RGB LEDs for color enlargers

[DREW WILEY]

Interesting stuff, that's for sure. But questions remain. First, how to scale up until you've got an LED colorhead with enough punch to compete with traditional commercial halogen heads for sake of large prints, and with the extra ooomph potentially needed for any supplemental registered mask density. Second, how reliable and easily replaceable are these components likely to be a decade or two or three down the line? - what will be the maintenance headaches? Third, what are real world optimized objective tests on available RA4 paper going to look like?

Of course, when it comes to old Durst 138 condenser heads, and any sizing and heat issues caused by cramming things into that confined space, it's fairly easy to adapt a significantly larger device atop the negative stage anyway. I never have used the old condenser head on any of my Durst units; it's still stored on a loft somewhere.

I'll be following all this out of curiosity, because it does seem LED's are the path forward for most. But at my age, I'm already well equipped for the duration with both additive and subtractive 8x10 and 5x7 halogen colorheads, which themselves took a lot of work to perfect. But those were designed or custom modified back in the Cibachrome days, which needed way way more lumen punch than RA4 papers.

 

[Lachlan Young]

First, how to scale up until you've got an LED colorhead with enough punch to compete with traditional commercial halogen heads for sake of large prints, and with the extra ooomph potentially needed for any supplemental registered mask density.

From hands-on with a current commercially available unit: seems level with 4-6x ELH, delivered within a 4x5 area with additive filtration (i.e. almost all the output in narrow bandpasses, no wasted output from a halogen running through a filter cutting off most of the output spectrum). Makes Fomatone less of a waiting game. No worse than an Ilford MG500 at the high contrast end - and you can stick a G5 under the mixed white-light option. Colour is mostly a question of understanding the how/ why of the control interface - and there's always the option of simply doing RGB sequentially, especially if you are using masks anyway. 4x5 is relatively affordable, anything bigger gets expensive really fast - still, a fraction of the cost of many dichroics in the past.

 

[koraks]

A little off topic, but I saw on your blog you had a durst print processor. Is that what you are currently using?

Yes, a lot of the time I use the rcp20. Sometimes for some quick prints I just use trays though; quicker to set up and put away when I'm done.

 

[DREW WILEY]

Lachlan - I use a pair of ELH's in my ordinary Durst CMY 8X10 colorhead; and that has plenty of punch for 30X40 inch RA4 prints. But it's overkill for typical black and white printing, which I generally use a high-output 12X12 V54 cold light for. Sequential RGB color printing is a horrible chore, though I do have the right kind of registration gear to do it, even from black and white color separation negatives. My true RGB SIMULTANEOUS 8X10 system involves some complex and sometimes cantankerous circuitry; but it's powerful enough to have done even 30X40 Cibachromes on a routine basis, yet far more efficiently in terms of heat and energy than my old 2000W Durst color mural head - better color too. I don't know how much more use it will get, however, since the standard CMY Durst head I'm using now is ample for large RA4 prints, and the color rendering is nearly as good. I also have a customized simultaneous RGB head on my 5x7 L138 Durst chassis, which I use for a lot of things, including precision internegatives, duplicates, and in relation to 4x5 and smaller film up to 20x24 print size.

I really can't see LED's replacing time tested halogen for big work any time soon. Besides, the high-production commercial labs have bought expensive laser printers anyway for large print RA4 purposes. Several varieties of those machines are still being made, so I hope Fuji doesn't pull the rug out from under them down the line by dropping RA4 papers. That's not likely to happen anytime soon, however. If I was younger I'd probably tinker with LED's too; but at this point in life, I just want to enjoy the equipment I've already seriously built.

 

[eli griggs]

@eli griggs that will probably work in combination with dichroic (subtractive) filters. This thread is about additive filtering using red, green and blue LEDs, however. There have been threads about replacing incandescent bulbs in enlargers with white LED lights; do a search on that, you may find something color-related as well. If not, maybe make a new thread on it.

So, using the built in RGB software for mixing lights, instead of just plain white lights is not a possible, even paired with a colour analyze designed for used with dicro heads, etc?

How about using these tools for B&W printing?

 

[koraks]

@eli griggs if it's an actual RGB source it may work. You mentioned adjustable white balance, so I assumed a white panel, not RGB. If you can give the specs of the product I could be more specific.

In principle, an RGB bulb with adjustable color will sort of work, although the wavelengths of the LEDs will be a poor match to the paper. You may not get particularly clean hues this way. The apps or remotes that come with these LED products will generally be a pain to use for printing and the filter resolution is likely insufficient for color work (only 8 bit).

Whether a color analyzer will work, depends. Some color analyzers don't work with PWM-controlled LEDs and since PWM is the usual way to control them, this creates a potential issue. I tried it with a Color star and it didn't work at all with 1.5kHz PWM frequency.

Of course, you don't need a color analyzer to print color, so this is not necessarily a big issue.

As always, give it a try and see what happens. I'm not overly optimistic.

 

[koraks]

@StevenPaul great, can you share some details of your setup? How do you control your LEDs? Which ones do you use?

 

[eli griggs]

@eli griggs if it's an actual RGB source it may work. You mentioned adjustable white balance, so I assumed a white panel, not RGB. If you can give the specs of the product I could be more specific.

In principle, an RGB bulb with adjustable color will sort of work, although the wavelengths of the LEDs will be a poor match to the paper. You may not get particularly clean hues this way. The apps or remotes that come with these LED products will generally be a pain to use for printing and the filter resolution is likely insufficient for color work (only 8 bit).

Whether a color analyzer will work, depends. Some color analyzers don't work with PWM-controlled LEDs and since PWM is the usual way to control them, this creates a potential issue. I tried it with a Color star and it didn't work at all with 1.5kHz PWM frequency.

Of course, you don't need a color analyzer to print color, so this is not necessarily a big issue.

As always, give it a try and see what happens. I'm not overly optimistic.

Yes, I am speaking to RGB LEDs.

Just Google "Lume Cube RGB Panel Pro" to read what it's about.

Cheers, all
Eli

 

[koraks]

Yes, I am speaking to RGB LEDs.

Just Google "Lume Cube RGB Panel Pro" to read what it's about.

OK, well, the Lume Cube specs are not as detailed as you'd hope for this purpose, but it looks hopeful.
I *guess* that they use five channels: R, G and B and two white channels (low and high K) to manage full-spectrum color plus adjustable white balance with high CRI.

Anyway, what's hopeful is that they imply 24 bit color depth and this should be ample for accurate filtering. From a practical viewpoint, the question is how convenient it is to adjust the 'filter' settings with their app; you're in a much better position to say something about this than I am.

It's also nice that it should be a fairly well-diffused panel, and if necessary, you could always fit a sheet of milky plexiglass over it.

The Achilles' heel of systems like these is usually that the blue LEDs are at 455nm, which isn't ideal for color printing, where a slightly shorter wavelength is better. How much of a problem this is, you'll have to find out by giving it a try.

So all considered, it should work. Is it the perfect solution? Likely not, but "in a pinch", as they say, it'll produce color prints, I'm sure.

 

[albada]

The Achilles' heel of systems like these is usually that the blue LEDs are at 455nm, which isn't ideal for color printing, where a slightly shorter wavelength is better. How much of a problem this is, you'll have to find out by giving it a try.

@koraks, have you been using such LEDs for printing color? I remember in your blog article on this topic, you weren't sure which blue wavelength should be used. The curves implied that under 450 is best, but the blue filter used by Kodak was longer. I bought some violet LEDs made by Cree, series XE-G. I haven't installed them yet, but they should be terrific for getting max contrast out of B&W papers. But I'm wondering what they'll do to color...

Mark

 

[koraks]

have you been using such LEDs for printing color?

Sorry, I made a typo. The regularly used blue in RGB systems is 475-480nm. Royal blue around 455-460nm is a better match for color paper.
It's been a few years, but I do recall testing the following:
* RGB COB LED with the regular 480nm blue and 620nm red. This gave overall poor color rendition.
* Discrete SMD LEDs at a rather low wavelength; I don't recall exactly, but it was something in the 425-440nm range. This gave poor, impure yellows, maybe due to crosstalk from the innate sensitivity of the other two emulsions, although the Fuji datasheets aren't clear on this.
* Discrete SMD LEDs royal blue at 460nm; this worked the best and is what I still use today.

Mind you, due to the way color paper is designed, you'll generally end up with crosstalk anyway at some point if you expose strongly enough with a blue light - you'll end up first activating the green-sensitive layer and ultimately the red-sensitive layer. The latter is really hard to accomplish, but the former is easy to see if you print a step wedge and expose it strongly enough. You'll see the yellow start to trend towards magenta at some point. This is not necessarily an error in the equipment or process used; it's an inherent property of the paper.

Note that in the Fuji datasheets, the spectral sensitivity is given for a density of 1.0 above dmin. It would be interesting to have the same plots for dmax of each color channel. I bet things look rather different, then. It would also be interesting to see the parts that were omitted because they were deemed irrelevant, such as the sensitivity of all three channels close to the 400nm vertical axis.

Theoretically, 480nm blue should work just fine, but for some reasons, my results with it were rather poor. Then again, in a similar vein, 620nm red should theoretically also work just fine, but 660nm worked better. Why? My speculation is that the 620nm LEDs had more secondary emissions at lower wavelengths (a peak in the green/yellow band) and thus caused crosstalk/crossover.

An interesting and also somewhat challenging effort would be to improve the green channel; in LEDs, 525nm green is the default, and green is (and remains) indeed the most challenging of the lot, from a semiconductor manufacturing perspective. A longer-wavelength green of 550nm would be better suited, but these LEDs are still not as common in high-power versions as the omnipresent 525nm types. I still intend to give this a try someday. Just haven't gotten round to it, yet.

they should be terrific for getting max contrast out of B&W papers.

In my experience, papers max out in contrast around the royal blue band or thereabouts and there's no utility in taking the blue wavelength further down. You'll just end up sacrificing some efficiency - which generally goes unnoticed since blue is never the bottleneck in terms of power output anyway. I had the same thought as you had back then, that a very deep blue would somehow be advantageous because it separates itself effectively from green. In practice, it doesn't quite work that way, or at least, that's what my attempts led me to believe. Let me know how you fare. I did most of my B&W testing with Adox MCP back then, and more recently with Fomaspeed.