using blue/green led arrays as an alternative to cold light

[vedmak]

I was wondering if anyone tried to use an array of super bright LEDs to print and could you share you experiences, tips, etc?

 

[guitstik]

Search: (there was a url link here which no longer exists)
(there was a url link here which no longer exists)

The search feature is handy.

 

[rmann]

Why not white and VC filters - much easier to assemble - do a google search for fish tank lights - there are a lot of people working on LED set ups for fish tanks and a few companies sell DIY kits.

 

[domaz]

Why not white and VC filters - much easier to assemble - do a google search for fish tank lights - there are a lot of people working on LED set ups for fish tanks and a few companies sell DIY kits.

You will cut output using filters. That said if you used a bunch of five-watt white LEDs it probably won't matter much. Using filters with the typical LED lighting kit is probably going to result in bad printing times though.

 

[onepuff]

You may well encounter problems using LEDs with VC paper as they are not "full spectrum" lamps. Each chip emits light at a specific wavelength which when pulsed by the controller to vary it's brightness and is mixed with the output from other chip(s) gives light at a given colour temperature. Unlike full spectrum conventional incandescent or TH lamps they do have significant gaps in their spectrum so will not render correct exposure results when using photographic paper. In this respect they present similar limitations to HID lamps. That isn't to say you can't experiment!

 

[rmann]

You may well encounter problems using LEDs with VC paper as they are not "full spectrum" lamps. Each chip emits light at a specific wavelength which when pulsed by the controller to vary it's brightness and is mixed with the output from other chip(s) gives light at a given colour temperature. Unlike full spectrum conventional incandescent or TH lamps they do have significant gaps in their spectrum so will not render correct exposure results when using photographic paper. In this respect they present similar limitations to HID lamps. That isn't to say you can't experiment!

Most LED manufacturers have web sites where they give the spectrum data - you can select a white LED that will give you both the green and blue you need for VC filters. Getting a set of filters is a whole lot cheaper and easier than building the control & timing circuits you would need to work with both a blue array and a green one - unless you are going to do only split printing. If you search the treads, I think you will find a number of attempts at going with blue / green LEDs. Not all of them end with a working unit.

It probably depends on your skill with circuit design and the amount of time you are going to invest which way you end up going.

 

[rmann]

Check the fish tank guys - they do very bright as they want to keep coral alive in their tanks without adding extra heat - they want it to be like sunlight. Their DIY kits are not the typical kind that are used for household lighting.

 

[konakoa]

I've made two homemade LED heads for large format 4x5 film using white LEDs. They work well and compare favorably to a Aristo cold light I have. I've got postings here on APUG about the LED heads I made and a website with how-to instructions.

Green and blue LEDs sound great in theory but don't work out so well in practice. I bought several discrete green and blue LEDs and tested them for VC work. Basically, what I found is green LEDs work great for low contrast and don't have any issues with VC paper. However, there are problems with blue LEDs. To the eye, they're an intense blue color--but the hitch is they're actually a combination of blue and green.

The slight green color hidden within the blue LED spectrum reacts with the low-contrast emulsion in VC paper and limits the VC papers I tried to about a grade 3 to 3 1/2. Hard contrast grades like 4 or 5 just can't be done with blue LEDs alone unless the green bias in the blue LEDs is filtered out somehow.

It's worth mentioning as well that the green and blue sensitive emulsions in VC paper do not have an equal response to green and blue light. Blue prints and builds up density very quickly and green takes much more intensity and time to reach a similar density. If you change the green to blue ratio even slightly, the the density will vary greatly. Be prepared to go through a lot of test strips and work prints.

I did a lot of step wedge testing in the darkroom with several LED colors and even made a spectrometer to study the discrete LEDs to come to these conclusions.

White LEDs with a VC filter set are much, much simpler to work with.

 

[ic-racer]

Why not white and VC filters - much easier to assemble - do a google search for fish tank lights - there are a lot of people working on LED set ups for fish tanks and a few companies sell DIY kits.

Or even better, a 120v halogen lamp and non-fading dichroic filters

 

[hrst]

You may well encounter problems using LEDs with VC paper as they are not "full spectrum" lamps.

Why would this cause problems? Please elaborate.

Silver halide are sensitized to two wavelength regions, blue and green, and you need to have those, and you will most probably have those. If you mean that sensitization dye absorption spectra would mismatch with LED spectra, I really doubt that. Neither are that sharp.

Using blue/green led arrays is naturally the easiest way compared to white-light source and dichroic filters. Mechanics are ruled out, it's electrically controlled, all you need to do is mix the light but that will be as easy or easier than in dichroic filter heads.

In fact it may work well enough even without mixing/diffusing but I would make sure it is mixed so that blue and green comes from same (all) directions.

Furthermore, with LED arrays it is possible to adjust separately exposure and also contrast at different locations! I have been thinking of for example 8x8 matrix for 6x6 film. This would work additionally to manual dodge&burn.

vedmak, have a go at it and report back how it works!

 

[onepuff]

Why would this cause problems? Please elaborate.

Silver halide are sensitized to two wavelength regions, blue and green, and you need to have those, and you will most probably have those. If you mean that sensitization dye absorption spectra would mismatch with LED spectra, I really doubt that. Neither are that sharp.

Using blue/green led arrays is naturally the easiest way compared to white-light source and dichroic filters. Mechanics are ruled out, it's electrically controlled, all you need to do is mix the light but that will be as easy or easier than in dichroic filter heads.

In fact it may work well enough even without mixing/diffusing but I would make sure it is mixed so that blue and green comes from same (all) directions.

Furthermore, with LED arrays it is possible to adjust separately exposure and also contrast at different locations! I have been thinking of for example 8x8 matrix for 6x6 film. This would work additionally to manual dodge&burn.

vedmak, have a go at it and report back how it works!

Discrete single chip LEDs emit light at a single wavelength while "white" LEDs contain 3 chips (usually chip on board these days) which when combined give white light at a specific colour temperature depending on the emitted wavelengths of the chips used and how much each chip is driven. This allows the colour temperature to be tuned. Colour changing LEDs use 3 chips also in red green and blue and are pulsed on and off by a controller to vary the brightness of each of the chips to mix the desired colour.

There are two problems when using LED light sources for photographic printing. The first is That the actual emitted wavelength from single colour LEDs is precisely that - one wavelength. Incandescent and halogen lamps produce light at many wavelengths and the better lamps have a flatter spectral curve giving a higher CRI. Filtration is used in enlargers to allow a band of a number of wavelengths of this spectrum to pass and photographic papers have a spectral response curve which will vary depending on the part of the spectrum received. VC papers produce differing contrast dependant on the part of the spectrum they receive. LEDs only produce a very narrow part of the spectrum so the paper's response will only be to this narrow part of the spectrum when exposed to light from an LED. This means that you will not get the same response from the paper as you would expect from a light source with a wider spectral curve combined with filtration with a wider band pass. The paper will only give the response within its spectral curve to the single wavelength it receives. This leads us to the second problem - colour mixing. If you have two un-mixed discrete single colour LEDs they will produce 2 wavelengths so the paper will respond to both of these and give a combination of the contrast you would expect from both wavelengths within its response curve. To get a single new wavelength of light from the combination the light needs to be collected via a lens system (the design would depend on the emission pattern from each chip) and needs to be mixed within a mixing rod or tube. To vary the contrast of the paper however you need to be able to tune the colour of the light (as you would using filters). To do this with LEDs means pulsing the LEDs to vary the brightness and this consequently varies the exposure time required. It is like trying to expose with a lamp which flickers at a varying rate. To overcome this difficulty it may be possible to build a separate LED module for each desired colour of light but this would be expensive or you could measure the output from the mixer for each desired wavelength and vary the exposure to suit.

I hope I have clarified what I believe are the difficulties of using LEDs as a light source for printing. I believe it is much easier to start with a broad spectrum lamp then use known filters to achieve the same result. As I said though - it does no harm to experiment.

 

[Lee L]

Here is konakoa's page on LED enlarger heads, with examples of prints made with a blue/green LED light source.

http://www.deadbread.com/crumbs/23c.html

Lee

 

[Steve Smith]

I have successfully converted a DeVere 5x4 enlarger to LED using the RGB PCB from an LED stage light like this: http://www.soundandlights.eu/large_pictures/LED_Par_64.jpg

Using just the blue LEDs will give grade 3.5, possibly 4. Someone here has managed grade 5 by including some UV LEDs into his design.

EDIT: The 'someone' I am referring to is in the link Lee posted above).


Here is my initial post: (there was a url link here which no longer exists)

And here is a post with some sample images: (there was a url link here which no longer exists)

There are two problems when using LED light sources for photographic printing.

I believe that results trump theory every time.

To do this with LEDs means pulsing the LEDs to vary the brightness and this consequently varies the exposure time required. It is like trying to expose with a lamp which flickers at a varying rate.

Pulse Width Modulation is a very common way of controlling LEDs and is what I intend to do with mine. The frequency can be in the tens of kHz and isn't really the same as flickering. However, limiting the current with a series resistor will work just as well. No need to over-complicate matters.

it does no harm to experiment.

Indeed!


Steve.

 

[RH Designs]

Most blue LEDs are not "blue enough" to match the sensitivity of the hard emulsion on VC paper. That said, there are LEDs available with the right spectrum. If you choose the LEDs carefully you can achieve a contrast range much wider than that available using white light and filters. I have obtained an ISO(R) range of 45-200 from Ilford Multigrade 4 RC using just blue and green LEDs of the right wavelength - this is quite a lot wider than with under-lens or dichroic filters. You don't need UV LEDs, just the correct wavelength blue ones. The bad news at present is that LEDs of that wavelength are not available in high-power versions, so you need rather a lot of them!

Onepuff: it is not true that LEDS emit only a single wavelength, a glance at the data sheet of any LED will show you that. Also, many white LEDs use a phosphor illuminated by a blue chip to emit white light, they are not a combination of RGB chips.

 

[Steve Smith]

Also, many white LEDs use a phosphor illuminated by a blue chip to emit white light, they are not a combination of RGB chips.

As a user of many white LEDs (at work) I would say that most white LEDs are modified blue types.

Have a look here: http://www.designledproducts.com/ to see the clever sort of things being done with LED lighting now (I'm not giving away any secrets though as we are subject to a confidentiality agreement).

Whilst white LEDs might not have a full, even spectrum, they are full enough that we can filter them with translucent coloured inks to give red, green, blue, orange and violet (see applications in the link above).

One of the salesmen/applications engineers from the company I linked to is also a photographer who has recently bought himself an enlarger. I know that he has also become a member of APUG. Perhaps if he sees this, he could add some further information.


Steve.

 

[Mainecoonmaniac]

Does any old timers remember the Minolta 45A? It uses xenon tubes instead of LEDs. I'm wondering varying the exposure of the different colored LEDs to change contrast?

Dead Link Removed

 

[hrst]

Discrete single chip LEDs emit light at a single wavelength . . . That the actual emitted wavelength from single colour LEDs is precisely that - one wavelength.

No. LEDs are far from monochromatic. For example, take a look at LED datasheets. They show the spectra. Typically, the bandwidth measured at 0.5*intensity is something around 20 - 70 nm. This is far from "one wavelength".

Furthermore, sensitizing dyes are not sensitive only to one wavelength - they also have a broader sensitivity spectrum.

This is why I don't believe the mismatch would be a real problem.

If the wavelengths (LED emit maximum vs. sensitizing dye absorption maximum) do not coincidence perfectly, it shouldn't cause any harm, just some loss in sensitivity, thus need for more light as some of it is wasted. If the mismatch is severe, it would prohibit the extreme contrast grades (00 and 5, for example), but I think this won't be a problem. The papers are anyhow designed for tungsten light sources that are very low in low-wavelength blues, thus I'd guess the sensitization goes to cyanish blue. And, I would give an educated guess that the green used in power LEDs is spot-on or very near.

 

[konakoa]

konakoa,
Can you please enlighten me a bit regarding "white" LEDs? What is the color temp of "white"? The reason I ask is that after reading Ilford's website info on the lightsource for their VC papers, they say that their filters will work OK with any incandescent source. Incandescents are about +/- 2700 deg-K. Incandescents are also going to become harder and harder to find in the near future, or so I hear from retailers who are stocked to the ceilings with CF bulbs as the current trend.

Silveror0, the white LEDs I used are rated at 4500K. The reason I choose that version was that they were the brightest available of all the white LEDs. It takes a lot of LEDs to get the equivalent brightness of a single fluorescent or tungsten enlarger lamp.

4500K does work well and I can print at all of the contrast grades with it. Yet as you noted VC filters are designed with much warmer tungsten lamps in mind. The effect I've noted is that VC paper prints to a higher grade by default with the white LEDs I used (4500K is cooler/more blue). A print in my condenser enlarger with no filter requires a #1 or a #1/2 VC filter in the LED enlarger to get the same contrast grade.

Coincidentally, the 4500K LEDs I choose are a near perfect match grade for grade for the Aristo V54 fluorescent lamp I have. Even though one is white to the eye and the other a blue-green, the filters print the same contrast ranges on each grade despite the distinctly different lamps. I thought it was rather interesting. This was with Ilford RC paper.

 

[alanrockwood]

I wish there were spectral response data available for available variable contrast papers. Unfortunately, that data does not seem to be available anywhere.

Alan

 

[RH Designs]

Ilford publishes the spectral response curve for Multigrade IV papers in its data sheet, here.

 

[hrst]

Ilford publishes the spectral response curve for Multigrade IV papers in its data sheet, here.

Unfortunately, the curves for B and G sensitive parts have been combined to create a "total" spectral response curve.

However, we can deduce that 500-520 nm for green should be spot-on. Do you agree?

Typical cheap power RGB led has green at about 520 nm and blue at about 460 nm (an example: http://www.satisled.com/3w-rgb-led-emitter-on-star-1w-for-each-color_p234.html ). However as I have shown above, the bandwidths are not that sharp, thus the G and B in typical RGB leds almost overlap.

AFAIK, blue exposure causes both blue- and green-sensitive layers to react in VC papers (in the other words, they have only-blue-sensitive and blue-and-green-sensitive layers). Thus, if our green is "green enough" not to expose the blue layer, and our blue is "blue enough" to expose the blue layer, it should work and give us all contrast grades from 00 to 5. I said earlier as an educated guess that the green in typical leds is probably spot-on, and this seems to be backed up by the evidence.

 

[Steve Smith]

AFAIK, blue exposure causes both blue- and green-sensitive layers to react in VC papers (in the other words, they have only-blue-sensitive and blue-and-green-sensitive layers).

It's the other way round.

All of the emulsions are sensitive to blue and some are sensitive to blue and green.

http://www.ilfordphoto.com/Webfiles/2010628932591755.pdf


Steve.

 

[Curt]

Or even better, a 120v halogen lamp and non-fading dichroic filters

Or even better one to three 120volt, correct wattage, correct color temperature halogen bulbs behind either RBG or CMY Dichroic filters that project into a Styrofoam lined tube with two Plexiglas diffusers.

Each "channel" can be use independently as in split grade or VC paper application or all three can be used for focus or white light with VC filters under the lens.

Run it off a CVS and put a fan in a box and run the hose to the head with a light tight baffle on the air intake side.

The possibility of industrial or commercial dimmers is a possibility also.

I'm nearly finished with this type of head for my Durst 138 enlarger. It's a diffusion head with a cooling fan as compared to a cold light diffusion head using a tube or tubes. I'm making a 5x7 head right now and will visit an 8x10 head a little later.

For my negative carrier; I really never liked glass; I'm using a "Zipper" glassless carrier. It mounts a 5x7 negative without scratching the hell out of the edges like the Beseler Negaflat does. I can also put a 4X5 negative in at 90 degrees and use the enlargers masking system, a great feature of the Durst enlargers.

For 8x10 I'd make one just like the Zipper carrier.

Best,

Curt "the 'ive had it with out of date bulbs, heads, lights and lamps that are problematic."

What kind of light source did that fellow in California by the sea use on his big converted studio camera enlarger? Old technology? So what? It worked pretty well for him and he had a nice Stetson hat to boot.

 

[alanrockwood]

Unfortunately, the curves for B and G sensitive parts have been combined to create a "total" spectral response curve.

However, we can deduce that 500-520 nm for green should be spot-on. Do you agree?

Typical cheap power RGB led has green at about 520 nm and blue at about 460 nm (an example: http://www.satisled.com/3w-rgb-led-emitter-on-star-1w-for-each-color_p234.html ). However as I have shown above, the bandwidths are not that sharp, thus the G and B in typical RGB leds almost overlap.

AFAIK, blue exposure causes both blue- and green-sensitive layers to react in VC papers (in the other words, they have only-blue-sensitive and blue-and-green-sensitive layers). Thus, if our green is "green enough" not to expose the blue layer, and our blue is "blue enough" to expose the blue layer, it should work and give us all contrast grades from 00 to 5. I said earlier as an educated guess that the green in typical leds is probably spot-on, and this seems to be backed up by the evidence.

hrst said what I was trying to say but didn't quite say it right, or to put it a different way, the spectral response curves and contrast curves for the two layers are not available separately.

As to the choice of green LED, you should probably use an LED with a center wavelength closer to 530 nm to be sure to not be in the toe of the blue curve while at the same time being fairly close to the peak sensitivity for the green layer. As to the blue LED, it is hard to say from the combined graph. However, I think an LED with a center wavelength of roughly 425 might be a good guess, or maybe even something close to 415. This would likely keep you away from the toe of the green curve while being close to the peak intensity of the blue curve.

 

[alanrockwood]

On second thought, taking a ruler to the paper and measuring the vertical distance between the curves, I think a blue LED of about 430 nm might be a little better than the values I gave earlier. This is about where there is a minimum vertical distance between the curves. This assumes that the curves are drawn fairly accurately, a possibly dubious assumption. However, if you had to shade the wavelength one way or another it should probably be toward 420 nm in order to stay away from the toe of the green emulsion.

For the green curve I think 520 nm might be pretty good, but if you had to shade that value one way or the other I would shade it more toward 530 in order to stay away from the toe of the blue emulsion.