My DIY LED "Expansion" Board

[TCT]

Pierre you are so concerned about not being a expert, can you just please make an official challenge for a D.I.Y. DTS already!?

[IntegrityShirts]

most LEDs come with a 120 degree lens. Assuming light distribution is conical round rather than elliptical, spacing between the rows should match the spacing between the elements on the strips. Otherwise you'd have hot spots. With a meter you should be able to change the distance up and down until you get even light readings. If you are hot above the LED the glass needs to be further and if you are hot between the element move the glass closer.

here's a nice timer for anybody looking to try building one of these:
http://www.auberins.com/index.php?main_page=product_info&cPath=16&products_id=368
(sorry, just found this last night)

I think alternating strips of different wavelengths would introduce the effects similar to hot spots as you would have different penetrating and crosslinking ability between the strips. It might be worth contacting a manufacturer in China and asking them to make strips with three different wavelengths on them (5050 strips have LEDs with three sources in each LED and can be made with RGB so it might be possible to put three different wavelengths in each unit. The problem is, all the lower nm LEDs I've seen are significantly lower Wattage than the 395-405nm).

from what I can tell, PhotoPolymer emulsions are really liking 340 or so for exposure, so chances are your times would be significantly lower with lower wavelength. See chart from Ulano below:


pierre

Yes, most all these strip LED's claim 120 degree lens. I did try to lay them out "square" meaning they'd have equal distance between rows and columns.

While I do think a 5050 with 3 different wavelength leds would be the best, I don't think anyone is currently using that setup. In fact, I'd be willing to bet that the starlight uses all the same LED's at the same wavelength. Wouldn't it be easy to tell by reading the resistance value off the strip's resistors? That would net us the voltage to the LED's and give us a good guess that they're all the same. The 395nm range have 270ohm resistors. What's the starlight use?

Distance from the glass is definitely something a UV meter might help determine IF its sensor is small/narrow/focused enough to pinpoint a change within a few nudges across the top of the glass when taking readings.

When I chose the 395nm range, I used this article from screenweb to form my "opinion" on what range would work best for a cheapie diazo emulsion.

If those charted images are accurate, lowering down to a 365nm LED, should yield LONGER diazo times and shorter Photopolymer times, right?

[alan802]

May be the best DIY tool ever shared on this site.  Simply amazeballs.

[3Deep]

Well hell who do you guys think build this stuff aliens ppl build this stuff and with the right tools DIY almost has no limit...might be ugly but works.  I LIKE it!!!!!! good job IntegrityShirts

Darryl

[blue moon]

most LEDs come with a 120 degree lens. Assuming light distribution is conical round rather than elliptical, spacing between the rows should match the spacing between the elements on the strips. Otherwise you'd have hot spots. With a meter you should be able to change the distance up and down until you get even light readings. If you are hot above the LED the glass needs to be further and if you are hot between the element move the glass closer.

here's a nice timer for anybody looking to try building one of these:
http://www.auberins.com/index.php?main_page=product_info&cPath=16&products_id=368
(sorry, just found this last night)

I think alternating strips of different wavelengths would introduce the effects similar to hot spots as you would have different penetrating and crosslinking ability between the strips. It might be worth contacting a manufacturer in China and asking them to make strips with three different wavelengths on them (5050 strips have LEDs with three sources in each LED and can be made with RGB so it might be possible to put three different wavelengths in each unit. The problem is, all the lower nm LEDs I've seen are significantly lower Wattage than the 395-405nm).

from what I can tell, PhotoPolymer emulsions are really liking 340 or so for exposure, so chances are your times would be significantly lower with lower wavelength. See chart from Ulano below:


pierre

Yes, most all these strip LED's claim 120 degree lens. I did try to lay them out "square" meaning they'd have equal distance between rows and columns.

While I do think a 5050 with 3 different wavelength leds would be the best, I don't think anyone is currently using that setup. In fact, I'd be willing to bet that the starlight uses all the same LED's at the same wavelength. Wouldn't it be easy to tell by reading the resistance value off the strip's resistors? That would net us the voltage to the LED's and give us a good guess that they're all the same. The 395nm range have 270ohm resistors. What's the starlight use?

Distance from the glass is definitely something a UV meter might help determine IF its sensor is small/narrow/focused enough to pinpoint a change within a few nudges across the top of the glass when taking readings.

When I chose the 395nm range, I used this article from screenweb to form my "opinion" on what range would work best for a cheapie diazo emulsion.

If those charted images are accurate, lowering down to a 365nm LED, should yield LONGER diazo times and shorter Photopolymer times, right?

that's funny, as I was looking for that article this morning to reference it and could not find it. It shows the curves for both the penetration and sensitivity unlike the other graphs out there.
In a quick recap it states the light should be 360-390 for SBQ and 390-420 for Diazo.
I think 365-380-410 would probably be the best to cover all the possibilities. According to the article, you should be golden with the Diazo!
Yes, the measuring portion of the UV meter is rather small, about the size of one LED, so you should be able to dial it in rather nicely!

pierre

[IntegrityShirts]

that's funny, as I was looking for that article this morning to reference it and could not find it. It shows the curves for both the penetration and sensitivity unlike the other graphs out there.
In a quick recap it states the light should be 360-390 for SBQ and 390-420 for Diazo.
I think 365-380-410 would probably be the best to cover all the possibilities. According to the article, you should be golden with the Diazo!
Yes, the measuring portion of the UV meter is rather small, about the size of one LED, so you should be able to dial it in rather nicely!

pierre

Yeah I kinda went into this with that article as a basis for my LED choice, not so sure that was a good idea or whether it is truly accurate across all diazo/photopolymer emulsions.

After some more testing of exposure times vs. emulsion thickness, I'm not so sure these LED's are the best solution.  They just don't seem bright enough to fully penetrate the emulsion in a timely manner. By timely I mean faster than my 5k watt Olec. If the emulsion is a thicker coating it is PAINFULLY slower than the Olec to get a full exposure. You can see the outline of the film pretty well which tells me not nearly enough light got through to the squeegee side.

 I knew going into this project that the LED choice was going to be the hurdle, as the rest is easy wiring. Think I'll request a couple samples from different manufacturers to do some more testing, which will delay a true verdict for a while.

[IntegrityShirts]

UPDATE with new LEDs and power supply!

I opted to order some more chinese 395nm LED strips and power them at opposite ends and power EACH row individually to avoid voltage drop on the thin strip material. Then, to further avoid voltage loss, I bought an adjustable power supply to pump these up to 13.8V to make sure they weren't thirsting for voltage.

The results? Well, pretty much the same: 150S with a 1/2 coat round edge Aquasol HV is right at 40 seconds for a good exposure.

Is it a good exposure? Yes, on Aquasol HV, it's a good solid exposure similar to 40 light units on the Olec 5k.
Is there some undercutting? Yes, a little. I need to move the LEDs closer to the glass but have to clearance some of the wiring loom at the ends to get them up higher which I think will reduce the undercutting.

Are these the best LEDs? NO. I'm not sure you can reliably get chinese UV LEDs that perform as indicated. It's a crap shoot.

[ebscreen]

I have the same DMM.

Barring more sophisticated test equipment have you thought of using the photocell from you Olec integrator
to compare output? Obviously the two light sources would be measured differently but you'd at least have a comparison.

[Northland]

Nice.... well done.  If the exposure time is the same, I'd be tempted to dial back the voltage.

When I turned up the voltage on mine, they got pretty warm & that's the number one cause of failure (overheating of the internal diode joint).

[Gilligan]

from what I can tell, PhotoPolymer emulsions are really liking 340 or so for exposure, so chances are your times would be significantly lower with lower wavelength. See chart from Ulano below:


pierre

Am I reading this wrong or wouldn't the 350nm be just the perfect wavelength for everything? (if you had to pick just one)   It hits SBQ at 87%, Fe+++(whatever that is) at 52% (not great but WAY better than 5% that the 405nm hits it at), Diazo at 93% and Diazo/AC (whatever that is) at 70% (again, better than 45% that the 405 hits it at).

If you had to pick ONE wavelength what would it be?

[IntegrityShirts]

I don't think that chart is accurate given what abuffington has posted since this thread started.

I was originally trying to figure out which wavelength to buy but I think the quality of the Led is just as important.

[blue moon]

from what I can tell, PhotoPolymer emulsions are really liking 340 or so for exposure, so chances are your times would be significantly lower with lower wavelength. See chart from Ulano below:


pierre

Am I reading this wrong or wouldn't the 350nm be just the perfect wavelength for everything? (if you had to pick just one)   It hits SBQ at 87%, Fe+++(whatever that is) at 52% (not great but WAY better than 5% that the 405nm hits it at), Diazo at 93% and Diazo/AC (whatever that is) at 70% (again, better than 45% that the 405 hits it at).

If you had to pick ONE wavelength what would it be?

that chart is only half of the picture (go figure!). Turns out there is penetration and activation. First one has something to do with how well the light gets through and the other is what light activates the best. They are quite a bit apart and it seem the best spot (in between both) is in the 385 or so range for most emulsions. I'll see if I can dig up the other chart.
pierre

[blue moon]

from what I can tell, PhotoPolymer emulsions are really liking 340 or so for exposure, so chances are your times would be significantly lower with lower wavelength. See chart from Ulano below:


pierre

Am I reading this wrong or wouldn't the 350nm be just the perfect wavelength for everything? (if you had to pick just one)   It hits SBQ at 87%, Fe+++(whatever that is) at 52% (not great but WAY better than 5% that the 405nm hits it at), Diazo at 93% and Diazo/AC (whatever that is) at 70% (again, better than 45% that the 405 hits it at).

If you had to pick ONE wavelength what would it be?

that chart is only half of the picture (go figure!). Turns out there is penetration and activation. First one has something to do with how well the light gets through and the other is what light activates the best. They are quite a bit apart and it seem the best spot (in between both) is in the 385 or so range for most emulsions. I'll see if I can dig up the other chart.
pierre

OK, read this:
http://www.screenweb.com/content/tips-optimum-screen-exposure

[IntegrityShirts]

Yeah I keep re-reading that article over and over again haha. Well, if anyone out there wants to try this project, I'd suggest going with a 385nm or lower LED strip to see how it compares to the 395nm strips North and I have already tested with similar results. I have a feeling the 10nm lower strips might be the ticket to a quicker complete exposure.

If I can source a non-shady supplier, I may give it another go here this month as it is the slow season now.

[mimosatexas]

I know this has been discussed here and there, but I would love to start a comprehensive thread about these DIY LED builds with standard formatting and an updated first post with details of each build and some kind of consensus summary.  After the new year I think I am going to try my own build and would like to further the development and discussion rather than rehash others mistakes and what not.

Can you detail the following in this format, and note after each what you might change or like to see tried for comparison:

LED Details:
-Source
-Cost (in total/per LED/shipping/etc)
-Wavelength (and any other details related to the light output itself)
-Electrical requirements/wiring diagrams, etc.

Build Details:
-Overall Size
-LED Layout (distance between bulbs,distance from glass, etc)
-Glass (thickness, type, source, etc)
-Peripheral info related to vacuum blanket, etc

Film/Emulsion/Exposure Details:
-Type of emulsion and coating technique
-Film brand and printer output
-Exposure Times and mesh used

Random thoughts, what you would improve, etc: