DIY light meter for darkroom

[radiant]

I've been experimenting building my own light meter for darkroom printing. Idea is to get the time to a "ballpark" and maybe reduce one test strip in the process. I've been using digital light sensor which should have "light ranges from 188 uLux up to 88,000 Lux".

For example on my Fujimoto G70 on f2.8 aperture on 50mm nikkor lens, distance 450mm, I get lux values ranging from 0.066 to 0.8 lux.

However I'm struggling how to process this information. If I convert this information to EV and assume paper "film" ISO is about 8, I get totally wrong exposure times. Let's just throw that the EV is -2 at that 0.6 lux. That gives 30 seconds exposure time on f2.8 at film ISO 100. Multiply that to whatever "paper ISO" you want and you end up to many minutes exposure time, which is totally wrong.

Is there a problem that I'm measuring illuminance and I should calculate luminance some how?

Any insights how to use lux meter for darkroom printing would be really appreciated

Datasheet for the sensor I'm using: https://cdn-shop.adafruit.com/datasheets/TSL25911_Datasheet_EN_v1.pdf

 

[koraks]

Are you sure you're controlling the sensor in the way you intend it to and correctly interpret its output? Troublesooting this requires at least a look at your microcontroller code.

 

[radiant]

Are you sure you're controlling the sensor in the way you intend it to and correctly interpret its output? Troublesooting this requires at least a look at your microcontroller code.

The lux amount is correct. When pointed out of window (now overcast) it shows 2000 lux which is in EV 9-10 and it's quite correct.

 

[bernard_L]

You are doing the exposure calculation as if you were taking a picture, in which case, for the same, for a fixed LV (not EV, btw) , the higher the f-stop number, the longer the needed exposure. In your case, the f:2.8 aperture of your enlarger is irrelevant in the calculation, because you are measuring the illumination on the baseboard. If you were to close your diaphragm, that would be directly reflected in the measured light level, without any need to make that aperture intervene a second time (in your calculation).

Rephrasing, your measurement is the analogous of a TTL measurement with the stop-down method; in that case the meter and lens construction are simpler, because there is no need to have an aperture coupling from lens to meter: the light behind the lens is what it is at the stopped-down aperture, the very same one that will be used for exposing the film; no linkage or calculation needed.

Given your stated goal "Idea is to get the time to a "ballpark" and maybe reduce one test strip in the process" I would not bother with "first principles" calculation and just calibrate against a known good print; your device can also help in selecting the proper paper contrast. But beware of a "technical" approach to print grade and exposure; it is the viewing experience that counts, possibly with the lightest print regions not being white, etc...

 

[radiant]

You are doing the exposure calculation as if you were taking a picture, in which case, for the same, for a fixed LV (not EV, btw) , the higher the f-stop number, the longer the needed exposure. In your case, the f:2.8 aperture of your enlarger is irrelevant in the calculation, because you are measuring the illumination on the baseboard. If you were to close your diaphragm, that would be directly reflected in the measured light level, without any need to make that aperture intervene a second time (in your calculation).

Given your stated goal "Idea is to get the time to a "ballpark" and maybe reduce one test strip in the process" I would not bother with "first principles" calculation and just calibrate against a known good print; your device can also help in selecting the proper paper contrast. But beware of a "technical" approach to print grade and exposure; it is the viewing experience that counts, possibly with the lightest print regions not being white, etc...

But of course. I don't understand why I've been thinking it as taking picture. Of course I'm just measuring light and the paper doesn't know the aperture at all, it's reacting just to the amount of light. And yes I should written at least EV(100), my bad!

My next test will be logging the lux value all around the baseboard and plotting it to excel, maybe some kind of historgram style. Then compare the print exposure time to this table. Maybe I can see the contrast grade for the negative too from this data, who knows? I just need to do quite a few prints to get enough data.

So maybe the best way is to do this in really practical way - I have touched some theories and those are really hard to implement / test that I have implemented the therories correctly.

If anyone is interested, I can share the data here too.

 

[bunip]

the ilford EM 10 works very well doing this job. 10 euro used from the bay

 

[Chan Tran]

The lux amount is correct. When pointed out of window (now overcast) it shows 2000 lux which is in EV 9-10 and it's quite correct.

When you said you point the meter out of the window you're not measuring lux. You are measuring luminance.

 

[btaylor]

the ilford EM 10 works very well doing this job. 10 euro used from the bay

+1 for the simplicity of this solution. I use one. It saves time and paper.

 

[radiant]

+1 for the simplicity of this solution. I use one. It saves time and paper.

Checked the manual and yes that helps in the darkroom certainly. That kind of meter can be achieded with lux sensor too easily so maybe I have to test similar approach: measure "correct" exposure for dark end of the print (bright on the board) and make that lux reading my target in future prints - if I understood correctly?

Cheapest EM10 is 15 euros + 12 euros shipping, so not exactly 10 euros - but still quite cheap!

 

[Chan Tran]

Checked the manual and yes that helps in the darkroom certainly. That kind of meter can be achieded with lux sensor too easily so maybe I have to test similar approach: measure "correct" exposure for dark end of the print (bright on the board) and make that lux reading my target in future prints - if I understood correctly?

Cheapest EM10 is 15 euros + 12 euros shipping, so not exactly 10 euros - but still quite cheap!

I am all for building your own meter knowing that it doesn't save any money but I think your approach is not right. I have never seen a darkroom meter that can actually measure in lux. All of them are relative meter that is you have to calibrate them for a light level that give you a certain print density. None of them would tell you how many lux that light level is.

 

[radiant]

I am all for building your own meter knowing that it doesn't save any money but I think your approach is not right. I have never seen a darkroom meter that can actually measure in lux. All of them are relative meter that is you have to calibrate them for a light level that give you a certain print density. None of them would tell you how many lux that light level is.

My idea is try to replicate the Ilford measurement method like it is. Meaning mine would be "calibrated" too. The sensor I am using is not a "lux" sensor, manufacturer writes following:

"TSL2591 is a very-high sensitivity light-to-digital converter that transforms light intensity into a digital signal output.. The device combines one broadband photodiode (visible plus infrared) and one infrared-responding photodiode on a single CMOS integrated circuit. Two integrating ADCs convert the photodiode currents into a digital output that represents the irradiance measured on each channel."

Lux is calculated from two different irradiances. If I say I will measure the light with photodiodes, does it change the game? The irradiances and lux is just a scaled output of the actual photodiode output, I believe.

Ilford devices way is really simple and probably enough effective for my goal. I can of course buy one from ebay but as I have my device up and running, it's worth trying to find these things out. Just to learn and just because I can

 

[Chan Tran]

My idea is try to replicate the Ilford measurement method like it is. Meaning mine would be "calibrated" too. The sensor I am using is not a "lux" sensor, manufacturer writes following:

"TSL2591 is a very-high sensitivity light-to-digital converter that transforms light intensity into a digital signal output.. The device combines one broadband photodiode (visible plus infrared) and one infrared-responding photodiode on a single CMOS integrated circuit. Two integrating ADCs convert the photodiode currents into a digital output that represents the irradiance measured on each channel."

Lux is calculated from two different irradiances. If I say I will measure the light with photodiodes, does it change the game? The irradiances and lux is just a scaled output of the actual photodiode output, I believe.

Ilford devices way is really simple and probably enough effective for my goal. I can of course buy one from ebay but as I have my device up and running, it's worth trying to find these things out. Just to learn and just because I can

When I said the EM-10 doesn't measure in lux because it doesn't care how many lux the light intensity is. It simply compares the intensity of the light to a previously calibrated value. Since the output of your TSL2591 is a digital value you can use whatever you use to read it to null the value. It actually easier than you think but it does require to make test prints.

 

[radiant]

When I said the EM-10 doesn't measure in lux because it doesn't care how many lux the light intensity is. It simply compares the intensity of the light to a previously calibrated value. Since the output of your TSL2591 is a digital value you can use whatever you use to read it to null the value. It actually easier than you think but it does require to make test prints.

For calibrating the meter I need to do test prints, no doubt about it. After that maybe refining test strips only when changing negative?

I tested the EM-10 logic with my DIY meter and I found it problematic because all my enlarger lens apertures are "stepped" - I can adjust only full apertures and that's too inaccurate for getting things work. There's just too big jump betweem apertures. Usually I do my test strips in 1/3 or even 1/4 steps. Half f-stop is too much. I don't understand how EM-10 can be used with stepped aperture control. Also I read this and I wasn't that sure is that the correct way either: http://www.film-and-darkroom-user.o...74cff4f27003d1306dddf83951&p=6591&postcount=5

 

[MattKing]

Some of my enlarger lenses allow me to turn off the "clicks".
All of my enlarger lenses let me set the aperture at an intermediate position between the clicks - which is how I use the EM-10.
If I need to adjust that by 1/3 or 1/4 stops, I adjust the exposure time.

 

[ic-racer]

I built an enlarger meter and electronic timer back in 1974. Currently, if I need baseboard readings I use the White channel of a color analyzer. Used to be people could not give those away. The one I use now was a gift.

 

[radiant]

I built an enlarger meter and electronic timer back in 1974. Currently, if I need baseboard readings I use the White channel of a color analyzer. Used to be people could not give those away. The one I use now was a gift.

Does that cotain a photodiode + voltage meter? So you just adjusted correct light with photo diode output?

My Fujimoto is color enlarger and I think I have the color head for it. Of course I could use it like you're using it - but dialing the correct exposure with aperture seems a bit cumbersome still .. If I need to open up then I loose the sharpness (see http://vedos.tuu.fi/how-does-enlarger-lens-aperture-affect-to-print/ )

 

[radiant]

If anyone is interested, here are some measurements.

I measured aperture change in lux on different column heights. I used clear developed acetate so I didn't need to focus the image + find the same spot from the negative. So these values are close to dmax, I guess?

Y-axis is lux values . So my lux meter works correctly, it really is measuring lumens on area. There is a clear relation between lux and aperture so now I know how to continue this project!

 

[John51]

To reduce the number of test strips, I make 3 contact prints from each roll. Exposed at the time for optimum black and one stop either side of that. Makes it easier to estimate the exposure for individual prints.

 

[koraks]

@vedostuu you can't say if a certain level of illumination is dmax - it depends on the actual exposure (and of course development). Your approach seems to rely solely on light intensity, but keep in mind that exposure can be expressed as lux•seconds - in your argument, the "seconds" part is missing. This is also the reason why you're running into problems with full/half f-stops - that doesn't matter as long as total exposure is the same between a benchmark and a specific test, and you can get the exposure identical through the combination of illumination (lux) and time (seconds).
If your measured lux values are correct, is hard to say, but I venture that it doesn't matter as long as your metering device is consistent and you use the same device for process calibration and actual use.

 

[radiant]

Your approach seems to rely solely on light intensity, but keep in mind that exposure can be expressed as lux•seconds - in your argument, the "seconds" part is missing. This is also the reason why you're running into problems with full/half f-stops - that doesn't matter as long as total exposure is the same between a benchmark and a specific test, and you can get the exposure identical through the combination of illumination (lux) and time (seconds).
If your measured lux values are correct, is hard to say, but I venture that it doesn't matter as long as your metering device is consistent and you use the same device for process calibration and actual use.

EM-10 solution relies only on light intensity too. And yes the "seconds" part I'm in the search for.

The discussion has mixed a bit now.. I'm not running into problem with full/half steps at all because I'm after the exposure seconds. The EM-10 solution is running into that problem. I don't see that EM-10 doesn't work, it's probably a good way to reduce test strips!

My lux measurements are consistent and I can easily replicate the EM-10 functionality already - but that is not what I'm after right now. I want to point my lux meter to lightest area (or some other "standard" lightness where I have measured good exposure previously) and get a rough time for exposure - no matter what the aperture or column is. This is totally possible for what I have now measured.

Could this be so simple that if I need to expose for example 8 seconds to 0.5 lux and if the lux value changes to 1.0 I need to expose 4 seconds to get same amount of exposure? Maybe I don't need to calculate this at all through f-stops! I really need to test this out, could it be so simple..

 

[koraks]

Could this be so simple that if I need to expose for example 8 seconds to 0.5 lux and if the lux value changes to 1.0 I need to expose 4 seconds to get same amount of exposure?

As long as reciprocity holds true: yes.
And within the same paper, development, contrast grade etc, of course. The contrast thing will be your next challenge. You'll have to work out what you want to calibrate for exactly, and how many calibrations you want to do. Dmax, middle gray and zone IX for each paper and eachgrade you use? Something like that? Sounds like quite an endeavor!

 

[radiant]

As long as reciprocity holds true: yes.
And within the same paper, development, contrast grade etc, of course. The contrast thing will be your next challenge. You'll have to work out what you want to calibrate for exactly, and how many calibrations you want to do. Dmax, middle gray and zone IX for each paper and eachgrade you use? Something like that? Sounds like quite an endeavor!

Great news! I must try this really soon, I'm so excited!

I'm pretty stoked if I get only this part working - the rest of the features can wait for sure. I don't need any paper compensation yet since I'm probably using same paper all the time. Also multigrade filter calculation is not needed - altough that would be interesting to solve too. If I can skip test strips with this feature and make directly one proof print I can save much time and trouble. I always make proof print first to see what the print needs to be done (multigrade, dodging, burning). It doesn't matter if this proof print isn't perfect, it just needs to have close enough to get overall idea of the negative and how it prints out.

Currently my lux meter shows minimum, maximum, median and mid-gray (11%) lux values when I scan the image with the sensor (it logs the readings and calculates these from the log). That information could be used to determine the needed contrast grade but I need more data first from multiple prints before I can really understand the values

 

[koraks]

Sounds like an interesting approach and I don't see why it couldn't achieve what you want it to do. There are easier ways, but it's a fun project and that in itself is of merit.

FYI: like @ic_racer, I have used a color analyser for the same purpose analysers also to determine the required contrast grade. It worked, of sorts, but in the end, a few test strips is for me a more straightforward approach.

 

[radiant]

Sounds like an interesting approach and I don't see why it couldn't achieve what you want it to do. There are easier ways, but it's a fun project and that in itself is of merit.
FYI: like @ic_racer, I have used a color analyser for the same purpose analysers also to determine the required contrast grade. It worked, of sorts, but in the end, a few test strips is for me a more straightforward approach.

I have considered buying RH Designs analyzers to do the job but sometimes it is fun to try solve the problem myself. I have done own f-stop timer with some (I guess) unique features such as multigrade-mixer timer (to help out chaning the contrast without tinkering with time at all). That kind of projects give hope that when I have established the base for this kind of project, I can innovate something completely new from the already tested base!

Could you describe how you determined the contrast with color analyser? Sounds really interesting!

btw: thanks for the supportive words!

 

[Anon Ymous]

While we are on the subject of a DIY darkroom light meter, I'd like to also ask a question. I don't own and I've never seen any of them. I suspect that they use a photodiode and there's one characteristic of the photodiodes that make things somewhat difficult. The angle at which light falls on the photodiode affects it's output. It progressively becomes lower as the angle becomes lower. The reading from the edges of the projected image may be lower than what it really is. So, do these meters have any kind of lens on top of the diode?