My LED enlarger lightsource instructions / Complete parts list and build

[Ulophot]

Peter, as I have recently posted, a friend made one for me, using 3-D printing to make the housing he designed and the controller housing (which he is going to print if he can get around to it...). We had difficulty with light evenness and tried two sheets of diffusion but that made insufficient difference. Lining the inner housing between the panel and the diffuser with white paper apparently made none. The solution was to taper the light from the LEDs, making it brighter toward the edges. don;t ask me how; I'm not the programmer. It's not perfect at 4x5 full-frame but quite good and easily manageable. (I have certainly used enlargers with a lot worse distribution.)

I believe the panel is an inch from the diffuser (perspex/white acrylic), which I purchased from online in a 12"x12" piece and cut. It's 1/8" thick. I would have to open the head to check, which is takes a while. In any case, we simply laid the panel facing up on a desk, turned the lights out, and positioned the diffuser up and down, looking for the point at which the individual lamps disappeared.

[peter brooks]

Philip, many thanks for these notes, very useful. I see also your thread LED enlarger head questions, elgatosuizo's design.

I think plenty of empirical testing is the way forwards at the moment. I can program a little so customising the brightness of individual LEDs in the matrix may be an option if I can't get the diffusion to be even.

[koraks]

Just a remark on using WS2812 etc. for color / RA4 printing. Long story short: it's not a viable option.

There are three problems with this, two are deal-breakers in their severity, one is merely a nuisance.
1: The emission spectra of the red and the blue channels are not appropriate for exposing RA4. Like virtually all RGB led products, the blue is too long in its wavelength and the red too short. The WS2812 (and similar) are just the same in this respect as RGB LED strips, RGB COB leds etc.
2: The WS2812 is essentially an 8-bit device. I.e. it allows for 255 discrete intensity steps on each color channel. This is grossly inadequate for any even half-way decent color filtering on RA4 paper unless you employ additional measures such as turning entire led groups on/off to 'fake' additional color resolution. This will come at the cost of light intensity and evenness issues.
3: (This is the nuisance rather than the deal-breaker): for color printing, you basically want a truckload of red, a lot less green and only a little bit of blue. In practice, this boils down to something like a convenient ratio of 4:2:1 or so. Using WS2812's, this means that you're maxing out the red all the time and turning the green and blue down way low, which comes at the cost of additional color resolution problems (and 8 bit already was insufficient).

For B&W, WS2812's may be plenty good enough with the added red being convenient for focusing. For color work, it's a dead-end street.

The above is not just theory. I spent many hundreds (probably thousands, in fact) of hours on digging through theory, datasheets, engineering, building and testing various parts of a viable LED enlarger. I've gone through 3 generations of LED color enlargers; the first was very short-lived, using COB RGB LEDs which suffer from the same fundamental shortcomings as WS2812 etc. and was the test bed that taught me the lessons above (and quite a few more). The further two versions were viable enlargers that I actually used for B&W and color work extensively, with the latest version mostly being a power and usability upgrade.

To the best of my knowledge, there are currently no addressable LEDs, RGB strips or RGB COB LEDs that will give color fidelity anywhere close to a dichroic head, nor do I think it's likely these will ever exist due to the total lack of any other application that would require the same combination. On the upside, 660nm red LEDs have become far more common because they turn out to be very nice for plant growth. This means that a decent/good color head can be made at acceptable cost using high-power SMD LEDs, providing one is capable of engineering and manufacturing the required light source. It's not enormously difficult, but by no means a beginner project.

Edit, a few days later: this thread inspired me to write a proper piece about this issue, because it's something I've spent a lot of time on, so I might as well share everything I can get onto (digital) paper about it. Here it is, for those who are interested: https://tinker.koraks.nl/photography...-ra4-enlarger/

[elgatosuizo]

Oh man! Not this again. Sure it‘s far to be perfect for RA4, but it works. Somehow, but it works. Period. I had similar discussions of the B/W grade about how it’s impossible to hit the highest contrast grade. With a kind of dissertation on UV LED. But in the praxis the guys from Ars-Imago tested it and found it valuable. That beats all theoretical calculations. My experience with RA4 is indeed that the fundamental colors are not proportional to themselves in density. I empirically compensated this in the code with a kind of matrix. This lowers the resolution even more. But I can agree with you that for RA4 a good old tungsten enlarger may still be the better option if you want perfection. My project is a low budget possibility to retrofit an enlarger. Using a set of different LED (with suitable wavelengths) with an DAC controller is much better. But an overkill for 99% of the readers in this forum.


Gesendet von iPhone mit Tapatalk

[willwilson]

Marco, this is truly a fantastic contribution to the LF community. Thank you. I am building your controller now.

Question, if I was going to use an external timer what would be the best way to do that mod? Maybe just use the light on switch connection, button 5 on your controller, or just drive the 110v power to the power supply with the timer and have the led controller always on. Thoughts?

Thanks -=Will

Will Wilson
www.willwilson.com

[Ulophot]

Peter, the distance from the panel to the diffusion is actually 1.5". Due to a recent mishap, the code was lost, and at the moment, only the assembled code is available. That caused a problem, because when my friend finally brought the finished controller over today, there was an issue when it was hooked up, which he could not trace from the wiring, which all looked fine. He's hoping that colleague of his can reverse-engineer the coding so that he can see what's going on. Fingers crossed.

[elgatosuizo]

Will: yes you can use the “Lcd on” input with an 110v AC coil relais connected coil-side to your timer an with the switch side (normally open) to the lcd-on input. It triggers the “s_4” variable with digital value read on that input.


Gesendet von iPhone mit Tapatalk

[Tin Can]

Will Wilson

Love your prints and

it was good to study the Howard Bond Reciprocity chart you include on your site.

btw, I bought Howard Bonds enlarger and met him briefly.

Bond DIY converted a 4X5 to 8X10.

Marco, this is truly a fantastic contribution to the LF community. Thank you. I am building your controller now.

Question, if I was going to use an external timer what would be the best way to do that mod? Maybe just use the light on switch connection, button 5 on your controller, or just drive the 110v power to the power supply with the timer and have the led controller always on. Thoughts?

Thanks -=Will

Will Wilson
www.willwilson.com

[willwilson]

Tin Can, thanks for compliment and very cool about Howard.

Marco, relay is a great idea. Thanks.

Will Wilson
www.willwilson.com

[willwilson]

I completed my build. Everything works great! Marco, you are the best. Great project and great documentation.

I used an Aristo 1212 cold light as a base. I included some photos of this in case it was helpful to others. This worked great with 4 - 256 pixel LED (16x16 LED, https://www.amazon.com/gp/product/B01DC0IOCK/) and a 60a 5V power supply (https://www.amazon.com/gp/product/B01D8FLZV6). I removed the Aristo light bulb assemby, this is where the LED panels go. Next I cut the wiring holes in the Aristo housing (hardest part, but not too hard). The power supply fit in the top of the Aristo head, see first photo. I used a solid state relay to wire the controller up to my Stop Clock. I had this relay on hand at work, https://www.automationdirect.com/adc...ssr610-ac-280a, probably a smaller option out there. Two things, wish I had made my aristo wiring holes bigger. The double sided type is very unforgiving. You can see the slight misalignment in the LED example photos. I can't imagine this will matter but it hurt my pride a bit, lol. Second thing, I really wish I had gotten a larger project box for the controller. It would of been much easier to install and wire up. This thing is bright. Printing test next!






Thanks again Marco. Omega color head...here I come.