As a kid I used to occasionally get to play with disposable film cameras and, as with all things optical, mechanical, and electronic, I am ever intrigued by their inner workings. More recently, I found occasion to shoot with one of these plasticky wonders again and kept the body after getting the film developed. I have taken to rescuing these disposables from the landfill recently, intending to cannibalize them for parts; this is an exploration into just one of many ways to make use of this 'garbage'.

One Person's Trash...

I have saved several disposable cameras of different makes and models, each unique in its own way. Disposable cameras make for good project fodder, containing several different useful optical elements, and occasionally some useful springs and electronics. The ones with a built-in flash are especially cool, since it is difficult to integrate a xenon flash into a project without getting it out of something else, in my experience.

Disposable cameras are often sent away for recycling, but primarily for the electronic components. My town's biggest film developing house told me that the recyclers do not value the all-mechanical ones as highly as those with xenon flashes.

...Another Person's Camera

While the electronics are fun and hold a lot of DIY project potential, I wanted to recycle the photographic optics into a homemade lens for digital cameras. This is achievable using the magic of reverse-engineering, 3D printing, and experimentation.

One thing that I discovered early on is that all the Fujifilm-brand cameras I have appear to use very similar, if not identical, optical designs despite the camera bodies and plastic housings appearing different on the outside. This means that they share the same molded plastic imaging lens, as well as similar film box designs. This is good since I can make multiple lens adaptors without needing to reinvent all of the mating surfaces each time. I only have one Kodak-branded disposable so while this may be true of that brand, as well, I wouldn't know.

I harvested the molded plastic optics from several cameras and took measurements of the mating surfaces. I consider these to be the features of the optical element that come into contact with the body of the camera, which are used to hold the optic in place. I did my best to replicate all of the relevant dimensions in CAD so that I could construct the model of the adapter around the optic.

Designing the Lens Body

To be able to create a successful lens, the primary consideration is focal length. More specifically, the flange focal distance (a.k.a. FFD, which I have discussed before), but in this case the final lens will consist of the single optical element taken from the disposable camera, making the flange focal distance equivalent to the focal length of the optic with the subject at infinity. In essence, how far the optic is held from the camera's sensor will determine whether or not the image is in focus.

Finding Flange Focal Distance

To determine the FFD, I took some measurements from the various disposables I had dissected to estimate a number based on the original cameras' designs. It seemed as though the optics were positioned approximately 28mm from the film plane in the Fujifilm disposables. This makes sense, as the images they take are fairly wide-angle, suitable of a ~28mm focal length lens. Some internet sources would claim the focal length of disposable cameras like these is around 30-32mm, so a measurement like this is in the right ballpark.

My first printed prototypes held the optic about 27.5mm from the camera sensor for reasons that will become evident later on.

Targeting C-mount

I had originally intended on targeting the Nikon F-mount ecosystem as I have designed things for this mount before, but because the disposable element's focal length is shorter than F-mount's FFD, I decided to instead target C-mount. This decision was based on multiple factors but was mainly because C-mount is:

• Easy to model in CAD (it's a simple standard thread)
• Easy to adapt to other mounts, including:
  • Nikon Z
  • Pentax Q
  • Samsung NXm
  • Sony E
• Natively compatible with Raspberry Pi Pro Cameras

With C-mount, I could adapt the lens to several of the camera ecosystems I frequently use, making it a convenient intermediary mount to design around. I have also found that, while possible to CAD and print a bayonet mount like those used by most interchangeable lens cameras, it is much easier to print the helical threads on a threaded mount like C-mount or M42-mount, then use a nice metal adapter ring to serve as the mating interface between lens and camera body. This strategy typically gives a much more robust final product with fewer manufacturing headaches. This is an instance where it is good to know the limitations of your tooling technologies and accordingly design for manufacture.

Version 0 - Prototype

Since the FFD for C-mount is 17.526mm (0.69in) it is easy to mount an optical element with a focal length of ~28mm by simply accounting for the discrepancy in the lens body model. My original design (V0) added an extra 10mm to C-mount's 17.526mm to place the optic at about 27.526mm from the sensor plane. Being slightly less than the nominal 28mm I estimated from the original disposable body, there would be room to attempt to focus the lens by partially unscrewing it from whatever C-mount adapter it is mounted on. I could then later refine the dimensions if adjustments were necessary.

The first iteration on the lens design used a threaded retaining ring to hold the optic in place, and featured a maximum-aperture design. All of the designs would be fixed-aperture since including a mechanical diaphragm would increase the cost of the project too much and one of the reasons to recycle disposable optics in the first place was to keep costs down. Using the largest aperture possible with the optic was bound to result in a soft image, but I wanted the greatest-possible depth of field in order to determine if my focal length estimates were sound. A greater depth of field would mean the resulting image would be more noticeably soft if out of focus.

This initial attempt was certainly interesting to play with, but was by no means the ideal end product. Images came out far too soft with excessive amounts of bloom, but I believe this was the result of the wide-open aperture design. This initial experiment did tell me that the FFD estimations were nearly perfect, though, so I decided to proceed with refining the design.

Test Shots

Version 1 - Refining Aperture

After my previous proof-of-concept, I decided to experiment with aperture sizes. I decided to print multiple versions of the lens mount with different diameter apertures. For the sake of simplicity in calculating the aperture values, I chose to assume that the aperture diameter is the same as the entrance pupil in the following formula:

Fstop = entrance_pupil_diameter / focal_length

So, with an estimated focal length of 28mm the following table shows the various optical arrangements I compared:

The apertures on the disposable cameras themselves seemed to be around 2-2.5mm in diameter for an aperture of around f/14-f/11, interestingly enough. The results were certainly the sharpest with the 2mm aperture, though some contrast is lost under certain conditions.

Test Shots

You can easily compare images taken at different aperture settings using the buttons below:

Image Quality Observations

While the lens would never be called 'sharp' I still think the results are interesting.

The wide-open f/3.5 version is extremely soft with exaggerated bloom around highlights in areas of high contrast, almost to the point of unusability. I still think there is a time and a place for this lomographic effect, possibly in situations where one might want to impose a dream-like quality to footage or stills. Knowing when and where to use this version is key to success, and some dehaze in post goes a long way.

The slowest f/14 version is definitely a midday-sun-shooting type of lens since it doesn't let a whole lot of light through. I think the disposables likely contained a high ISO film at 800 or so, with a similar aperture to this one, and a shutter speed of around 1/100s. While the corner sharpness is still low on a mirrorless camera, my guess is this is partially mitigated in the original cameras by the fact that the film plane is actually curved. Center frame sharpness is actually pretty good, for a molded plastic optic, which just goes to show why this is even a viable product in the first place. Oddly enough, the lens at f/14 seems to actually lose some contrast over the f/9.3 and f/7 versions under certain shooting conditions.

The f/9.3 and f/7 versions fall somewhere in between the two extremes in terms of sharpness which of course makes sense, though the f/7 is such a marked improvement over the f/3.5 that they seem like a different lens altogether. I actually like the chromatic aberration that is more apparent towards the corners in this middling aperture range.

I think I prefer the f/9.3 version, though it might be perfect to design and print an f/11. I think something around there would make for the ideal balance of speed, sharpness, and lens character that I am looking for.

Version 1.1 - Aesthetic Revisions

I also decided I did not like the external appearance of the original lens with the weird under-cut which made it look too knobby. I also occasionally experienced difficulties unscrewing the optic's retention nut, so I modeled in some knurling. The refined form lead to what I think is a much more pleasing appearance.

I may make some additional revisions to the design, such as creating an f/11 version, designing a lens hood for flare control, and adding space to mount a second optical element facing backwards behind the front element. Adding a second element in such a way would realize a symmetrical optical formula similar to those classically used in short-focus wide-angle lenses. It would likely increase sharpness while halving the focal length of the optical system, among other things.

With the major decisions made, I decided to dub this design the DIY-Dispo Lens.

Why, though?

In the process of performing this exercise I have been asked this question (by multiple people) to which my response is, quite frankly,

"Why not?"

I'm deeply fascinated by optics, lenses, and camera systems, and it seemed like a worthwhile exploration in reverse-engineering a product that works while putting my own spin on it. I have definitely had fun along the way. Many might question the sanity in reusing optics from cameras that are literally designed to be thrown away, and which are notoriously unpredictable and known to produce mediocre images at best. I found this experiment useful in informing future lens design considerations, and I think any experience in design can be made good experience; one's got to start somewhere.

Furthermore, most modern lenses use a complex and painstakingly formulated combination of multiple elements to achieve a final image that is bright, contrasty, and free of as many aberrations and as much distortion as possible. While I would love to design such a lens one day, I honestly don't have the time or experience for that, right now. Realistically, I'm working with what I've got. The DIY-Dispo's lens design here - being single-element, fixed-aperture, and fixed-focal-length - is one of the simplest out there. I didn't even have to design the optics myself, just the lens body. The only way to make it simpler would be to remove the refractive optic altogether, making a pinhole camera. That said, I am still impressed with the image quality you do get from a single molded plastic element.

While the DIY-Dispo isn't winning any awards for image quality, I don't think that was ever the point. This project has gained me some more experience in optical design, 3D CAD, designing for manufacture (3D printing), and physical tolerance management. On top of that, I find the result rather amusing and capable of some curiously creative photography. Some might say I've crafted a trash lens, but I started with trash, and I quite like the results. I do think I will iterate on this concept in the future, but for now, here are some more sample shots:

Gallery

Files For Download

Though I know others have done this before, I thought I'd include my CAD models for this project if you would like to try them out for yourself. The models are ready to print and will probably be compatible with most Fujifilm disposable camera optics, though I would really like to invite you to attempt to design your own version of what I've done here. I made design choices that make sense for me, but that doesn't mean they need to work for you. I think this is a great entry-level introduction to optical system design with a very low bar to entry and low stakes. I guarantee you will learn a lot in the design process if you've never done something like this before.

Iterate! Improve! Innovate!

Thanks for reading,
~Joseph

View this post on Instagram

A post shared by Joseph Coston (@rare.optics)