Building a Variable Contrast LED head for a Durst 138

[Larry Gebhardt]

I got the sheet of Makrolon LD from Randy yesterday and cut it to fit my diffusion frame today. Compared to the 1/16th" acrylic it is 1.3 stops dimmer. Until I saw the Makrolon I thought the acrylic was doing an excellent job with diffusion. I've never had an issue with unevenness, but if I look closely I can see very slight areas where the green and blue are not as blended together as they could be. The Makrolon looks perfectly even, and given I usually have a ND filter over the lens I think I'll start using it permanently, though it will make large prints a bit slower.

Thanks Randy!

I'd still like to test out a thinner sheet of Makrolon if it exists.

[Tin Can]

Any plastics shop can cut you a piece of diffusion acrylic. What is important is not to order plain"white" plexiglas or whatever, because it will be too dense, but what they generally market as "sign white", because it will be translucent, to a degree obviously dependent upon the actual thickness. Thicker plastic can hypothetically be ground down at the edges and corners to match any illumination falloff due to your specific system, lens, etc, though there are a few tricks to doing that. Otherwise, concerning another question, Chinese LED arrays tend to be temperamental due to fragile wiring. They don't seem to last very long.

Look at this, LIGHT BALANCING DIFFUSION DISKS, ETC. Scroll down.

[Larry Gebhardt]

That's pretty slick to machine a piece of diffusion plastic to convert condensers. I wonder how well it works.

I cut a small piece of the Makrolon and compared it to the opal glass I use when using the condensers. The Makrolon is .9 stops less bright, and really no more even. I was hoping to be able to get the LEDs closer to the diffuser to get more light, but it doesn't look that way.

[Tin Can]

That's pretty slick to machine a piece of diffusion plastic to convert condensers. I wonder how well it works.

I cut a small piece of the Makrolon and compared it to the opal glass I use when using the condensers. The Makrolon is .9 stops less bright, and really no more even. I was hoping to be able to get the LEDs closer to the diffuser to get more light, but it doesn't look that way.

When I was using 900 bulb LED panels and Makrolon, I found 2 thicknesses was necessary and I turned my panel down to 10% power.

Then I went to the Artogragh.

[jbmia]

All practicality aside and in the spirit of experimentation, I have a few questions for Larry and others who have worked with LED's...

As I mentioned in another post, if I were to convert the Durst 138s head into a diffusion head, with a large mixing chamber, as is illustrated here: http://glennview.com/jpgs/durst/57/g...head/big_3.jpg , with a light source in the standard position, the mirror and condensers removed and lined with 1/2" styrofoam sheet, and an opal glass (or similar) in the base where the lower condenser usually sits, what lumen output (if that is the correct measure) would I need to achieve similar EV's to any of the compatible CLS heads? Researching a similar head, the Chromega E dichroic , which was widely used on these enlargers from what I understand, it takes an ELC 250 watt bulb, which I understand is approx 900 lumens. Given that's not the same design I'm suggesting and the size, distance to negative, and structure are different, should I shoot for something higher, say 1500+ to avoid building something under - powered? You can always dial down, but not up past max...

Given the XP-E royal blue, green, and red cree led's are lower output (max of around 515 lm), could I just arrange 4 in a tight group to achieve that EV range I'm looking for or do I need a single LED source with that output intensity to really achieve the EV I'm looking for? Say the CXA1520 (2k-4k lm's) or similar? The problem with these arrays is that they only seem to come in some variation of white with it's limitations (MG filters required)... Also, if I went with something like the CXA1520, what would be the best temperature, 3500k ? (As a side note, have you guys seen the CXA3590? 10K-15K lumens!! Wow.)

So, to wrap this up, Larry has built 3 different LED implementations.. one uses the latico condenser assemblies where lower power LED's are used up close to the negative and the others use either the bright white point source or blue/green LED's, but using the condensers (and maybe a diffusion layer in there somewhere)... I"m looking at something in the middle which would seem to approximate the more traditional mixing chamber, but using an LED source.. I was thinking this might provide for more uniform light distribution. The problem on the opposite side being the ability to achieve the required intensity...

What say you gentlemen.. LED wizards and lumen engineers:

What lumen output to achieve good EV's similar to the type of dichro heads that would normally be used here?
Can a group of lower power LED's achieve the same intensity as a single higher power LED?
If a higher power LED (only available in white), what color temp? 3500k?

[Larry Gebhardt]

I tried the approach you show in your linked photo. I wasn't able to get even illumination, which is why I went with the approaches I did. It also sucked up a lot of light. I'd say you want something in the 2000 lumen range for larger prints at a reasonable time.

Looking at the data sheet you would want the the 4000K version of the CXA1520, since it has strong peaks in the blue and the green. That should be well balance for use with VC filters or fixed grade papers. The warmer bulbs have much of their light in the red spectrum, which is useless for printing.

The CXA3590 looks pretty cool. I'm sure it would be bright enough, if you can get the chamber even enough.

If you want XPE green and blue I'd think you'd need at least 12 of each color, and probably twice that. Using the condensers are much more efficient, and placing them in the top condenser slot is also more efficient and probably easier than densely packing them to shine out the bulb window in the head.

[jbmia]

Thanks for the feedback Larry.. That's the kind of detail I was looking for.

It sounds like you're approach of placing the LED's in the condenser slot is probably the most sane approach at this point... The only thing holding me back on that route is figuring out a way to control the LED's using my existing RH Designs F Stop Timer... If I could figure that out it'd be a pretty easy build.

On the other hand, I'd be fine building a version of your controller/timer as well, but it's the Arduino code that's tough for me... Any updates on packaging that up for download like Mr. Tyrell with a BOM yet? No pressure. :-)

[Peter De Smidt]

Look at this, LIGHT BALANCING DIFFUSION DISKS, ETC. Scroll down.

De Vere was doing that, machining their diffusion plastic, at least 30 years ago.

[Larry Gebhardt]

Thanks for the feedback Larry.. That's the kind of detail I was looking for.

It sounds like you're approach of placing the LED's in the condenser slot is probably the most sane approach at this point... The only thing holding me back on that route is figuring out a way to control the LED's using my existing RH Designs F Stop Timer... If I could figure that out it'd be a pretty easy build.

On the other hand, I'd be fine building a version of your controller/timer as well, but it's the Arduino code that's tough for me... Any updates on packaging that up for download like Mr. Tyrell with a BOM yet? No pressure. :-)

I hope to have a set of circuit boards designed, and a BOM for several options developed at some point soon. Work is taking up most of my spare time. That and snow removal. But it will be done.

If you go with white LEDs and the right driver they will work with the RH Designs (or any) timer. But I'd rather have both fstop timing and variable contrast.

[Larry Kellogg]

It would be terrific if the LED head would work with an RH Designs timer!