Making a UV projector for alt-process prints

[AndrewBurns]

Some of you may be aware, but if not for some background I've been making alt-process prints using an LCD screen as a 'digital negative' via contact printing for close to a year now. My previous thread covering this method is here: https://www.photrio.com/forum/threa...egative-in-alt-process-contact-prints.207526/

This method has been working fine, and I've been getting great results with it, however it's obviously limited by the size of the LCD screen which in my case means I can't really print larger than 8x10". Obviously a bigger screen would mean I could make larger prints, however there are good reasons that make this essentially impossible (primarily, nobody makes a larger monochrome LCD screen and colour LCD screens block far too much UV light). The end result is that I've embarked on making a 'UV projector', which is basically just a UV enlarger except with an LCD screen instead of the negative.

This isn't really a new concept, in fact older video projectors used exactly the same setup as I'm building except with white light and a colour LCD screen (or RGB lights and monochrome LCD screens). For inspiration I found this video of somebody making a DIY 4k video projector, I'm basically doing exactly the same thing except with a UV light source rather than white:



Here's a simple diagram of what I'm making, really just a condenser enlarger with an LCD screen.



This is of course not the only way of projecting a digital image, these days most projectors (particularly powerful ones like cinema projectors) us 'DLP' technology, which is basically reflecting light off an array of tiny mirrors rather than shining it through an LCD screen. The big advantage of this technology is that it's much more efficient, the energy lost reflecting off a mirror is tiny vs. the energy lost passing through the polarizing filters of an LCD screen.

I did look into this technology for UV printing but decided against it primarily due to price, a 4k resolution UV projector (which I'd consider the minimum for a decent print quality) was getting close to $10k USD. I also found out that the DLP micro-mirror chips are actually significantly limited in terms of how much optical power they can handle, particularly in shorter UV wavelengths. So to get a reasonable print time the expensive micro-mirror chip might only last hundreds of hours before failing.

UV enlargers have recently become a bit of a trend, the advent of very powerful but also affordable UV LED lights means that it's possible to get 'reasonable' exposure times (in the tens of minutes, rather than hours). The big difference between an enlarger and a projector is the LCD screen is vastly less efficient than shining light through a negative. Best-case the LCD blocks 90% of the UV light, so a projector will be even harder to get reasonable exposure times out of.

To get around this problem, I'm attempting to 'throw power at the problem" with an absurd 400W 380nm UV light source.



This light source will generate so much heat that I've decided to water cool it, and made a custom water cooling block to mount on the back of it.



The condenser lenses are a pair of large acrylic fresnel lenses I got from Edmund Optics, they were kind enough to send me the transmission profile for their acrylic and at 380nm it still transmits 87% of the light. For a projection lens I settled on an EL-Nikkor 210mm enlarger lens. Nikon's marketing information for these lenses mentions "EL-Nikkor lenses are specifically designed for ultraviolet transmission in the 350-450nm range". I'm making the frame of the projector from aluminium extrusions and acrylic sheet.



Here's a view of the LED light source through the condensing and enlarging lenses.



Today I hooked up the water cooling to the LED block and ran it up to 350W of input power (probably about 100W of radiant output @ 380nm) and honestly it's kinda scary.





The water cooling system worked perfectly, I ran the LED for a few minutes and the cooling block temperature was stable at just under 40 degrees Celsius, all of the other parts of the system are a different story! The energy absorbed by the fresnel lenses was enough to significantly heat them up, causing them to expand and buckle, and the metal enlarger lens body was quite hot to the touch after a few minutes. Putting your skin in front of the output results in a very noticeable heating effect, I assume that prolonged exposure would probably give you sunburn. As a side note, I'm doing all of this testing wearing laser safety glasses that are very effective blocking UV wavelengths!

I had already planned to include powerful blower fans to cool the LCD screen but I'll also need to include some for the fresnels as well, I think managing heat is going to be a real challange even though the LED itself is easily cooled, just due to the sheer amount of radiant energy flying around in there. I can always reduce the output power of the LEDs as well if it just proves too much, with a subsequent increase in exposure time.

As far as exposure times go, I haven't done any testing yet so I can only speculate based on some calculations, but I'd assume on the order of 20 minutes to expose an A3 paper area of traditional cyanotype formula, which all things considered is pretty fast.

Next steps will be to sort out cooling for the lenses and then build the LCD screen carrier and focus mechanism!

 

[calebarchie]

Please make sure you have an appropriate extinguisher on hand!

 

[reddesert]

How big is your LCD screen "negative"? 8 inches or so? This might be a step in the wrong direction, but once you have accepted enlarging, you might be able to reduce the size of the "negative" to make the optical system more manageable. For ex, it looks like your acrylic fresnel condensers are maybe 8-10 inches square (200-250mm). If you used a smaller negative and glass condenser lenses, you might be able to find glass condensers in the 100-150mm diameter range that are somewhat affordable and much more heat-resistant.

If you can find glass condensers made out of common crown glass like BK7, they are likely to have very good UV transmittance (BK7 is pretty good down to 320-330 nm). I don't know about the UV transmittance of plate glass, but may still be superior to acrylic. I would look for such lenses on surplus sites or similar to save money. If you reduced the "negative" to 4x5", you could use the condenser unit from a 4x5 enlarger.

 

[AndrewBurns]

How big is your LCD screen "negative"? 8 inches or so? This might be a step in the wrong direction, but once you have accepted enlarging, you might be able to reduce the size of the "negative" to make the optical system more manageable. For ex, it looks like your acrylic fresnel condensers are maybe 8-10 inches square (200-250mm). If you used a smaller negative and glass condenser lenses, you might be able to find glass condensers in the 100-150mm diameter range that are somewhat affordable and much more heat-resistant.

If you can find glass condensers made out of common crown glass like BK7, they are likely to have very good UV transmittance (BK7 is pretty good down to 320-330 nm). I don't know about the UV transmittance of plate glass, but may still be superior to acrylic. I would look for such lenses on surplus sites or similar to save money. If you reduced the "negative" to 4x5", you could use the condenser unit from a 4x5 enlarger.

Yeah my LCD screen is 10" diagonal which is roughly the same dimensions as a 5x7" negative. You're correct that the condenser fresnels are roughly 10" square and to get a conventional glass lens of that size is quite expensive.

I did look at smaller LCD screens because as you say it kinda makes sense if you're already enlarging because you can then reduce the size of your optical elements. I've mostly been focusing on screens with resolutions of 8k+ to allow for good print sizes at high resolution, and there are some ~7" diagonal screens with this kind of resolution, however there are two main downsides that I can see with going to a smaller screen.

1. The small high-resolution screens seem to have lower transmission vs. the larger ones for the same resolution. Like a 7" 8k resolution screen might transmit 4% UV light vs. 9 or 10% for a 10" 8k screen. I assume this is because the ratio of pixel area to non-pixel area gets worse as the pixels get smaller.
2. Assuming you're making the same size print and you want the same exposure time you need the same LED power, but as the LCD screen gets smaller the power density at the screen goes up, possibly to the point where the LCD screen is damaged from heating (as the vast majority of the heating will take place at the LCD as it blocks the most light) I may already be at the damage threshold for my 10" screen, I won't know until I start testing with it (and I will only test it once I have a cooling system working).

Another option I've been thinking about is using parabolic mirrors to condense the light source rather than lenses. It's possible to buy large parabolic mirrors made for telescopes that could potentially do the job, but they're quite expensive and you can't get anywhere near the same range of sizes and focal lengths as you can with the fresnels.

 

[reddesert]

The light path with mirrors is likely to be a nuisance. I don't think you'd need a parabolic mirror - condensing systems don't need to deliver great image quality, and a spherical mirror should be acceptable, I doubt the spherical aberration would cause a problem. But setting up an unobstructed path from light source to mirror to LCD screen would be. Using mirrors off-axis is likely to create some uneven imaging of the source and uneven illumination of the negative, although perhaps a combination of mirrors could be rigged to compensate.

It might be possible to use glass Fresnel lenses if you can find them big enough. Of course Fresnel's original lenses were glass, and I once saw a Fresnel-type made out of discrete prisms that was over a meter in diameter at a lighthouse museum. While searching a bit for glass Fresnel lenses, I found that they exist for use in theatrical lights, like spotlights. In retrospect this shouldn't have been a surprise - hot lights are hot, deliver substantial power, and would eat up plastic light modifiers. So that's another place to look for large glass condenser-type lenses. You'll need to have them very out of focus, but that's also true of the acrylic Fresnels.

 

[AndrewBurns]

I got myself a UVA/B radiometer recently and decided to take some measurements of my light source and optical path. I had intended to use the radiometer to measure my currently contact printing system so that I could get an estimate of the relative power of the projector, however the contact printing setup uses 405nm LEDs and the meter I got has a filter that hard-cuts anything above 390nm so I can't measure my existing system at all.

The radiometer has a maximum range of 40mW/cm^2 so I was only running the projector LED at about 60W or so, and much to my surprise it seems like the transmittance of the acrylic fresnel lenses I got from Edmunds Optics is significantly less than it should be at 380nm. I was measuring about 18mW/cm^2 on one side of the fresnel and only about 3mW/cm^2 on the other, which is something like 25% transmittance when the data they gave me indicates that it should be between 86% and 91% at 390nm! I've emailed them about this and they're going to check with their supplier to see what's going on, but no wonder the lenses absorbed so much energy that they warped and buckled when I was running at high power.

I've started contacting other suppliers to see if I can get lenses that will actually work, but I'm also starting to consider if the borosilicate glass fresnel lenses used on stage and theatre lights might be more appropriate if I can get them in the right size and focal length. They'll be more expensive and larger/heavier, but also more likely to actually transmit my wavelengths and much more robust to high power levels (even if I get acrylic that works, there's no guarantee that it won't go brittle or hazy over time from UV damage).

 

[avandesande]

You could try a 4x5 condenser set from a enlarger. It won't cover your whole LCD but it could at least work as a proof of concept. I've been working with point source enlarger and it is amazingly efficient, with a 150 watt bulb it exposes a 8x10 Adox Lupex contact paper print in about a second.