they could and still do a reasonable job.

Why do people burn the edges of a print? Three reasons; light source falloff, lens falloff, and a general preference for it visually. The light source falloff is considerable when you get to the edges of the field with a lot of enlargers...

It may be a bit hard to understand without a diagram, but drawing it up makes it pretty clear.

Second, once you have produces a light source that is reasonably uniform, you can mitigate a hotspot by cutting a piece of frosted mylar or tissue paper (bumwad) to fix in the middle. It will be cut into an oval or circular shape most likely. You'll have to do testing to figure out how much, what size, and the number of layers that may be required. It may affect the CT of the light, unless you get a neutral material, though.

I suspect that if you build the integration chamber correctly (large relative to the size of the area you want for the film), you will not need to make any adjustments to the light emitting surface.


---Michael

[Peter De Smidt]

That was an early design for basic feasibility testing. I agree with Michael that making the integrating chamber much bigger is the simplest solution. You might want to check into the output spectrum of Xenon flash. It might not be as continuous as one would expect, and flash-to-flash variation can be an issue. Adding more diffusers is unlikely to help.

You could've added these questions to the existing DIY lightsouce thread.

[Larry Gebhardt]

Sorry I can't offer any tricks to make the output even. However I must ask why you don't switch to another light source? I've used a color head for DSLR scanning. You get even output and you can dial out the orange mask at the same time. Also a standard flash unit from a few feet back, through a closer diffuser, will also provide reasonably even illumination.

[madumi]

Thanks for replies so far.

@ Peter, OK, I'll post to the DIY lightsource thread once I get my prototypes up and running--I'm just looking for suggestions which way to take the design for now...

@ Michael, Thanks for the input... If I designed a unit that used reflection at the source of the backlight for the negative, a lambertian matt reflector would work well, I'll stick with transmissive sheets for now though.

@ Larry, the only reasons I'm avoiding dichroic enlarger heads is because I figured the illumination would slow down the shutter speed (shooting at 100 ISO), and my preference would be to create a digital color profile to guide the conversion of the film types. If I fiddled with the color balance for each shot, that would mess with the accuracy of the color profile. Gels can bring the color balance within about 7% density--hoping that proves close enough...

One idea that has been running through my mind to even out the edge illumination would be to use illumination from the diagonal & block the illumination directly behind the final opaque acrylic. With the inverse square law, varying the angle & distance of the light from the sides should give some flexibility on "countering" the edge darkness... Maybe something similar to a ring flash for the light source--with the light projecting onto a circular acrylic disk below it...

Any other suggestions?

[Jim Jones]

My 5x7 Elwood incandescent diffusion enlarger had uneven illumination. Tracing paper and a little work with graphite pencils made a mask to lay on the diffusion plate that worked well.

[Peter De Smidt]

In my experience, making the light source significantly bigger than the negative helps. My current source's area of illumination is 6x8". It's great up to 5x7". With my source, I found that putting reflective material on the outside walls did help with edge fall-off. As Jim says, you can make a mask to even out the source, possibly even using film to expose and develop a mask, or using an injet printer with transparency film. After you get the light source as even as you can with construction, you can use software, if needed to even it out even more. See: http://blog.phaseone.com/scan-negati...ure-one-pro-8/

With my De Vere 504 enlarger, the acrylic diffuser is cnc'd to be thicker in the middle.

Oh, and supposedly sintered teflon reflects light very evenly. I don't know the availability or cost. I just found out about it recently.

[madumi]

@ Peter, Thx, good to know about the sintered teflon... For what it's worth, for high efficiency (97% reflectivity), Anomet also has an interesting film that can be applied (adhesive backed--not good for high temperatures though)... Here's a link for interest's sake: http://anomet.com/cgi-bin/online/storepro.php

[Peter De Smidt]

That's good to know, Bert. Thanks.

[Drew Wiley]

Getting your diffuser even can take quite a bit of work. You can double diffuse with two sheets of translucent white acrylic with space between them, like Durst does with halogen point light sources, or grind a thicker piece to match the characteristics of the mirror box, which takes a tad of gear and practice. I don't know how hot your light gets, but assume you've got some way of cooling the head. There are also some more exotic ways to do this which I won't try to explain here.
But bagging an efficient even light box on your first try is pretty difficult. Usually there is some prototyping involved with consequent refinement. Even formal
manufacturers of colorheads have come out with mass-produced lemons from time to time. In terms of assembly, high temp tapes and caulks can be found, but
you want to surf an appropriate industrial dealer like McMaster for these kinds of things, not your local hardware store. All kinds of high-temp products are used by aerospace and hi-tech industries, so you have to think in terms of their potential suppliers of odds n' ends.