LED enlarger head questions, elgatosuizo's design

[Ulophot]

By some miracle, I have a new neighbor who is both a "shop guy" and a programmer, and is excited about building an LED head for my D-2, something way out of my ballpark.

A number of people said they were going to build elgatosuizo's design (https://www.largeformatphotography.i...list-and-build). Has any of you done so? And if elgatosuizo's listening, by all means chime in.

I print only B&W, using VC papers. I have given my neighbor the relevant PDF as well as the 8x10-model PDF by BJ Harlin, which uses only specific blue and green LEDs. We're barely in the planning phase. We'll build the box around the 256-LED grid indicated by elgato. Neighbor has a 3D printer and I'm thinking to build the light box of wood or metal and 3D print a piece that will attach to the top and hook it to the three D-2 housing screws.

Here are my immediate questions:

1. Have you experienced any glitches to pass along?
2. How does the light output compare with "legacy" sources, e.g., Aristo/Zone VI round coldlight (my present rig) or 212 bulb and condensers or diffuser
3. Would I be better off using the blue and green LEDs in such a roughly 6.6" square enclosure?
4. Can I continue using my Zone VI Compensating enlarging timer with the sensor? (Note: I do not use a split-filter printing approach, i.e., pure green and then pure blue; I want the mixed-color light for each contrast.)
5. Does the output remain relatively constant as different contrasts are dialed in?

Thanks.

[Eric Woodbury]

Disclaimer: I have not made the elgatosuizo design. My comments are based on five LED sources that I've designed and used.

I'm using a 6x6 array of 1Amp LEDs made by Cree mounted on a 1/4" aluminum plate that is 10x12". Very efficient LEDs. Half green and half royal blue, same wavelenths are those above.

2. It is about 6x brighter than my V54 1212 Aristo coldlight. Part of this is that I use an efficient diffusor. Highest contrast is about 1/4 point higher than I could get with Aristo coldlight and the sharpest dichroic cutoff filter I could find.

4 & 5. If the LEDs are driven with controlled current sources, then a closed loop timer is not needed. I use a MetroLux, but in SECONDS mode without the sensor. The pure blue/green LEDs will mess with the sensor's ability to control exposure since the sensor has different sensitivity to those two colors of light. Furthermore, the design should be such (and I don't know that it is) that the light output is constant regardless of contrast. An example highest contrast should be 100% blue, no green. Lowest contrast should be 100% green, no blue. Middle contrast should be 50/50. This gives a consistent highlight exposure across all contrast settings.

It is important that the LEDs don't change temperature much. The hotter they get, the less light they produce, which will throw off your exposure. So they should be well heatsunk. Try to hold the temp change to less than 10C. Also, heat reduces the life of the LEDs. Driving the LEDs with a current source insures that the intensity remains constant (assuming the temp is constant). Given that this design runs on 5V, the LEDs are not being driven on individual current sources. If it is isothermal and the LEDs are all from the same lot, then you are probably fine. Also, since these are lower current LEDs and there are so many of them, the consistency is not as important.

I really like LED light sources. The contrast adjustability is wonderful and quick. The light is bright. It will be good to get a few reviews of this particular circuit that you are considering. Hope this helps.

Good luck. -e-

[Ulophot]

Thanks very much, Eric. I'm forwarding your reply to my builder friend for reference. He has had me order elgato's 16x16 grid, along with a convertor, jumper cables, and arduino board, and a apiece of plexi for testing light output. If it's not sufficiently bright, we'll be back to blue and green LEDs. I have a funny feeling that that's where we'll be. Do you have a link to spec and/or photos of your LED layout?

Also, can you say anymore about temperature? We have no central AC, so in the summer it a window unit in my upstairs darkroom and I typically keep room temp around 70-72 while working. Do you have a single heat sink, such as finned aluminum(?) one that you can recommend?

[Eric Woodbury]

https://www.ledsupply.com/leds/cree-...power-led-star

Lots of info here and there is a link to the full blown datasheet.

With 36 LEDs on 12x12 aluminum, it is a simple 6x6 grid, alternating blue green like a checkerboard. Since the LEDs are sparse, the illumination at the first diffuser is uneven. Even with the common, dense, white sign plexiglass used in Aristo coldlights, the LED 'spots' come through. I changed the diffuser design to be two diffuser plates made of Satinice (>90% transmittance), first one is 1.5" from LEDs and the second is the top of the negative carrier, as is done with the Beseler 12x12. For my smaller negatives in standard neg holders, then the 2nd diffuser is very close to the negative. Regardless of the exact design, the first diffuser should be 1.5" from these LEDs (because of the little lens that is part of the LED) and the second should be at least 2" away from first. Double diffusion produces even illumination and with over 90% transmission for each diffuser there is more light than with a single white plexi diffuser.

Photos show 12x12 aluminum plate. 3/16" thick. One side is the LED pattern and the other is the connection side. LED 'stars' are glued with thermal epoxy to the plate. Making these plates with LEDs is not hard, but it is tedious and time consuming. The little extension wires must be soldered to the LED stars before gluing to the plate. Connection wires are then soldered to the extension wires once it is all assembled.




I've also made a smaller version on the LED Supply heatsink kit. For that I used "3-up" LEDs, but this is unnecessarily bright.

I should say that having a very bright light source is nice if you want to print those very dense negs.

With only 36 LEDs, I use four current sources of 1A. LEDs are wired 9ea in series and the main current is derived from 36V. My electronics are all analog style.

For a client, I have a design for making very large prints that is much greater. It has proven to be difficult. Control signals are small and the total power controlled is very large. This has made grounding a pain in the ass. It has 144 LEDs. I wear special blue blocking glasses when I work with it. It's scary bright.

[Larry Gebhardt]

My first LED head was just 3 watt white LEDS in an even grid for 5x7. I had issues with evenness where the center was too bright. I forget the exact details on the number and spacing, but it all fit in an old Omega E2? head so height was limited. With a pencil mask I was able to get it evened out. Test for evenness as early as possible in the build.

My next was a VC head for a Durst 137S that I documented at http://www.trippingthroughthedark.co...8s-condensers/. That has worked well for me.

[Ulophot]

Okay, this is getting exciting! With basic parts received, my builder friend has assembled the arduino controller, 16x16 LED pad, and power source. I tested the light coming through a piece of 1/8 white plexi at a distance adequate to even out the light, and the reading with my light meter is the same as my Zone VI head. So, the illumination level, which was my main concern, is fine. Since each of the 256 LEDs is individually adjustable, falloff toward the edges can be compensated, if I find that necessary, though expect I may not -- not any more than my coldlight head may produce now.

He is going to add an inline switch so that I can use the unit with my Zone VI compensating timer, as long as -- I have to double-check this -- the dry-down compensation cuts time, not voltage, which I am fairly certain it does. The power supply and arduino board can be installed under my enlarging table, and an LCD readout combined with a couple of potentiometer dials will allow me to get the light color where I want it. I will continue using my Ilford VC filters, so this will just allow me to find the setting I want.

He will design and 3D print a housing that will attach the light head to my D2.

[Ulophot]

New question, from the builder.

The inline switch: He said he's having difficulty finding an AC-DC solid-state relay. The job is to switch the light on and off without turning the power supply on and off. Anyone have a source/idea for this?

Again, this is elgatosuizo's design (linked in my OP, above). I guess the difference is, he is using a controller directly connected to the LED pad. I'd prefer to keep using my familiar Zone VI timer, with it's very useful foot switch and dry-down compensating circuit -- unless this can be built into a controller.

Any ideas?

[Ethan]

A foot switch should be very easy to add to the controller. You could wire another switch to the arduino and have both pins activate the same timer loop. Or, maybe simpler, set the switch circuit to be normally closed, and then have both switches in series, and when either of them gets hit it will open the circuit and start the timer.

At least circuitry wise, I think it will be easier to add the dry down compensation to the arduino controller. You would have to do some tests to figure out the math for dry down compensation however. I've never used one of those Zone VI controllers, but I imagine it's just a multiplicative factor that decreases the time of exposure. Once you know what that factor is for your paper/papers, it should be a simple matter to having a button which multiplies your exposure by that factor. I suppose the factor might change depending on what the densest part of you image is, but if that's the case I think you could just reference a function for the multiplicative factor rather than a constant. If the factor is a constant it would be fairly simple to program it into the controller using the keypad when you use a new paper, if its a function it would probably have to be programmed into the controller code, but now that I'm thinking about it, you could also input several factors with the keypad and have a program which builds the function from that.

Back to what you asked, what I'm saying is I would probably build the compensation into the arduino controller rather than have a relay to activate the enlarger with your Zone VI timer. It would mean modifying the code, or perhaps writing new code though, so depending on where your skills are finding a relay could be easier if you haven't done much programming.

If you do decide to stay with your Zone VI timer, here's what I think may work:

Does the Zone VI timer output 120V AC power? Your builder probably already has ideas on how to connect it to the arduino controller, but if I were doing it my first instinct (and this may or may not work, I'm not as knowledgable on this type of circuitry) would be to treat the timer as a switch, and use a voltage converter to turn the 120v AC output of the timer to 5v DC, which is what the arduino's logic level is. Then I think you could wire it to one of the arduino pins as a switch, and when it gets 5v power from the voltage converter, it would see that as the circuit closing and turn on the LEDs.

[Ethan]

Thinking about my voltage converter idea, I'm guessing it wouldn't actually work since the output would probably fluctuate too much and could cause inconsistencies. However, looking at Mouser, I see many relays which I think could work. Do you know what the specs you need are, or why your builder is having trouble finding one?

[Eric Woodbury]

Dry relay easiest. Check your ratings.

I used a PFET and an opto coupler. Tricky business.

Whatever. You want ground isolation from unit to unit.