Converted 8x10 Dichroic to LED!

[Geary Lyons]

Well after about 2 months of toasting $3 Luxeon III LED’s like they were popcorn, I have successfully completed the conversion of a Durst 184/301 head to LED light source.

Firstly, a big “Thank You” to Huw, whose site and emails helped me to get started. His DeVere 504 conversion was my inspiration. I modified Huw’s layout to accommodate my 10”x10” Durst head and used Luxeon III’s, rather than Luxeon I’s.

My printing process is pretty much some graded and mostly VC. For VC, I use the split grade technique utilizing a RH Designs StopClock. So my design elements were SoftMax and HardMax, as simply as possible. What a joy, no more “spinning the dials” on my color head. Just flip 1 toggle for Soft, green, 1 toggle for Hard & Graded, (blue), both for focusing.

My previous 8x10 working head, a 1000 watt quartz halogen dichroic, worked very well. Quartz halogen lamps produce most of their light in the red range combined with copious amounts of heat. I was looking for a simpler interface, less heat and less noise, (fans), So the comparison, easy, minimal heat and perhaps, no fan needed.

What about illumination efficiency? The LED head is 2 ¾ stops FASTER! That makes the LED’s illumination equivalent to my old head with 2750 watts. Think about it. It would take 11 quartz halogen lamps and 2 ¾ time the heat to generate the same hard/soft printing power!

The Durst head was really a joy as a base platform for the conversion. Lot’s of room in the fan compartment for the 24 volt power source and wiring harness runs. I left the internal fan in place, just in case I need to dissipate some heat faster than my heat sinks alone will accomplish. It has 10x10 and 5x7 mixing boxes that slide in like big drawers. Take the mixing box out and the led panel is readily accessible.

I am still sorting things out, but the initial testing and process run was very, VERY encouraging.

I will post more as I get more experience with the new LED head, but so far, I’m sold.

Cheers,
Geary

[ic-racer]

This is really fantastic, but can you post pictures and a schematic?

[BarryS]

Sounds great. I'm mostly a split grade printer myself, so the single switch LED source sounds like it would be a good upgrade. I'd *really* appreciate some solid documentation of your project. The commercial market is not really doing much in the way of darkroom technology development, so these homegrown projects are one of the only ways to move the art forward. Definitely check out some of ic-racer's excellent posts for a good example of project documentation.

[Geary Lyons]

This is really fantastic, but can you post pictures and a schematic?

I am working on a more detailed article with pictures & schematics.
Cheers,
Geary

[Jon Shiu]

Wonderful news, Geary! Hope to see some of your prints sometime!

Jon

[Geary Lyons]

Sounds great. I'm mostly a split grade printer myself, so the single switch LED source sounds like it would be a good upgrade. I'd *really* appreciate some solid documentation of your project. The commercial market is not really doing much in the way of darkroom technology development, so these homegrown projects are one of the only ways to move the art forward. Definitely check out some of ic-racer's excellent posts for a good example of project documentation.

Well, I will do my best! I will keep my "report" generic. That way the basic principles could be applied to other heads/enlargers. It will have, perhaps, more practicality than a brand/model specific article.

Regards,
Geary

[rob]

What LED color do you use? I 've been planning to do the same using luxeon royal-blue and green for my 45mx and vivitar VI.

[David A. Goldfarb]

Sounds excellent. I could see a homebrew 8x10" enlarger in my future.

[Geary Lyons]

What LED color do you use? I 've been planning to do the same using luxeon royal-blue and green for my 45mx and vivitar VI.

Rob,

Those are the colors that I used. The green wavelength is right on for Soft contrast. The royal blue is the best available choice. The royal blue is just a slightly longer wave length than the peak Hard contrast sensitivity, but in split grade printing you will not notice.

Cheers,
Geary

[Turner Reich]

I was checking and the Leds you mentioned are listed for $5 and up in price, where can we get them or $3.?

[Geary Lyons]

I was checking and the Leds you mentioned are listed for $5 and up in price, where can we get them or $3.?

Future Electronics is the direct supplier.
http://www.lumileds.com/products/line.cfm?lineId=2

Wow! Timing is everything. They now list them at $4.86 in cut quantities. I paid $2.86 for my first batch, then $3.17 a month later. That was about 3 weeks ago. The price was the same for cuts as a reel of 250. Now they have raised the price of cuts. They have really raised the price over the last 60-90 days!

Glad I learned how not to destroy them on the "cheap ones"!

Geary

[Turner Reich]

Glad I learned how not to destroy them on the "cheap ones" and will share that knowledge with all of you.

[Geary Lyons]

Glad I learned how not to destroy them on the "cheap ones" and will share that knowledge with all of you.

Well, everyone but you!

It is pretty simple. Do not be a heavy-handed klutz. These LED's are not designed for point-to-point hand soldering. The contact paddle is attached to the body with a very narrow, thin lead. They break off easily. The LED's are also quickly susceptible to over heating from too much or too long solder iron heat. They quickly transform into NLED's. Then the dead-leds must be removed from the circuit and hopefully without destroying the paddles on the lead-in and lead-out partners. I quickly, (just not quite quickly enough :eek: ), learned to "flash" each LED before and after soldering into the circuit.

My soldering experience comes from tube audio equipment repairs and and upgrades . I needed to develop far more finesse in handling this miniature buggers.

Cheers,
Geary

[Geary Lyons]

The Durst CLS 301 was a great head to convert. There is ample room. The head is well constructed. Each panel can be removed easily for access. The mixing boxes are heavy drawers lined with highly reflective aluminum so the light distribution is very even.

Also there was no guilt. The CLS 301 really sucked as a 10x10 head since it was designed for the 5x7 138 chassis and fitted, as an economy head for 10x10, to the 184. It only cranked out a whopping 600 watts of quartz halogen light. So it was well under powered for 8x10.

The basic lighting design is 3 arrays of 6 green Luxeon III LED's, (18) and 2 arrays of 6 royal blue Luxeon III LED's, (12). The power supply is a dual channel 24 volt regulated DC. Each array, per electrical specifications, was terminated to ground with a 1 ohm resister. The array calculated to 1 watt dissipation, but I used 5 watt power resisters. It is cheap insurance.

So each array looked like this:

24VDC => +O-+O-+O-+O-+O-+O- 1 ohm 5 watt resister.

I built color coded wiring harnesses and used terminal blocks to facilitate assemble, testing and repair.

The LED layout was inspired by Huw, but I slightly modified the layout to better accommodate the 10x10 head. I hand-wired my board, the LED's attached with high thermally conductive adhesive to an 0.1625" thick aluminum board with the finned heatsinks attached with the same high thermally conductive adhesive. A good daughterboard with traces would have greatly simplified the assembly and cut my LED losses.

I built color coded wiring harnesses and used terminal blocks to facilitate assembly, testing and repair.

More to follow,
Geary

[Geary Lyons]

The LED’s were laid out on a 10MM grid. Huw calculated this pattern to provide even illumination for his DeVere 504 5x4 conversion using Luxeon I’s. I used the essential layout with the Luxeon III'x to accommodate both 5x7 and 10x10 mixing boxes on the CLS 301.

[ic-racer]

My soldering experience comes from tube audio equipment repairs and and upgrades . I needed to develop far more finesse in handling this miniature buggers.

Cheers,
Geary

Way off topic: Cool, I just finished a 100w tube amp project.

More on topic: I see that your system is based upon using the existing head's mixing box. Do you have any thoughts on how a 'cold light' replacement system might work. Specifically, many more LEDs would be needed to cover the 10x10 area. I could see this as being problematic in many respects from getting the leds close enough to each other, heat and evenness of illumination. Do you think it would be feasible?

[Eric Woodbury]

This seems like a very interesting project and more feasible with these new high power LEDs, but I'm concerned about getting very high contrast, 4+, from LEDs. The high power ones that I saw were 450 nm. Can you get a grade 5?

As far as a coldlight retrofit, I was wondering about a hybrid system: using a blue tube so that one could hit the grade 5 and then green LEDs. I had a dual grid VCL8100 for a while and it was a good idea, but not very bright, especially the green. For now I'm happy with filters.

[Ash]

I'd love to remove the bulk and heat from my Dichroic Chromega F. Thanks :)

[Geary Lyons]

Way off topic: Cool, I just finished a 100w tube amp project.

More on topic: I see that your system is based upon using the existing head's mixing box. Do you have any thoughts on how a 'cold light' replacement system might work. Specifically, many more LEDs would be needed to cover the 10x10 area. I could see this as being problematic in many respects from getting the leds close enough to each other, heat and evenness of illumination. Do you think it would be feasible?

Off topic: What power & driver tubes? I'm running modified 6L6GC/6SN7 based single ended mono-blocks, about 80 watts each.

On topic: It really doesn't matter whether one is replacing condenser, diffusion or coldlight heads. I started this project as a custom "built from scratch" LED head. I was going to use a 6" depth head box on an base of .1625" aluminum sheet with 2 1"x1/4" reinforcement bar running the length to strengthen the cantilever .

The Durst 184 uses the bottom of the head/mixing box unit as the upper guide for the neg carrier. The head/box are cantilevered off of the "camera" head sub platform. I decided to use the Durst head box because it is very well constructed, it eliminated any engineering concerns on the neg carrier-to-head fit and I had it!

The LED layout was developed by Huw and fit into a very shallow depth box. IIRC the depth between the LED's and diffusion sheet was only 40mm. The distribution was very even. I checked distribution at varying heights and it was always even across the projected image within .1 stop, (.1 stop is the limit of my meter).

The Luxeon's lambertian radiation pattern is very tight and predictable. My original plan was to increase the LED grid spacing from 10mm to 20mm and put the LED board into a head/box that was 6" tall. I was going to line the box with highly reflective material. This would give me both direct and reflected fill to the diffuser. In the "final" head, I left the grid at 10mm because I have the 5x7 mixing box for the Durst. The 10 mm spacing is tight for hand soldering, but very doable.

The grid I posted earlier has 12 royal blue and 18 green LED's. This works for both 4x5 and 10x10. You do not need more LED's for 10x10, as long as you are using high lumen LED's with tight wavelength spec's. The Luxeon's, (I's for 4x5, III's for 5x7 -10x10), fill these critical requirements nicely. As I posted earlier, the LED head, as built, tested out as 2 3/4 stops faster, (hard and soft channels tested individually!), than my 1000 watt dichroic head in max yellow, (soft) and max magenta, (hard).

Heat has not been an issue. I built in plenty of heat sink area, .1625" thick aluminum main board, .1625" aluminum daughter board and approximately 5.5"x 7" X 1" finned area. The fins and and the LED's are permanently bonded with a special highly heat conductive 2 part adhesive. The main and daughter boards have highly heat conductive grease between them. I also have the fan available from the Durst Lacobox. Simply put, massive overkill!

I plan to test heat rise at the LED base, but haven't got to it yet. But, the "touch test" after about 5 minutes with both channels continuously on would indicate that there is no concern!

I will not get the chance for serious printing until after the July 4th holiday. So far all of my playing around with the new head has been positive and just reward for the cost and some of the development/build frustrations.

Cheers,
Geary

[vann webb]

Wow. Thank you for posting this info, Geary. I would love to build one of these for my own use.

[Geary Lyons]

The LED’s were laid out on a 10MM grid. Huw calculated this pattern to provide even illumination for his DeVere 504 5x4 conversion using Luxeon I’s. I used the essential layout with the Luxeon III'x to accommodate both 5x7 and 10x10 mixing boxes on the CLS 301.

A good question was raised. Who is Huw?

Huw is an Enlish engineer who first published enlarger head projects using LED's. His url is http://www.huws.org.uk/ . His last project was to convert a DeVere 504 to a LED head with built in microprocessor and integration of the control electronics to be driven by an RH Designs analyser. Far more sophisticated then my split grade design parameters!

Cheers,
Geary

[ic-racer]

[QUOTE=Geary Lyons;364969]Off topic: What power & driver tubes? I'm running modified 6L6GC/6SN7 based single ended mono-blocks, about 80 watts each.

This one is a Guitar amp.
http://ceriatoneforum.com/index.php?topic=201.0

Again thanks for posting your experience, it is inspirational!

[Geary Lyons]

The saga continues!!!! I bumped the Hard contrast range into Grade 5 by adding 4 1watt, near UV/Violet LED’s, (400nm). (See the grid attached for placement.) I ran step wedges to test the contrast. I used Ilford MGIV RC. Looks like I am getting about ISO 65, so well into Grade 5. (See wedge attachments.)

VC paper was designed for tungsten light sources and is sensitive to below 400nm wavelengths, that is, sensitive to UV light. Tungsten and fluorescent light are broadband and extend into the UV range. The very tight spectral response of the LED’s requires more thought to the paper sensitivity and contrast range. One of the concerns in using LED’s was the inability to get to Grade 5. The Phillip’s Royal Blue produced the hardest contrast, but, in my tests, only got into the lower Grade 4 range.

We all must be concerned with using the near UV wavelengths. There are some basic safety procedures to avoid issues. Do not look directly at the lit LED, even for a few moments. The spectrum of these LED’s is very narrow with the peak at 400nm. This is really violet and the short wavelength at which light is normally categorized as “visible” for the human eye. But never forget that you only have 1 pair of eyes and always proceed with due caution

The Soft Grade is very extended, about ISO 215, Grade 00 with MGIV RC. I will run some tests with non-extended range papers, as well. I am using a 21 step wedge. I would like to try a finer gradation, as I narrow down the spectral response.

The Hard is now about .5 stops faster than the Soft. I may try to achieve a better balance by removing some of the Royal Blue LED’s. Although the near UV LED’s are not “as powerful” , in wattage, the shorter wavelengths are higher in energy.

The head is really nice to print with, fast, quiet and cool! The elimination of the fan noise and the head heat is a true joy. I did not realize how much I had “environmentally internalized” all that goes with pouring 1000 watts of mostly heat into the DR.

I have ordered some small modular boards that use the new SMD Rebel LED’s. This will allow easier scalability. The work continues and the LED budget starts to look like the current US Federal budget!

[ic-racer]

Excellent, thanks for posting your results. Just to clarify, are you using a conventional timer and two exposures, like 'standard' split grade printing?

[Geary Lyons]

Excellent, thanks for posting your results. Just to clarify, are you using a conventional timer and two exposures, like 'standard' split grade printing?

I am using the RH Designs StopClock Pro. It provides a 2 channel f-stop timing routine for split grade printing. Most any timer will work. I prefer to work work with f-stop timers rather than linear.

Cheers,
Geary

[Peter De Smidt]

Geary,

This is a great idea, and you executed it very well!

[Turner Reich]

Do you think that you will share the actual making of the light source or are we to admire you and keep complementing you on your progress?

Have you ever hear the phrase
if you're going eat candy in front of everyone then bring enough for everyone?

[Steve Goldstein]

Thanks for posting the LED maps. I've always been fascinated by the LED distribution that Huws calculated and that you followed. One of these days I'll have to verify it for myself. It's not intuitively obvious that it can provide even illumination. How high above the diffuser do you need to place the light source? My guess is there should be a minimum height below which the nonuniformity becomes noticeable, though with most normal subject material it might be hard to detect this. Have you experimented with this? It might be easiest to detect using the highest contrast settings, and might be different with the UV diodes active or not. Just four UV diodes doesn't seem like enough to be uniform over the whole field. I guess you could look for it by contact-printing step wedges in various positions on the paper using identical exposures.

BTW, who makes the UV LEDs? Could you post part numbers for all the LEDs at some point? And have you considered adding red LEDs for focusing?

[Geary Lyons]

Steve,
Good questions!

How high above the diffuser do you need to place the light source? My guess is there should be a minimum height below which the nonuniformity becomes noticeable, though with most normal subject material it might be hard to detect this.

Huw used a height of 40mm, from the plane of the LED's to the diffusion plate, in his Luxeon I 4x5 head.

Have you experimented with this?

My iniatal go was to use the 3 watt Luxeon III's, double the horizontal spacing and increase the height from LED's to diffuser to 80MM. This worked quite well. I did some quick measurements with a light meter at the baseboard level and the light was very evenly spaced. I think the greatest fall-off was about .2 EV in the far lower right-hand corner.

It might be easiest to detect using the highest contrast settings, and might be different with the UV diodes active or not. Just four UV diodes doesn't seem like enough to be uniform over the whole field. I guess you could look for it by contact-printing step wedges in various positions on the paper using identical exposures.

The spatial displacement of all of these LED's is a Lambertian pattern. Typical LED's project a very tight beam with most of the light concentrated within +/-15-20 degrees from the center axis. The LED's that I used project a fairly wide beam, providing 80% relative intensity at about +/- 50 degrees from the center axis. (See attached chart). The "usable" intensity of the light increases with the decrease in wavelength. You need fewer Royal Blue than Green and fewer UV than the Royal Blue. For a lot of reasons, I want to have just enough of the near-UV to push up the contrast.

I like your idea of testing evenness of the UV with step wedges. I will try that! My light meter really doesn't respond to the UV band.

And have you considered adding red LEDs for focusing?

Initially I included 2 red Luxeon III's. My observation is that you really do not need the RED for focusing. Most of the high powered white LED's are just a mix of blue & green. I focus with all of the LED's on. It is very bright and under safelight illumination, quite white looking. The red LED's have different electrical characteristics requiring a separate array. I may add them back in the final go, but more for critical positioning of image, but I really have had little need for this.

BTW, who makes the UV LEDs? Could you post part numbers for all the LEDs at some point?

I used:
Philips Luxeon III
Royal Blue LXHL-PR09
Green LXHL-PM09
Red LXHL-PD09

SuperBright LED's INC.
400nm I watt XR7090UVV

I do not consider this board my final product. It really is a working mule, though quite usable as the end product. Philips have introduced a surface mount device LED, Rebel, that is very small and a bit more powerful than the Luxeons at 1/3 the forward voltage. I have ordered some small circuit boards upon which to mount the Rebels. This will provide modularity and scalability.

Cheers,
Geary

[Skorzen]

Just a heads up to anyone considering using Luxeon LED's, while not shown in their pricing there is a substantial discount for orders of 5+. Right now the Luxeon III's are showing up as 5.8852, if you order 5 the price becomes $4.5908 at 10 it becomes 4.2038. There is another break at 50 and 100, but I expect that most people here would be buying smaller quantities. There is a similar break on the other Luxeon LED's (Rebel, I ect) that I looked at.