Printalyzer - Darkroom enlarging timer & exposure meter

[dkonigs]

I don't remember whether you are using the Printalyzer's meter to measure rise/fall times.

For the enlarger calibration, that's exactly what I'm doing. Remember, I just need to measure the behavior of the light as it is, not as its dimmed for a normal print exposure. So you drop the head, open the aperture, and make it as bright as possible just for the purposes of doing this calibration. I then configure the system to sample the light (using the sensor's "clear" channel) every 5ms while running the enlarger through several exposure cycles, do some basic statistics, and ultimately produce the rise/fall data at a ~10ms resolution.
I2C is absolutely fast enough for this.

(I don't hate 8-bit MCUs. I just don't think its the best idea in the modern era, to try and cram fancy floating point math into them. Which honestly isn't needed in most of the applications they'd still be used for.)

 

[koraks]

(I don't hate 8-bit MCUs. I just don't think its the best idea in the modern era, to try and cram fancy floating point math into them. Which honestly isn't needed in most of the applications they'd still be used for.)

You're right, of course. Not that eg a 328p at 16MHz couldn't handle this job - it does so with ease. But if you already have an stm324xx as your main CPU, what would be the point?

 

[albada]

I must agree about 8-bit MCUs. My LED-controller lets you set LED-powers and time, both in log2 units using look-up tables. And that's about the limit of what an 8-bitter is suitable for (an Arduino/ATmega328p in my case). I wouldn't want to make an analyzer out of one.

Here's an idea of a way to determine paper-curves without a densitometer: Have the user take a picture of test strip(s) with a DSLR or phone, and give you the jpg file via USB. As long as the strip includes paper-white and D-max, with the steps in a specified order, you should be able to figure out the tones between them, and knowing the exposure times and intensity, you can calculate the curve. Or was this idea already discussed? It's been long enough I don't remember.

 

[dkonigs]

Here's an idea of a way to determine paper-curves without a densitometer: Have the user take a picture of test strip(s) with a DSLR or phone, and give you the jpg file via USB.

No, that idea has never been discussed. However, it is probably the hardest possible method in terms of data processing. Just getting even lighting and perspective across the image is damn near impossible. Now scanning that strip on a flatbed scanner would work a lot better, and I'm pretty sure people have explored the idea of doing "poor man's densitometry" that way.
But still, I'd rather not go down the rabbit hole of fiddly image processing myself just yet. There are so many more important things to focus my time on with this project. Also, you could probably get in the ballpark by comparing the strip to some sort of reference material like RH includes. I'm also still considering building a simple reflection densitometer probe as an accessory for the device, but I haven't put much more than casual research into that idea as of yet.

 

[albada]

No, that idea has never been discussed. However, it is probably the hardest possible method in terms of data processing. Just getting even lighting and perspective across the image is damn near impossible.

If the test strip has a paper-white patch on one end and D-max on the other, your tone-scaling will automatically remove the effect of typical uneven lighting. That leaves perspective, which would complicate your patch-finder algorithm. That algo will give your 32-bit processor a workout!
I just thought of a reason to avoid jpg-files from cameras: Cameras often change light and dark areas to white and black in order to boost contrast. And they alter tone-curves in unknown ways. Realizing that, I think the densitometer probe would be best.

 

[dkonigs]

Okay, time for another teaser
I finally decided to go ahead and put my second prototype PCB into an enclosure, so I could have a dedicated "development unit" that was separate from my "darkroom unit."

Now I really need to get to work on that video actually explaining and demonstrating this project, if I can figure out how to write an intelligible script.

Of course as I've mentioned in the past, all the hardware and software is open source and I'm keeping it in this repository:
https://github.com/dkonigsberg/printalyzer
(still need to write documentation, though)

 

[albada]

The case of the right meter-probe looks injection-molded instead of 3d printed. In fact, everything about that analyzer looks professional. Are you keeping track of how many hours you're spending on this project?

 

[koraks]

In fact, everything about that analyzer looks professional.

It sure does. I remain very impressed and inspired by this approach.

 

[dkonigs]

The case of the right meter-probe looks injection-molded instead of 3d printed. In fact, everything about that analyzer looks professional. Are you keeping track of how many hours you're spending on this project?

The meter probe on the left was made on my (FDM) 3D printer.
The meter probe on the right was made by Shapeways and their big-fancy-expensive (MJF) 3D printers.
If I make these in any sort of quantity, I'm likely to do something like the one on the right, albeit from someone else with lower prices. However, I'm probably still a few design iterations away from that part being finalized.

I'm not really keeping track of my time, but I don't have much else to work on right now. (my actual photography, along with all my other out-of-house activities, has taken a nosedive this past year for obvious reasons).

My main motivation for putting together the second complete unit was finally getting firmware updates via USB stick working, which makes having a separate "in the darkroom" unit suddenly a lot more practical.

Also, there's a decent chance that the next major revision to the main unit will have a different case design that's better suited to quantity production. (Similar outer shape, but probably a U-shape sheet metal design like ProtoCase can make.)

In any case, I'm rapidly getting to the point where I need to actually track down some sort of useful consultant to advise me on how to proceed if I want to make this thing "distributable."

 

[albada]

...useful consultant to advise me on how to proceed if I want to make this thing "distributable."

I suggest:
1. Talk to the city hall of your city and ask them for the forms and procedure for starting your own business.
2. Talk to a tax expert who works with businesses about correctly handling the money-side of the business.

For all the rest, such as where to advertise, I suspect that you know more than any consultant.

 

[dkonigs]

Finally, I have a video showing off this project!
Introducing the Printalyzer


Its kinda long (like everything I do tends to end up), but I added time indices for the various sections. There are portions that could probably be expanded into standalone videos in their own right, but I wanted to just recap as much as possible in one place.

 

[koraks]

Looks like I'm going to have some popcorn tomorrow
Looking forward to watching it!!

 

[albada]

Let's review what you've accomplished:

• Designed an attractive case (using mech CAD).
• Designed circuitry and a circuit board (using EE CAD).
• Also did the above items for the meter probe.
• Written much software.
• Created a thorough 37-minute video.

How do you do it all?!

 

[distributed]

Thanks for posting the video. You have built an impressive system!

I am impressed by how well you integrated 21st century features into a device that looks - and probably feels - much like an appliance. The different visualizations, e.g. the one in test strip mode, look informative and pretty, and take away much potential for error. In my control program running on a PC I have millions of pixels, yet missed the opportunity to present as clearly. Bravo!

From a system perspective, having a USB host in the device is a very clever move. Apart from enabling all the peripherals you already mentioned, you and users will be able to add more in the future. Plus, I expect USB to be reasonably future proof!

 

[dkonigs]

So I've recently begun working on a little side project to the Printalyzer, with the intent of it being another accessory to accompany the device. However, I'm building this part with the goal that it can also be a completely independent product.

I know one of the side discussions in this thread may have alluded to the idea, and I know I keep complaining about "scope creep," but I figured this was one I really needed to give a try.

So what is this? Well, a reflection densitometer!
You see, calibrating a device like the Printalyzer to a reasonable level of accuracy kinda requires one. But the only real options out there are buying an ancient used unit (of questionable calibration and consistency) or forking out $$$ for a new Heiland unit (that almost feels too simple for its price tag).

Therefore, I'm going to attempt to build my own, and I'm going to try to make it cheap. Its going to include the bare minimum of hardware (light source, sensor, buttons, and a microcontroller) and connect via USB to some external device that can handle display and configuration functions. That external device could be the Printalyzer itself, or a computer or smartphone, depending on what software I (or anyone else) feels like writing.

As of today, I've built the first "development prototype." The goal of this device is to basically flesh out the concepts, experiment with the "shell" design, and write some of the software. There will be some changes in the final design, but I'm hoping that the core components won't change much.

This is how the components fit together... There's a PCB, a "cone" that blocks the sensor from the LEDs on its sides, and the main body:

Here's what the PCB itself looks like:
Its using the same sensor as the Printalyzer's meter probe (TCS3472) which is probably sufficient for doing "visual" (a.k.a. B&W) reflection densitometry. Its then surrounded by four "white" LEDs that cover the right portion of the light spectrum. Beyond that, there's just a couple of buttons, some driver circuitry, and a microcontroller (with USB interface) to control everything.

Here's what the top of the board looks like right now, as one would typically work with it for testing:
There's no top as of yet, and these are obviously the easiest/simplest buttons I could use. The real device will have more care and attention to these parts, but they're not important to the core functionality.

Finally, here's the side profile:
Since this will be more like a "puck" and thus won't be part of a "unit" like most densitometers, I wanted to make sure that you held it by a piece that could lay flat on the table. This will help keep the light/sensor cone as flat and low as possible, to make readings more repeatable.

As far as components, its basically a PCB with ~$10 worth of parts (more or less depending on quantity), and a 3D printed shell. I'm contemplating adding some diffusing materials inside that cone, but I haven't yet made up my mind on the specifics there. First I want to see how far I can go with just this.

Right now I can say that it powers up, doesn't let out the magic smoke, can run the LEDs, and the sensor gives reflection readings.
Of course turning those readings into proper densitometer data is the next big part of the project. But that's all software

 

[koraks]

Very neat! I've toyed with this idea myself as well, but it's low on my list of priorities atm. Let us know what your observations are!
You're quite a fan of those stm32's I see

 

[MattKing]

At first glance, I thought the PCB looks like a skateboard!
Feel free to run with that thought if you are looking for a name for it.
A very interesting project.

 

[albada]

Great! Just what the doctor ordered.
One IMPORTANT suggestion: Add a small 7-segment display showing the density the sensor is seeing.
The added cost is almost zero, and the display would (1) greatly future-proof the product by not relying on an operating system that will be incompatible with the software in two years, and (2) I suspect many darkroom workers, including me, don't want to bother having a computer or smartphone in the darkroom. USB power-adapters are widely available, so powering it solo won't be a problem.

 

[koraks]

I second the notion of a small display being a useful addition. Personally I wouldn't opt for a bulky 7-segment but one of those tiny monochrome oleds; they will fit in the device more easily, interface conveniently with I2C and cost about a dollar. And they're more flexible in terms of what you display on them of course.

 

[ic-racer]

Amazing work!

 

[dkonigs]

Right now the device I'm showing off is essentially a development prototype. Its purpose is to prove out the critical parts of the system (sensor, LEDs, microcontroller, physical design of the read head, etc). The "final" version is going to have to get bigger, if only to make it easier to hold and to add hardware to hold it together.

Also, making this thing usable as a completely standalone device would probably add more cost than many of you think. First, LCD/OLED displays aren't *that* cheap (if I source from a reliable/reputable vendor). Second, I'd probably need to squeeze in at least 4 buttons just to support configuration/calibration functions.

That all being said, there are essentially two use cases for a device like this:

1. Something cheap enough to provide as an "accessory" for the Printalyzer, where it is entirely reasonable to expect the Printalyzer to act as the "user interface" for configuring/controlling it.
   
2. A standalone device, that can send readings to the Printalyzer, but which can operate by itself.

Also keep in mind that everything I do will be documented and open source, so you really aren't beholden to "some janky old Windows program not maintained in a decade" just to be able to use version 1 w/o the Printalyzer.

Regardless, I could absolutely see making two versions of this device. Not just so I get my "cheap accessory" while still having something I can provide as a "standalone gadget", but also because there are two different sensors I'm planning to experiment with for this.

Okay, now for some technical fun...
According to my research (through ISO 5-3), there are two different types of "visual density" (B&W density) that may be applicable to a project like this:

• Standard Printing Density, centered around 570nm
• Standard Type 2 Printing Density (silver halide), centered around 430nm

I've experimentally determined that the visual channel on my X-Rite 810 densitometer is most likely reading the 570nm one, and my Heiland TRD-2 densitometer can't actually distinguish colors that should appear different on a "visual" reading.

Also, in case anyone was wondering, the X-Rite uses a separate set of photodiodes for "visual" density. Its not just getting it from combining the RGB channels. (Yes, I've disassembled the read heads from mine to see how it works.)

So sensor I've started with on this side-project is the TCS3472. Its the same sensor I use for the meter probe of the Printalyzer. Its also cheap enough to make this work. It really can't handle 430nm, but it can probably handle 570nm just fine. Its not that far from the center of what's used for "lux", so it may just be a matter if figuring out how to adjust the formula that combines its R/G/B channels to shift things slightly. It won't be perfect, as the TCS3472 isn't a proper spectral sensor, but I think it'll be good enough.

Now the other sensor I'm going to be trying out is the AS7341. That one actually is a proper spectral sensor. If I can figure out how to process its data correctly, I may be able to not only do a higher quality "visual" density reading, but I may also be able to implement "Status A" color reflection density readings. The only real downside of this sensor is that its approximately 3x the price of the other one. So while it might be a bit too much for a "cheap accessory," it may be just fine for a "standalone full-featured device."

 

[koraks]

If I can figure out how to process its data correctly

Yeah, that's a bit of a math puzzle. Its not so bad if you stick with readings from just this sensor, but it'll be slightly challenging if it needs to match up accurately with other sensors that aren't XYZ.

 

[DMJ]

Derek,
If you made a device that didn't switch mains (the enlarger), would you be able to ship it without having to go through all the regulations? This is such a great project and I'm thinking if the device only sent pwm signal or serial data (usb) to another simple small device that would actually do the switching...an arduino keeping it open source and easy to build. So we could actually buy the main device (the brain) from you and build the small one (great project for high school electronics/photography/"steam"). Then of course release, when all is clear, a fully functional device for the ones who want an all in one solution.

 

[koraks]

if the device only sent pwm signal or serial data (usb) to another simple small device

The design could be easily modified to accept power from an external supply. Something like a laptop charger or a not too flimsy wall-wart adapter.
Of course the device sending an on-pulse while it's powered down would be somewhat challenging However it's the approach I adopted for my new led enlarger light source, but it requires its logic boards to be permanently powered (standby mode). Not optimal for environmental and safety reasons of course.

 

[albada]

LCD/OLED displays aren't *that* cheap (if I source from a reliable/reputable vendor). Second, I'd probably need to squeeze in at least 4 buttons just to support configuration/calibration functions.

Small 3- and 4-digit 7-segment LED displays for clocks are cheap.
It only needs one button: ZERO. The user will first put a white section of a test strip over the sensor, and press ZERO. After that, all measurements will be white-relative, which are the densities users want.
On the Printalyzer, the user will press its Enter button to enter the current reading. Or, add an Enter button to the densitometer, making it two buttons.

Why use a color sensor? Use a mono sensor having a good green response, and illuminate the paper with green LEDs that are always on while the unit is on.
Calibrate it once in your lab. If sensors are decently consistent, that one cal will be suitable for all units. If not, you'll need a cal method in the factory, adding a few minutes to assembly-time.
You might ask Nicholas Lindan how he calibrates his meters.

An ATMega or PIC is cheaper than the STM32, and is well-suited for this light-duty chore.