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

[koraks]

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.

Just take a blue power LED with a decently-sized die. That's basically the same as your white LED with the phosphor removed (again, it's not a UV LED hiding underneath the phosphor).

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.

Which for your measurement is basically irrelevant. The net efficiency is probably significantly higher with the phosphor removed. Keep in mind that a LED used as a photodiode is sensitive to wavelengths shorter than its peak wavelength when used as an emitter. So the blue diode inside a white LED is sensitive to UV, and in your use case, all the phosphor does is convert a little UV to green and red light. This is essentially lost to the sensor diode, which is blind to those longer wavelengths.
So in short, you're better off with a simple blue LED. Of course, white LEDs are plentiful, so there's a practical reason why you might prefer those; that's fair enough.

and you can add more diodes in series to get higher voltages that are less affected by the impedance

Or just build a simple transimpedance amplifier. Plenty of circuits all over the net.

there is modulation...not unexacted

Not sure what you mean here. Many light sources will indeed flicker at e.g. 50/60Hz grid frequency, or at a substantially higher frequency depending on how they're driven.

 

[imgprojts]

Just take a blue power LED with a decently-sized die. That's basically the same as your white LED with the phosphor removed (again, it's not a UV LED hiding underneath the phosphor).



Which for your measurement is basically irrelevant. The net efficiency is probably significantly higher with the phosphor removed. Keep in mind that a LED used as a photodiode is sensitive to wavelengths shorter than its peak wavelength when used as an emitter. So the blue diode inside a white LED is sensitive to UV, and in your use case, all the phosphor does is convert a little UV to green and red light. This is essentially lost to the sensor diode, which is blind to those longer wavelengths.
So in short, you're better off with a simple blue LED. Of course, white LEDs are plentiful, so there's a practical reason why you might prefer those; that's fair enough.


Or just build a simple transimpedance amplifier. Plenty of circuits all over the net.



Not sure what you mean here. Many light sources will indeed flicker at e.g. 50/60Hz grid frequency, or at a substantially higher frequency depending on how they're driven.

Alright, blue led it is...the modulation is interesting. I'll plug the LED on an oscilloscope and post a trace. There's a 60Hz signal in there but more stuff that I didn't dig into too much.

 

[koraks]

There's a 60Hz signal in there but more stuff that I didn't dig into too much.

There's likely loads of noise because it's a super high impedance circuit.

I tried this approach as a LED-as-a-sensor for a DIY UV dosimeter/integrator, but quite quickly decided it was so far inferior to a simple and cheap UV sensor that I abandoned it. The method I used did actually use an amplification circuit before feeding the signal into the A/D of a microcontroller.

As said, there are plenty of sensors that would be up to the task. One of the simplest ones is the ML8511 or GY8511, which appears on a number of DIY modules in the $5-$15 price range. The similarly-priced LTR-390 has an I2C interface and a slightly different spectral sensitivity. It also appears on several DIY-ready modules.

I'd seriously consider getting some kind of DIY UV module instead of mucking about with an LED, which is going to result in measurement problems one way or another.

 

[AndrewBurns]

I was just looking at UV sensors as I will need to include one in my UV projector system. I haven't bothered for my contact print setup because the amount of UV getting to the paper is quite consistent and so I've found just timing the exposure to be adequate. For the projector on the other hand, because I can change the enlargement and make prints of whatever size I want, the UV getting to the paper will be proportional to the degree of enlargement and can vary substantially. I could just measure the level of enlargement and have a calculation to work out exposure time, but it would be more accurate to have a UV sensor at the paper to integrate exposure.

The LTR-390 is really designed for peak sensitivity in shorter wavelengths and has almost no sensitivity in the 365 - 405nm range that most people will be printing in:

The AS7331 includes three different sensors, and the UVA one seems to have a more useful spectral response:

It's not super cheap but not too expensive either. Sparkfun has a breakout board for the sensor, and it communicates over I2C as well which makes connecting it to a microcontroller or raspberry pi pretty easy.

SparkFun Spectral UV Sensor - AS7331 (Qwiic)

The SparkFun AS7331 Qwiic Spectral UV Sensor measures UV radiation on three channels, UVA, UVB, and UVC, with high sensitivity and accuracy.

www.sparkfun.com

 

[koraks]

There's quite a few UV sensors, some with I2C, some with a simple analog output. The AMS sensors (ASxxxx) are attractive; several of them will measure UV in various bands. The one you posted looks nice. They're all very well documented, which helps a lot. It takes a little getting used to working with them, but when you get the hang of it, you'll notice they all follow the same principle.