Build my own color head?

[Donald Qualls]

I'm interested in color printing, and already have a 4x5 enlarger, but it has a cold light head that I'm told will cause me to tear out what's left of my hair if I try to print with a filter drawer (and never mind there's no sensible place to put a filter drawer).

Buying a color head to fit has turned out to be hit or miss on price and availability, and would require an upgrade to the working mechanism. Buying a whole enlarger would pretty much depend on finding one within driving distance -- shipping over a longer distance will cost nearly as much as the complete enlarger.

I've seen a number of cases where people have built multi-contrast LED lamphouses for existing enlargers -- and it occurred to me that with high power LEDs in red, green, and blue (if I can get the correct wavelengths, at least) I ought to be able to build a color head, using time-slice (aka duty cycle) dimming to vary the mix of the three primary colors, much like dialing a filter setting on a color head's controller. That is, dial some cyan filter, and red gets dimmed a matching amount. Dial yellow, and you cut down blue.

The electronics for this (DC choppers, same as motor speed control) aren't hard to buy, and high power LED modules also are readily available. The big tricks will be getting the right colors to match the sensitivity peaks of RA-4 paper, and getting the light patterns from the LEDs correctly mixed to give even illumination on the negative.

Any advice ("harder than it looks" should please include why), experiences, etc.?

 

[Tom Kershaw]

I think the member 'Koraks' is the person you want.

 

[Doremus Scudder]

Donald,

This isn't "harder than it looks," rather "easier than it may seem":

You don't mention what brand or model of enlarger you are using, so I don't know what OEM-compatible color head might work for you. That said, here are a couple of work-arounds to get rid of your cold-light-head problem:

First, get a regular condenser head with a tungsten light source (you may have one already?) and use CC filters. Once you zero in on your filter pack, you usually only have to make small changes from print to print. Slipping in a 10CC color filter is almost as easy as dialing in the change.

Finding an enlarger with a color head may be easier than you think. Omega D-series enlargers with color heads are really plentiful and not so expensive these days. Keep an eye on Craigslist and auction sites and be patient. Something will almost certainly turn up.

Or, you can do what I have done: adapt a different brand of color head to your enlarger. I have a 5x7 Chromega E head mounted on my Beseler 45MCRX. It took a bit of machining, filing and cobbling things together, but it works perfectly and the larger size gives me more even illumination for my 4x5 negs than the (questionable) Beseler color heads.

And, if by any chance you have a Beseler 45, I've got a Beseler color head that worked last time I checked that I'd be willing to send for the price of shipping.

I'm not trying to talk you out of building your own color head, but LEDs may present some spectral-response issues and end up being more difficult to deal with than you anticipate...

Best,

Doremus

 

[Donald Qualls]

@Doremus Scudder What I have now is an Omega D2/DV with a Zone VI cold light and its diffuser mounted below where the variable condenser glass would have been -- that is, the diffuser is just above the mask that contacts the top surface of the negative carrier. Works great for B&W, and I've split printed with blue and yellow gelatins held below the lens on a multicontrast filter fork (though it's been more than a decade since I've powered up, I presume the light still works -- if not, I've got another reason to need some kind of upgrade, since Zone VI cold light parts are hard to come by).

I got the enlarger in this condition, with a very few negative carriers, in 2003; I've bought some additional carriers and still need more (I shoot a bunch of different film formats). No condenser glass, no lamp house. To use the Chromega D series heads, I'd have to upgrade the lifter arms (not impossible, they show upon eBay and are apparently available from KHB), but I'd also have to find a working D size Chromega head that I can afford. To use an original D series lamp house, I'd have to buy one and a set of condenser glass (not sure about the lamp house, but the condenser glass was a couple hundred bucks ten years ago).

Spectral response is the core of my statement about "correct wavelengths". Red, green, and blue aren't hard to come by, paper sensitivity curves are published -- the spectral output curves of color LEDs aren't as easy to find (usually have to drill all the way back to the module manufacturer, which can be hard information to find for a commercial lighting unit).

Any old blue, red, and green should work to some level, given the paper response curves are bell-shaped and LED outputs tend to be fairly narrow spikes -- worst case, I'd expect to need some adjustment to find neutral, and I'd have to do that with filters, too.

 

[DREW WILEY]

Yeah ... Just shoot that rhinoceros enough times with a BB gun. By the time it finally falls, you'll have spent more on wasted paper & chemicals than the cost of a real colorhead.

 

[koraks]

@Doremus Scudder
Spectral response is the core of my statement about "correct wavelengths". Red, green, and blue aren't hard to come by, paper sensitivity curves are published -- the spectral output curves of color LEDs aren't as easy to find (usually have to drill all the way back to the module manufacturer, which can be hard information to find for a commercial lighting unit).

Any old blue, red, and green should work to some level, given the paper response curves are bell-shaped and LED outputs tend to be fairly narrow spikes -- worst case, I'd expect to need some adjustment to find neutral, and I'd have to do that with filters, too.

Well, my experience is that not 'any old' set of LEDs works well. Spectral peak does matter, particularly for the red and the blue ones, with the red ones being the more challenging ones. The problem is that most red leds are too close to the green side of the spectrum (620nm), but you really need 'deep red' at 650nm or preferably 670nm, but I personally wasn't able to source the latter (at least cheaply and easily). For blue, it seems that royal blue or near-UV works OK, but you don't want to go below 425nm (too close to UV). Problems I experienced with LEDs that were too far from the spectral response peaks of the paper were uncorrectable purity issues with primary colors as well as color crossover between shadows and highlights. I settled for a set of LEDs as explained in my thread of a few months ago where I documented by color LED enlarger build: https://www.photrio.com/forum/threads/craziness-using-leds-to-print-ra4-and-b-w.171911/
This device, integrated into a very old Durst 138S enlarger, has been my 'daily driver' for B&W and color in all formats from 35mm up to 4x5" since December 2019. Apart from a minor software update somewhere in January I did not change anything about it and it has served me very well.
Currently there are a few issues that I'd like to resolve in due course, but this hasn't been a priority since basically, the thing just works and gives me prints that are entirely on par with those I got from a 'real' (dichroic) color enlarger, but the LED contraption is more convenient in daily use since I could of course tailor the user interface to my personal needs. The remaining issues are:
* Somewhat annoying user interface response issues with the rotary encoders and buttons sometimes 'missing a beat'; i.e. they sometimes don't register a movement or button press event. This is due to the combination of fairly computation-heavy software (it turns out that in its full functionality, the whole thing involves quite a complex software implementation that pretty much runs the controller hardware to its limits) and suboptimal interrupt service routines (too computation-heavy, resulting in missed interrupts). This is mostly just an annoyance and the controller works fine apart from it, so I haven't done much about it.
* For B&W work, it turns out that it would be nice to have significantly more green light than I currently do. The problem is likely due to the 'green gap' in LED technology, which results in the efficiency of green LEDs being way less than that of blue and red ones. With neutral papers, this is not much of an issue, but when enlarging 35mm negatives to larger sizes on warm-tone (slow!) paper, exposure times tend to get too long for my taste. The solution is obvious: build a new light source with much more green LEDs. I've got the parts for this, but haven't gotten to it yet. I'd also need to upgrade the green driver channel, which is technically not complex, but just one of those chores that haven't taken precedence over taking pictures and making prints.
* For some odd reason, in color work, if I use the red channel (along with blue & green) for focusing, this influences the light output of the red lights for a subsequent exposure within the next few minutes, resulting in unpredictable color shifts. Apparently something heats up (either the current limiting electronics, the power supply or the LEDs themselves), creating temperature-stability issues. If I'm going to construct a new head, I'm going to drastically improve the cooling system to improve temperature stability in the hope that it helps. For now, I just only use blue + green for focusing in color mode, which turns out to be a totally effective workaround.

Overall, your thinking is sound, at a conceptual level. R, G and B LED arrays, dimmed using PWM, driven with current-limiting hardware, and some calibration to make sufficiently precise filtering possible. I worked out those things for my use case and it just works. In fact, the filtering precision of my implementation is far better than that of a typical dichroic head - i.e. 1cc of change on a dichroic heads is something like 2-3cc change on my LED head. This additional precision isn't really required, but I 'got it for free'; i.e. there was no penalty in implementing it in this way. To date, I have not been able to find any problems with consistency; the light source behaves predictably and gives good prints. I have not run colorimetric tests on it (I barely heave the means to do so, or in fact not at all), so I can't tell if the results are 'perfect'. However, they are certainly good enough for my eyes. I've been told by a fellow photographer I'm rather color-sensitive, implying that fairly good results (putting it cautiously) are achievable with this approach given the fact that I can get the colors the way I'd want or expect them in my prints.

If you want to undertake something like this, I'd recommend going through my threat that I linked to above and check out the various challenges and problems I ran into and (mostly) solved or worked around. With this concept, the devil is really in the details. The basic approach isn't very complex if you have an average understanding of electronics and digital control systems. I'm not an electrical engineer by any means, so fairly well-developed layman's knowledge can be adequate. Just be prepared to run into unanticipated issues and systematically tackling them. It's not a matter of 'if', but 'when and which' problems you will run into.

 

[koraks]

By the time it finally falls, you'll have spent more on wasted paper & chemicals than the cost of a real colorhead.

This was certainly not the case for me. Financially speaking, the majority of the investment was in electronics, on which I spent a couple of hundred euros when all is said and done. I haven't run the numbers on it and I certainly ordered all kinds of stuff for later use, but I estimate the costs including stuff used for testing that I didn't end up using in the final system at around €300. In terms of paper and chemistry, I spent probably less than €50.
The major investment is really in time. I worked on this on and off for a few months. I probably spent around 3 weeks fulltime (spread out over a ca 6 month period) on designing, testing and building. If you put a price tag on those hours, it dwarfs the material costs, obviously.

 

[Donald Qualls]

Hmm. Obviously (to me, at least) I won't be able to start anything on this until my darkroom construction is finished. I was looking for an analog solution on the controls -- essentially, dimmer knobs or sliders calibrated for cc. I have programmed, but my take is, why get a computer involved if a potentiometer can do the same job? @DREW WILEY I guess we should all just go back to our caves -- I'm pretty sure there are some marrow bones waiting to be cracked open, if we dig in the midden...

 

[koraks]

I have programmed, but my take is, why get a computer involved if a potentiometer can do the same job?

You're not going to get the required precision with a set of potentiometers and you will run into insurmountable problems with linearization. Yes, it can be done with analog circuitry, but it will be vastly more complex both in terms of hardware and in terms of engineering than a microcontroller and some software. I predict you will inevitably hit a brick wall and/or end up with an unusable 'solution' if you go the analog way.

 

[Donald Qualls]

Honestly, if I have to have a computer in it, I'll probably never build it. I'm not much of an electronics person, and surface mount components are for machine assembly, not human handling. I can cobble up an SCR controller, probably even a voltage regulator, and after they're breadboarded I can fabricate a circuit board, but I think you have an inflated assessment of "average" electronic ability. The "average" person's electronics knowledge may or may not include checking whether the device has power; further troubleshooting generally involves dialing a phone to talk to a support technician or returns department.

 

[koraks]

Well, yes, I have to admit it's not really a simple project. But the electronics aspect isn't overly complicated, neither in terms of design, nor in terms of physically building it. No SMD soldering was involved in my build, no custom PCB manufacturing; it's all a combination of ready-made building blocks and fairly simple discrete electronics soldered onto experimentation board. Heck, I didn't even use an oscilloscope (let alone a digital signal analyzer)...
However, this kind of project still does require a bit of knowledge about electronics and I agree that this 'bit' is more than what the average person has at their disposal. So I should have worded it a little differently.

And yes, it likely does have to have a computer in it. As I said, it can be done without one, but that will make it all way more complex than just chucking a little arduino (or whathaveyou) in there and solving the more challenging stuff in software. Getting a perfectly fitting curve-shaped response (as I needed) was a matter of a handful lines of code and a PWM controller with the possibility to modify it based on tests, while it would have cost me hours and hours of SPICE simulation and soldering to get it done in discrete and analog components.

 

[Donald Qualls]

For me, "Adruino" is out of the question. Embedded programming is far from anything I've done, I'd have to learn it from scratch. I'd have a chance, of sorts, with a Raspberry Pi or equivalent, with an OS to deal with if something crashes and a display to see the results before I turn out the lights -- but even then I'd probably walk away.

Hmm...

 

[koraks]

I had to learn Arduino too for this purpose, but I had the advantage of having dabbled in some programming and scripting since my teens (basic, java, php etc.) The good thing about Arduino is that the web abounds with examples, libraries, forum posts etc.
I suppose you could go the Pi route if you so desired. However, I'm not sure if it'll actually be any simpler than an Arduino tethered to a desktop or laptop computer for debugging/monitoring purposes.

 

[Donald Qualls]

I learned BASIC in 1975, but the last time I programmed was when I tried to simultaneously relearn C and learn Python about 6-7 years ago. I don't really have time to learn a whole new way to program, or the technology of Arduino (or Pi, for that matter) -- not retired yet, probably not for another ten years.

 

[DREW WILEY]

Well, I have built a couple of very serious precise-color custom additive colorheads and know what kinds of issues are involved. Anyone thinking about rheostatic control is the one living in a Pleistocene cave, not me!

 

[Doremus Scudder]

Do you want to print, or do you want to engineer and build an LED color head? It's your decision how you spend your time. It seems that the cost of a color head would be about the same/less than a DIY LED head.

Whatever you decide, have fun!

Doremus

 

[koraks]

It seems that the cost of a color head would be about the same/less than a DIY LED head.

It is. I couldn't find a suitable head when I was looking for one, so necessity was the mother of invention in my case.
You're certainly righy about the time investment. My project involved the "opportunity cost" of a significant amount of darkroom time.

Anyone thinking about rheostatic control is the one living in a Pleistocene cave, not me!

Well, he didn't suggest a rheostat. That is indeed caveman technology.

 

[Donald Qualls]

I most certainly did not suggest a rheostat. My darkroom is connected to the house central heating and air, I don't need a resistance heater added. I suggested a chopper type circuit, like a DC motor speed control, aka PWM. I think it would be possible to get the necessary precision that way,perhaps with long travel sliders or switches with calibrated resistances, but as @koraks pointed out,it'd be more electronics design work, building, and testing than an Arduino based system. I could buy the parts I need (electric bicycle throttles would work) but then I'd have to match voltages at various stages. Almost certainly easier to keep shopping for a bargain on a color head that will just bolt onto my D2.

 

[Chan Tran]

If you have problem with the precise dimming you can use 3 timers and expose the 3 color for different fimes. At the start all 3 lamps would be on and then the color needed the least will shut off first and then the next. It would be more precise than a potentiometer.

 

[DREW WILEY]

I suppose one can find some old Popular Mechanics article telling you how to build a nuclear submarine using spare washing machine parts. Shop enthusiasts have fun with all kinds of projects, including me. There is nothing wrong with making something that sorta works. But if someone is expecting high-quality color reproduction from anything LED at this adolescent stage of that particular technology, all I can do is roll my eyes. Yes, you'll get something - but what? Sufficient for some fun perhaps. More affordable than simply going out and picking up a used halogen colorhead? Very unlikely. Able to render hues as accurately? - even more unlikely. Besides the electronics issues, filter nm bandwidth has to be quite narrow, or else everything else that gets through those filters constitutes residual "white light" affecting all three dye layers, hence contributes to color contamination, muddiness, or crossover. You are also going to have a size limitation of small prints due to a lack of serious lumen punch; but not everyone needs big prints. I applaud anyone attempting to obtain a good cool light source using LED's, but do feel it's a bit premature for serious color printing, and more realistic for VC black and white printing applications at the moment. Kinda like the Wright Brothers still testing gliders rather than actual airplanes. I built my own colorheads specifically for very high color accuracy using a pulsed halogen system. One of these units was bright enough to punch forty inch Cibachromes from 8x10 masked originals, so orders of magnitude more powerful than needed for color neg printing that size. But the extra lumens are helpful composing the image through that dense orange mask. I allotted a $15,000 materials budget for that particular colorhead and its support, and used up every bit of it. Ironically, I have a somewhat smaller Durst 8x10 subtractive (CMY) color enlarger right next to it that originally cost around $26,000, but that someone gave me for free. The difference in color quality per se is detectable, but rather minor.

 

[Donald Qualls]

If you have problem with the precise dimming you can use 3 timers and expose the 3 color for different fimes. At the start all 3 lamps would be on and then the color needed the least will shut off first and then the next. It would be more precise than a potentiometer.

Of course, be sure to do any dodging in the first part of the exposure...

 

[btaylor]

If you enjoy this sort of thing, building your own color head, then by all means you should do it. The “opportunity cost”’will be worth it.
But if you just want to get printing, a proven used color head seems like the better choice. Keeping an eye out on CL and eBay often yields the bargain. I have purchased multiple working color heads for my Beseler 45 for under $100 in the last couple of years. I would think a DIY conversion with ordinary hardware store stuff wouldn’t be terribly difficult. My 2 cents worth.

 

[koraks]

You are also going to have a size limitation of small prints due to a lack of serious lumen punch

This is certainly not the case. For color work, the led head I currently use gives printing times that are on par with a 300-500W incandescent/halogen source. It just dissipates only 20-30% of the power.

As to color rendition - as you havent done side by side comparisons with your undoubtedly discerning means and neither have I beyond some informal testing, I'd be hesitant to adopt your stance of "it really can't be good". We just don't know for sure.

 

[Chan Tran]

Well as far as color I think it's even possible to expose modern paper with 3 monochromatic light bulb. That is 3 light bulb and each only emit light of 1 wavelength. I think new paper is optimized for that as they are used in digital printer that use laser beam to print on RA-4 paper.

 

[DREW WILEY]

Koraks, that's below the luminance level I'd want for a serious pro enlarger, but probably OK for garden-variety 4x5 work. I haven't seen a tad of evidence so far of LED's even existing capable of the narrow nm output ideal for color printing. With tungsten-halogen, you've got enough continuous blackbody spectrum to section out specifically what's needed. It's a different problem than blending LED's in architectural or display bulbs to fool the eye that the color is actually balanced, whereas the spectrum is discontinuous just like in fluorescent lighting. It's getting there; but color enlargers are probably very low on the list of LED mfg priorities.... Chan- many current RA4 papers are optimized for both RGB laser and conventional enlarger colorheads. I happen to mainly use true RGB colorheads which give exceptionally good response; but recent subtractive CMY heads also perform well. My design uses sandwich "trimmer" filters which cut off light either side of each of the three independently-controlled spectral peaks. There's a distinct trick to it because heat build-up changes the characteristics of dichroic filters somewhat, and too narrow an nm band can be counterproductive. Additive enlargers are more difficult to design than subtractive ones. But all subtractive CMY colorheads have a bit of white light spillover which can contaminate the other dye peaks a bit, especially if the filters have gotten old and some of their coating has spalled off. Having the cooler light that LED could hypothetically provide is of course a nice feature; but there are still a number of issues that complicate any simplistic substitution. It will be fun to see what transpires over the next decade or so. But I'm anticipating all my own enlargers lasting well past my own lifetime, so probably will not myself be experimenting with any new designs.