Using an LCD screen as a 'digital negative' in alt-process contact prints

[AndrewBurns]

Yeah I'm not ready to give up just yet, I still have a few more things I want to try and I had a similar battle with the light field correction that now works quite well. I think the hardware and software parts aren't insurmountable but the physical and chemical processes of actually producing the print are the really complicated and difficult parts.

I've been thinking with cyanotype in particular the chemical coating changes colour quite significantly during the printing process based on level of exposure, and I wonder how much the particular colour of the chemistry at a given time effects how much UV light it absorbs and as such the rate of development. So the paper initially starts at a light yellow colour, which you would imagine would be quite absorbent of UV light, then in the mid-tones changes to a blue colour that I'd think would be more reflective of UV, and finally the paper 'inverts' to a green-brown colour which would again tend to absorb more UV. So the rate of change of tone at constant UV irradiation would initially be rapid, then slow, then rapid again, which would tend to explain the S-shape calibration curve I use.

In any case I really like using the LCD screen as a digital negative, that side of the process works very well and has merit, it's just unfortunate that the particular type of LCD you want (monochrome and high resolution) is only cheeply and easily available in quite small sizes, necessitating the tiling. If I could buy the same type of screen but large enough to expose an A3 sheet without tiling, even if the resolution were the same, I'd do that in a heartbeat rather than having to go to all this effort. Maybe one day that will be possible, or the DLP route will become cheap and high-resolution enough to be practical as well, but for now this still seems the best shot.

 

[koraks]

I've been thinking with cyanotype in particular the chemical coating changes colour quite significantly during the printing process based on level of exposure, and I wonder how much the particular colour of the chemistry at a given time effects how much UV light it absorbs and as such the rate of development.

This is usually referred to as 'self masking' behavior and it's a notorious parameter in salted paper and carbon transfer prints. I'm sure it does 'something' in cyanotype as well, indeed.

If I could buy the same type of screen but large enough to expose an A3 sheet without tiling, even if the resolution were the same, I'd do that in a heartbeat rather than having to go to all this effort.

I wonder if you could manage to somehow get hold of a bare 4k computer monitor LCD.

 

[AndrewBurns]

I wonder if you could manage to somehow get hold of a bare 4k computer monitor LCD.

The issue with that is any lcd screen with an RGB colour filter blocks substantially more UV light than the mono screen (particularly the R and G pixels), and the mono screen is already letting through only 10% of the UV at best. Also the UV degrades the colour filter layer quickly, the mono screens are supposed to last 10 times longer when used in 3D printers.

Would be nice if I could convince an LCD manufacturer to make me a special one-off but I'm sure there'd be a price tag attached to that! Maybe one day the hobby 3D printer market will demand larger resin printers and economies of scale will make it happen..

I ran some quick numbers for the 4k DLP projecting onto an A3 sized area and it works out to around 230 dpi at that scale, which is actually pretty reasonable despite the relatively low resolution. A high-res DLP modified with a UV light and exposing the whole print at once is probably the most likely/logical path forward for digital alt-process printing at this point IMO. Still continuous tone though, don't think there's any way of making reasonable half-tone exposures without re-creating a direct to plate image setter, and that's probably beyond a hobby project...

 

[koraks]

Hmm, good points, I hadn't thought about the RGB filter layer yet. Come to think of it, additional layers in the sandwich may be an issue, too. For a one-off LCD, you'd be looking at some kind of prototype supplier/workshop. Not economically viable, I'm afraid.

DLP projection remains an interesting proposition, given a sufficiently powerful light source and a suitable lens. I think for direct digital imaging, it's probably the way to go at this point, not in the least because it also rids you of much of the issues related to collimation (since the projector 'manages' beam shaping to begin with) and screen-to-print contact. The flexibility in magnification is also nice. I do have some doubts about sharpness in the corners, but maybe that's not so bad in practice.

Edit: I wonder whether it might be feasible to do a spin on the laser projector concept, using a laser source and two mirror galvos. The main issues I can think of are obviously the safety aspect as well as a sufficiently powerful UV laser, although I imagine the latter should be available in some way.

 

[calebarchie]

Still continuous tone though, don't think there's any way of making reasonable half-tone exposures without re-creating a direct to plate image setter, and that's probably beyond a hobby project...

Has been done long ago exactly as a hobby project but DTF not DTP(?). Look up photoplotters, which are still in commercial use.

I wonder whether it might be feasible to do a spin on the laser projector concept, using a laser source and two mirror galvos. The main issues I can think of are obviously the safety aspect as well as a sufficiently powerful UV laser, although I imagine the latter should be available in some way.

Was attempted with a polygon mirror setup which would be even faster, did not eventuate I think due to the same kind of issues here.

C

 

[AndrewBurns]

Yes I have a 2-axis galvo and f-theta lens assembly here that I bought for a film scanning project that never happened. I was thinking about it and looked into it, but it comes back down to questions of power and data throughput. When you're talking about reasonably high DPI values and you want printing times in the minutes and not hours range you start to get into the 100+ kHz range for modulating the light source and 'placing' dots, which makes the electrical hardware and firmware side a bit trickier. You also start needing laser powers in the tens of watts optical output range to keep up with the scan speed, which is basically impossible without going to a fibre laser source.

There are UV galvo laser systems available at a hobby level, and software to drive them that could probably be used for photo printing, but even the cheapest ones are several thousand USD and I couldn't guarantee they'd be able to do what I want them to do at the end of the day.

 

[AndrewBurns]

I made a vernier scale pattern which I've been using to try to fine-tune the alignment of my exposure tiles. The following prints were made as two exposures, with the overlapping region running horizontally along the middle of the frame. One exposure draws one half of the scale, and the other the other half. The large markings are every 10mm, and the small marks are every 1mm. The vernier scale indicates 0.1mm increments, and the lines are 3 pixels wide (about 0.14mm).

As you can see over the three prints, I gradually dialled in the alignment, the third is the best I can do, within 1 pixel in each direction. (They're cyanotype prints, I just converted the photos to B&W for clarity).







I just made a test print strip of a photo with the improved alignment and it's definitely made the blended region better, but it still needs some tweaking to be acceptable.

 

[AndrewBurns]

Well this is the best result I've managed to get trying to tile together multiple contact prints, and it only 'works' because the photo is so busy that the small line along the overlap is less visible (but still there if you look for it).



Ultimately I'm going to abandon the concept of tiling multiple exposures, it's just too inconsistent. It seems that with no overlap it's impossible to stop the paper moving a tiny bit between exposures which results in a super-obvious seam. And with overlap I haven't managed to get the curve linear enough for it to work properly. The system works perfectly fine for A4/8x10 sized contact prints which is what I'll keep using it for (I want to give gold-toned Kallitypes a try next).

I've already started shifting my focus towards making a digital UV projector for larger alt-process prints, but I'll make a new thread for that when I have something to show.

 

[mshchem]

Well this is the best result I've managed to get trying to tile together multiple contact prints, and it only 'works' because the photo is so busy that the small line along the overlap is less visible (but still there if you look for it).



Ultimately I'm going to abandon the concept of tiling multiple exposures, it's just too inconsistent. It seems that with no overlap it's impossible to stop the paper moving a tiny bit between exposures which results in a super-obvious seam. And with overlap I haven't managed to get the curve linear enough for it to work properly. The system works perfectly fine for A4/8x10 sized contact prints which is what I'll keep using it for (I want to give gold-toned Kallitypes a try next).

I've already started shifting my focus towards making a digital UV projector for larger alt-process prints, but I'll make a new thread for that when I have something to show.

Digital UV projector sounds promising. You might need to get an aluminized Nomex suit and welders goggles if you intend on dodging and burning

 

[AndrewBurns]

Digital UV projector sounds promising. You might need to get an aluminized Nomex suit and welders goggles if you intend on dodging and burning

Haha yeah exposure times would be long (in the 20-30 minute range probably) so you'd also need a lot of patience! But for real, I'll probably be getting close to the power throughput damage threshold of the LCD screen to get exposure times to a reasonable length.

 

[mshchem]

Haha yeah exposure times would be long (in the 20-30 minute range probably) so you'd also need a lot of patience! But for real, I'll probably be getting close to the power throughput damage threshold of the LCD screen to get exposure times to a reasonable length.

I have an old Arkay Speed Dodge contact printer that has 20 independently switchable Argon lamps. It emits a violet color light that's got some UV too. It's too fast for enlarging papers, like 1 second. Azo type contact papers are usually 2-3 seconds. I think these were used by publicity and marketing departments to churn out hundreds of 8x10 "glossies" . Max print size is 8x10

Good luck with your projects!!!!
Best Mike

 

[AndrewBurns]

Found something useful to do with the big stack of failed prints I had with lines down the middle, turn them into covers for sketch pads that I can give to people for Christmas! Basically just taking the cover off a stapled-together sketch pad and saddle-stitch binding on my failed prints instead.

 

[MattKing]

Great gift!

 

[koraks]

@AndrewBurns I love that - looks great!

 

[Graham06]

You could just buy a bigger lcd screen. Here is a 14” 16k screen https://www.aliexpress.us/item/3256...05008251776483&gatewayAdapt=glo2usa4itemAdapt

Dunno if that is the largest.

It occurred to me that I could construct a beseler 45 head adapter to mount my uv lamp in my enlarger. I have a 150w led lamp from Amazon and I see they have 300w now.

I just bought a10” 8k screen and just got it to show usable pixels this evening. The display setting are not yet right but I should be able to enlarge a 4x5 crop of my 10” screen without the pixels becoming too large (someone in that thread said 16x20 on their 8x10 enlarger works fine) If I was only going to enlarge things the 6.9”? 9k screen would have been my best option

 

[Graham06]

You could just buy a bigger lcd screen. Here is a 14” 16k screen https://www.aliexpress.us/item/3256...05008251776483&gatewayAdapt=glo2usa4itemAdapt

Dunno if that is the largest.

It occurred to me that I could construct a beseler 45 head adapter to mount my uv lamp in my enlarger. I have a 150w led lamp from Amazon and I see they have 300w now.

I just bought a10” 8k screen and just got it to show usable pixels this evening. The display setting are not yet right but I should be able to enlarge a 4x5 crop of my 10” screen without the pixels becoming too large (someone in that thread said 16x20 on their 8x10 enlarger works fine) If I was only going to enlarge things the 6.9”? 9k screen would have been my best option

Reading the middle of this thread I see you have enlarger ideas

 

[koraks]

You could just buy a bigger lcd screen.

He'd also have to make or buy a bigger light source. He's working with one he's already got. I think point of the exercise was to try and keep using it.

 

[AndrewBurns]

You could just buy a bigger lcd screen. Here is a 14” 16k screen https://www.aliexpress.us/item/3256...05008251776483&gatewayAdapt=glo2usa4itemAdapt

Dunno if that is the largest.

It occurred to me that I could construct a beseler 45 head adapter to mount my uv lamp in my enlarger. I have a 150w led lamp from Amazon and I see they have 300w now.

I just bought a10” 8k screen and just got it to show usable pixels this evening. The display setting are not yet right but I should be able to enlarge a 4x5 crop of my 10” screen without the pixels becoming too large (someone in that thread said 16x20 on their 8x10 enlarger works fine) If I was only going to enlarge things the 6.9”? 9k screen would have been my best option

Yeah my current screen is already effectively 14", and the extra resolution isn't necessary when contact printing. You can now buy 16" mono LCD screens but I figured that wasn't a big enough difference to what I currently have to be worth it, and as Koraks mentioned my light source is only as large as my current screen.

It's annoying/unfortunate that all of these mono screens tend to be a very wide aspect ratio, while most of what I print tends to be either 2:3 or 4:5, which means that you're never utilising the full screen size or resolution. This is less of an issue with an enlarger/projector design as you can simply design the system to only illuminate and focus on the area of the screen you actually care about, but for contact printing there's not much you can do.

Yes I'm currently working on a UV projector, which is more complicated than contact printing but a lot more flexible as well. I think if a mono LCD screen as large as maybe 22" or 24" existed I would probably be happy with that size of print, but then of course getting a collimated light source to cover such a large area semi-evenly is a bit of a challenge as well.

 

[avandesande]

This is my dream panel. I've tried to contact the manufacturer but haven't had much luck. It's got everything- it's 8bit, square pixels, high resolution, normal aspect ratio. It will fit in 8x10 enlarger.
I don't know if it is real or just a specification someone if fishing with....

https://www.panelook.com/DBM141M12K01-14-1inch-12K-MONO-TFT-LCD-11520-8640-MODULE-51PIN-MIPI-for-3D-printer-LCD-display-detail_183822.html

 

[koraks]

It will fit in 8x10 enlarger.

Image area is slightly bigger, even, but I don't doubt that it'll fit in a modified 'film' carrier.

I don't know if it is real or just a specification someone if fishing with....

Hard to tell, of course, but I don't see why it wouldn't be real.

 

[imgprojts]

Yeah my current screen is already effectively 14", and the extra resolution isn't necessary when contact printing. You can now buy 16" mono LCD screens but I figured that wasn't a big enough difference to what I currently have to be worth it, and as Koraks mentioned my light source is only as large as my current screen.

It's annoying/unfortunate that all of these mono screens tend to be a very wide aspect ratio, while most of what I print tends to be either 2:3 or 4:5, which means that you're never utilising the full screen size or resolution. This is less of an issue with an enlarger/projector design as you can simply design the system to only illuminate and focus on the area of the screen you actually care about, but for contact printing there's not much you can do.

Yes I'm currently working on a UV projector, which is more complicated than contact printing but a lot more flexible as well. I think if a mono LCD screen as large as maybe 22" or 24" existed I would probably be happy with that size of print, but then of course getting a collimated light source to cover such a large area semi-evenly is a bit of a challenge as well.

I've been using a large venue projector from eBay. Its 1080p. I can print using the "new cyanotype" formula within 15 minutes or so depending on the overall shade of the image. I'm venturing into ferric gum and ferric carbon with some success. But man, the cyanotypes kick quite a bit of butt. Is since the new cyanotype formula has such a long continuous tone, I've been playing around with the way I expose. I've settled on a threshold exposure. Basically I go to gimp, and preprocess my image, and there I set timing values from the threshold. I wrote a webserver that then serves a stack of images to the projector. So then you just place the sensitized paper at focus, you set your total time and hit go. And you can adjust the negative if you want to push the highlights or the shadows around. The detail is pretty good. I haven't seen other cyanotypes but I can print from a 5 second ~1"x1.5" to 24"x24". I guess I could go to 180" if I really wanted. At the 11" x 15" size I don't see the pixels. One thing has been hard... Focus. I usually use a piece of paper on a ramp and I set my focus doing that. Recently I got a "ZB2 BG3 UV IR Dual-band 380nm Pass Filter Violet Glass" Filter from EBay and I use that to focus the blue light. I've been thinking about removing the color wheel but having 2nd thoughts. I saw a guy who put together a UV projector using an LCD screen and I think that might be better? I just wanted to connect with like-minded people. Howdy.

 

[avandesande]

I'm pretty sure that with a UV COB and condenser set it can be made to work. I've been building a point source enlarger, and with a 150 watt projector bulb I can expose adox lupex contact paper in 1/2 a second. I'm stunned at how intense it is. I've ordered some UV COB to experiment with and see if it can work reasonably fast.

 

[imgprojts]

I'm pretty sure that with a UV COB and condenser set it can be made to work. I've been building a point source enlarger, and with a 150 watt projector bulb I can expose adox lupex contact paper in 1/2 a second. I'm stunned at how intense it is. I've ordered some UV COB to experiment with and see if it can work reasonably fast.

I looked into some sort of universal way to measure dosage without breaking the bank and I think I got something. We want to know the intensity of our UV light source and compare it to other people's. LED's can produce a voltage if exposed to light of higher frequency than their color. However you don't want just any light to pass thru. Only UV. And well you could buy a UV LED I suppose. But what is very much available is regular light bulbs. Under each LED white phosphor, you'll find a UV LED. Usually burnt out lightbulbs will have many LEDs that are still working. Anyway, you just solder wires to one or more LEDs in series and monitor the voltage on a readily available digital Voltimeter or DM. If you and your buddy use the same LED from the same lightbulb model... Maybe even different ones, you should see the same voltage given that the LED is rated to the same voltage. Now you take that number and multiple it by the spot area. The resulting number can be used to compare almost any broad band system making UV. This is just for a ballpark figure. I've compared my projector output to a 50w COB for example and categorized a bunch of different lightbulbs at home to see if they were safe for cyanotype or to calculated the exposure length. Indeed as you place the LED under any UV source it glows, powering the LED under it. But normal light cannot make the phosphor glow so it is blocked. Up close to the projector's lens I measure 0.161V whereas for the 50W COB I measure 14.5V. The difference is huge.

 

[koraks]

Under each LED white phosphor, you'll find a UV LED

Blue, actually. Not UV. Doesn't matter; the principle still works.

However, there are other problems with your plant to compare light sources using a simple photodiode approach:

Firstly, when used in the way you've set it up (i.e. a multimeter connected directly to one or more LEDs-used-as-photodiodes), there are multiple issues that stand in the way of getting comparable results across different people. Firstly, the impedance of the multimeter is not fixed across all multimeters; some may present a load as low as several hundred kOhm's, while others may be up to several hundred MegOhms. Also, what you say here isn't entirely correct:

you should see the same voltage given that the LED is rated to the same voltage

No. The signal output depends as said above on loading as well as on surface area of the LED and conversion efficiency, which is moderated by aspects such as the phosphor layer, the packaging etc.

So the first problem is that your measurement setup will in fact not give comparable readings if others try to replicate your approach.

Indeed as you place the LED under any UV source it glows, powering the LED under it.

No, the UV lights up the phosphor. That's a different effect from 'powering the LED'. Both happen at the same time, although I wouldn't call it "powering the LED". The light does create a tiny little current that tries to flow between the P/N junction of the LED, and this current translates to a voltage across the junction that depends on the load.

What you've got is a poor man's photodiode without a proper signal conditioning circuit. It detects something, but with very little consistency. It's not particularly useful for anything except detecting the presence of a substantial amount of UV. It's not suitable for accurately detecting the level of UV, and it sure isn't suitable for cross-user benchmarking.

Secondly, you're referring to cyanotype as a process to benchmark with, but different processes have quite different spectral sensitivities (or conversion efficiencies). You can have a light source that gives nice and short exposures on cyanotypes, but that still struggles with DAS-sensitized carbon (this is particularly problematic when using LED light sources, which are narrow-band). The practical implication is that you would actually have to generate a spectral power plot (as opposed to a single number) in order to characterize a light source. This is evidently not possible with your single photodiode setup; it would have to involve some kind of prism that works in the UV band down to around 320nm at least, and a setup where you can accurately sample the lights of different wavelengths.

So your experiment is interesting, but it will not do what you hope it will.

PS: there are cheap UV sensing modules on the market of various kinds, ranging from fancy little chips with signal conditioning and an I2C bus that you can easily interface with using e.g. an Arduino or Raspberry Pi, down to simple UV diodes that you need to add your own signal conditioning and A/D conversion circuitry to. You might have a look at that option; some of these will give far superior results than your present setup.

 

[imgprojts]

Blue, actually. Not UV. Doesn't matter; the principle still works.

Sorry I was oversimplifying it. Agreed to the points you made. One other thing about it is that it functions as a kind of scintillator for the incoming light. The phosphor glows in UV and we don't know what wavelengths the UV diode below actually would react to. But that's the interesting thing. My first approach was to go buy a proper photodiode, then I thought, well I don't really need to compare to anyone, I just need this to work. So I thought about scratching the phosphor off. But when I realized just how hard that is, I just gave it a go without modifying the darn thing at all. But then I realized that any light from outside that makes the phosphor glow, will then emit some light of a different kind inwards to the diode. So I thought that's kind of cool, basically they gave us a diode with a filter and a scintillator. The yellow coating also acts as a filter, if you look at it thru the special blue filter, the LED's look black.

Anyway, at least for a single DM and light source the idea works:

and you can add more diodes in series to get higher voltages that are less affected by the impedance. One thing I that also found is that there is modulation...not unexacted... coming from the various light sources. So that also affects the universality of this.