Please, if not interested in the story behind this thing, feel free to move on to Step 1 bellow.
A few years ago ( 2015 ), excited about having our first -and only- child, Anahita, I decided to do something special to celebrate and announce the news to friends and family. As an artist, inventor and engineer, I wanted to do something new and innovative, something that I could show her when she's older and tell her how much she meant to us! We also had friends and family in 5 continents, so it had to:
- Stand the test of time. ( So she could see it when she grows up! )
- Needed to be small and flat enough to fit in standard postal envelopes.
- It needed to be post-able without making security alarms going off on every conveyor belt!
- It needed to be sturdy enough to withstand the stress of an intercontinental postal journey.
- I wanted it to be unique, not a greeting card you can buy from any grocery store.
- I wanted it to be interactive ( have some sort of user input )
- I wanted it to narrate some sort of story
- It needed to be cheap to build as I was kind of struggling ( financially ) at the time!
At the time, I was in Paris, waiting for my papers to flight to St. Louis and join my pregnant wife! while there, I participated in a Hackathon and won an incubation prize-pack that included a two-month membership in a cool Maker-Space called Usine IO ( IO Factory ) which was truly like a candy-factory for me! with all sort of manufacturing machines available to members! I would be staying in Paris though, for only a couple more weeks max! so I had to start right away! I decided to go with my first idea and make a mechanical card, with user input, that would fit in an envelope and last forever!
This ended up being a very cool and lovely project, just what I needed in the difficult times of being stranded in Paris, away from my family and dealing with stupid bureaucracy!
It was such a great project that I thought, once I settled and more stable, I will share it with others! so I documented all the steps and took photos and videos, and, when, finally, was leaving the country, put the hard drive with all the documents, carefully, in my back-pack and jumped on the plane,... later-on, when I attempted to start writing the story and the making-of, the hard-drive did not work! it was saying that it's not formatted and was making a funny noise! all very sad signs of a corrupted hard-drive!
Long story short, I lost a ton of photos and videos and documents and memories ( not only this project but also captured moments of good, and bad times during the last couple of years! ), What was left was the stuff that I've shared on social media or those that were backed up on the cloud, which was, only the last part of the project!
And this is it! Here I'll attempt to re-create the making-of from what's left in this ten thousand miles journey of this project! I'll try to re-create some of the stuff that I can and for the rest, I'll do my best to explain the process.
Step 1: Sketching.
I had a generally visual idea in my brain about what I wanted to do, but because of the limitations in time and money, I had to find ways to first make it work and fast and then make it cheap, to build!
The thing about imagination is that it's cheap and fast but it's not iterative! meaning you can't build the V 2.0 of your imagination, based on its V 1.0 still in your imagination! your brain doesn't have a fixed version, a snapshot, of V 1.0 to analyze it and base the next version on it.
A simple and fast workaround is to sketch it! throwing it on paper, helps your brain to actually start making things up, based on a virtual representation of a real object that was once an imagination itself.
So you draw your first idea, then you look at it, thinking about how can it be manufactured, what can go wrong in this design, what can be improved and what material can be used to build it. you let your brain play with these variables and come up with the next version, and then, do the same thing over and over again.
After a few iterations, give it a break and get back to it in a few hours or the next day. Also, use external sources of inspirations, look at some other cool industrial design examples and see if / and how you can implement the same kind of design concepts in your idea and see how your brain surprises you with new, fresh things!
That's what I did, The first design was pretty boring, although, at the time, when I was putting it on paper looked very exciting! at first, I thought I could design a simple gear system with pieces of cardboard, but after drawing it on paper and thinking about its problems, I very quickly realized that paper/ cardboard was not the right choice, they could slip over each-other and also wear out very quickly by friction. I imagined they wouldn't be able to turn smoothly either. Then I thought about wood, and imagined -no matter how much I'd love making things out of wood- it wasn't sturdy, flexible and robust enough to withstand the pressure of a journey in the postal system, it also could have the same kind of wear and tear and friction problem as cardboards.
Step 2: Choice of Material.
thinking about different materials, I was leaning toward synthetic materials, Especially a glass-like acrylic sheet, I knew of from some of the architectural projects I've worked on in the past, as Plexiglass. Plexiglass is a trade-mark of a company selling these and soon I found out what I need is sheets of Acrylic.
Then I found out that there are two kinds, extruded and casted sheets. each with special characteristics. extruded sheets are more expensive, but they are better suited for laser-cutting as well because the thickness of the material is more homogenous, the casted sheets are stronger and clearer ( as there is less stress during the formation of the sheet ), but the thickness may be more variable and less precise. I also found out that they are kind of expensive! then, using my imagination, again, I tried to find ways to find cheap sheets somehow. after a few hours of research, I found this place, specialized in selling bulk acrylic products! I went there with the hope of finding some scrap sheets to experiment with and that ended up being my source. it was a great place with all sorts of acrylic, they had everything! I also learned right there that there is this special kind of glue that glues two sheets to each other as they've always been one! so I got the glue and a couple of different scrap sheets!
Step 3: Converting Raster Sketches to Vector Designs.
Now it was time to have fun with the laser cutter, but first I had to convert my sketches to G codes!
I've never worked with Laser Cutters before, Neither had I designed mechanical gears and machines. I had, however, a ton of experience with design tools and technics, both in 3D and 2D, I know how a laser cutter cuts, technically, and was a master at using drawing tools to produce Vector files, I knew Adobe Illustrator, Inkscape, and Corel-Draw, inside-out! Not only, how to use them, but how to make them work! I used to develop plugins for some of these tools!
So converting my sketches to precise Vector graphics, without losing precision when converting from one tool and format to others, was a breeze. There were times when some of these tools were not playing nice with others, and I simply opened the SVG file and edited some attributes manually to fix the conversion (especially from Illustrator to Inkscape, I remember, there was an issue ).
The process was pretty simple, import the sketch and use it as a background image with reduced transparency, draw the Vector are in the preferred tool ( AI in my case ), then export / convert to the Inkscape format ( which was the format accepted with the Laser Cutter Machine at Usine-IO ), double check the measurements and the color-codes and adjust/fix the values accordingly.
The laser cutter machine at Usine-IO was a really good one, you could send a file in multiple colors and ask it to treat each color differently, so for instance, Red could mean cut with a higher power (cut-through), blue could mean cut with a lower color ( more like a vector etch for text or marking ) and Green could mean Raster etching! so the idea was to put everything in one file and have the machine to pass through the design 2 times, once ( and first ) for etching graphics and once ( and second ) for cutting, which was very important decision as it would make the manufacturing process much much easier by eliminating the need to re-align the laser head in different passes and resulted in much more precision over-all.
At first, I was intimidated by the physics and maths of designing gears! so my first approach was to use 3rd party design tools and plugins to design gears with correct ratio, tooth and what not! but it was a nightmare! the free ones weren't as sophisticated as one would hope for and the commercial ones were either expensive or were completely limited in trial mode. so I did a bit of research, found out what makes good gears ( the angle of attack, the number, and thickness of teeth, the ratio and what not ) and tbh, not only it wasn't that complicated, I also enjoyed it very much!
So I ended up designing the gears from scratch in Adobe Illustrator, I would draw a circle with desired size, divide it to desired segments, extruding/beveling each other segment and smoothing out the edges, I would leave a tiny space in my design, so the teeth wouldn't fit perfectly in each-other ( Thought that may cause the system to lock ), and everything worked out beautifully! I was thinking I can introduce a bit lubricant to make things smoother, but even that wasn't necessary! It also gave me the option to make prettier and lighter gears, reducing the overall weight and letting me to cut some of the other elements inside the unused space of the larger gears.
So don't be intimidated by the amount of literature going around designing gears and moving elements and mechanical physics! most of that stuff is to perfect things and some of it may not have any effect on anything at all, except for satisfying a university that paid a Joe to produce a paper about that thing! As an insider, I can tell you that a considerable amount of these paper, may actually, be completely wrong! and for the most part, that's a normal and expected thing!
Back then, there weren't a lot of free and good tools to generate gear designs, nowadays, however, a little googling will yield more promising results! for instance I found this website very interesting and easy to use ( https://geargenerator.com ), give it a try! .. there are also other ones just Google, or DuckDuckGo, or whatever!
Step 4: Trial & Error With the Laser-cutting Machine!
At the time, Usine had four laser cutter machines, two big ones, and two smaller ones.
Turned out these machines are very popular and are booked very often! this turned out to be a problem, each of these machines has a different level of precision, power, and calibration and having to switch between them make pieces that are not compatible with each other. Also, depending on the time of the day, the settings needed to be different! What would have made a clean cut on the first batch, 8 in the morning, would over-burn the edges by 4 pm!
The other big issue and the one that took the most time to overcome was, design-adjustments to accommodate laser features and characteristics. As I said, this was my first experience with Laser and soon I learned that lasers are extremely good at following certain cuts and designs but they struggle with some others, For instance, it would be much easier and take much less time and create much less vibration, for the laser head to follow a curved angle comparing to a sharp corner, even though that curvature adds computational complexity to the design.
Also, I learned that over-all, the calibration errors are more pronounced as the head gets farther from the zero origins, so, no matter how small the error margins of a machine is, it's always worth when the head is the farthest from Zero base.
Also, if you are getting weird shapes, like a slightly elliptical circle or if the start and end of a closed path do meet, before anything else, check if your medium is secured well enough to the base ( usually a honey-comb or a metallic grid ), I used small magnets and metal weights to prevent the medium to move around when the powerful laser head makes rapid moves around the machine ( shaking everything! )
Another pro-tip would be to think about the movement of the head when designing your path, for me, this saved a lot of time and made the cuts way more precise. The thing is that, depending on the material, the heat produced by the laser may make the material expand, shrink, deform or bend a little bit, and if you have many passes at the same region this may make the effect more dramatic. keeping this in mind when drawing your lines make a much cleaner and more precise cut. It also helps the machine to manage the smoke and fumes better ( by distributing the amount of smoke at a given time ), again, resulting in cleaner and more consistent cuts and reduces the burns and mirror damages. An easy workaround for this problem is to use multiple stroke colors in the design, each for one pass of the head and assign all of them to the same power, speed, and cut settings ( only to force the head to cut a color first, then the next and so on, to distribute and dissipate the heat ). believe me, this helps a ton!
There are some other common sense design decisions that you may have to make knowing how the laser cutter works, for instance, if you have multiple nested complete cuts (closed paths shapes inside each-other) of material ( which is great to save both time and material ), you can't send them all to the machine and expect it to go through them in the right order. If the machine cuts the outer shape first, the piece may fall into the tray and not be there when the head decides to start cutting the inner shapes! again, as a simple workaround make different stroke colors and make the head to start from the inner-most shapes to the outer ones.
Step 5: Perfecting the Concept!
Making the design simpler / more minimalist:
As an architect, I learned to love simplicity!
I tried different designs with different mechanical movement principles, including a planetary gear system ( which I produced as a different project to create a limited edition business card! ), I also tried different mechanisms to transfer circular motion of the gears to linear motion for the heads.
The final design consisted of 7 gears, used to transform the movement in different directions and two slotted and routed shafts used to convert circular motion to reciprocating motion. The gear ratio was designed to make the movement smooth and user-friendly. I also exposed part of one of the bigger gears and used it as the user-input to engage and power the system.
Reducing the thickness:
First I used a 4-millimeter thick acrylic sheet for the outer layers and a 2-mm thickness for the inner layer.
I also had the outer layers etched and carved, each 1.5 mm, to create a housing for the inner layer ( the gears ) so the overall thickness would be 8 mm, but after a few experiments, I decided to use 2-mm sheets for everything and eliminate the internal carving entirely, committing to an overall thickness of 6 mm. this turned out to be a very critical decision in the success of the project as it reduced the design and manufacturing complexity and made the design more error tolerant, saving the project, as, after the final batch-cut, I realized that there was some inconsistency in the thickness of the sheets which would have caused serious problems during assembly with the previous layout.
Reducing the Layers:
All the rotary elements ( the gears ) were put in one layer, there was only one small shaft extending from one layer to another ( to convert the circular motion to reciprocating one ). if the thickness was not a concern, the ideal design would need to have 6 layers. I eliminated half of the layers by embedding their functionality in other layers with two tradeoffs, first one was that one of the outer layers would need to have a moving element ( not attached to anything ) that would fall out! I fixed this issue with a simple solution, I put transparent scotch tape on top of the moving element, putting another layer of the tape on the side and only masking the exposed parts that were facing the moving part ( so it wouldn't stick to it! ).
The other issue was that the heads ( that were moving back and forth, giving little kisses to each-other! ) ended up with no layer at all! which looked actually pretty cool, making the visuals more interesting and letting me reduce the overall size of the card ( as the heads, now, could extend outside of the card boundaries when farthest from each other ), with the tradeoff of making those elements more exposed and in-risk of breaking during the shipment. My workaround for this issue was to make the attachment of the head to the slotted square the weakest link ( connected with very little glue ) so if there were any parallel stress ( which is not very likely to happen in an envelope ) the head would detach from the shaft, rather than breaking, and this, indeed worked, and between all the delivered copies, there was only a one case (to my knowledge) were the head was not connected, and even in that case, the receiver figured they just need to glue that back!
In the original design, there were also two additional, paper thick, layers, cut from slippery plastic sheets to reduce the friction between the layers ( and also for visual aesthetics ) which were eliminated in the final production ( for the most part! ) as it wasn't really needed!
Step 6: Cutting the Final Batch
I spend 80% of this projects time on pre-production, and that was the best thing I could have done! because of that, the manufacturing and production of the final batch was a joy! I have tested and designed workarounds for everything that could have gone wrong or caused problems during production!
The only issue was to book them the machines! which you weren't allowed to book more than 1 hour per day! so I went there early morning because I knew nobody would come to Usine-IO or book the machine that early! I instead, booked my time at 10:30, and hoped nobody would decide to get out of bed earlier to come and laser cut stuff! and that turned out to be right, but even that wasn't enough time to cut and etch everything!
I ended up, going 2-3 days trying to use any single minute the machines were available to finish my job, the most difficult thing saw to calibrate the machines ( as I had to use different ones ), but by the second day, I figured ways to do that very quickly and efficiently as well! I decided to keep pieces cut by different machines separately though, as I thought ( correctly ), the laser beam itself may not have the same exact size in two machine ( due to calibration differences or different brands/build and capabilities of tubes with different powers ) which added a bit of time and complexity to the next step, which was the assembly!
In this final batch, I used only 2 mm sheets with two different colors, dark (for the inner layer) and light green ( for outer layers ), I also cut a couple more sheets just in case some of them had issues ( which they did! as an architect/engineer you should know that everything that can go wrong, will go wrong! ), I also had the gears tagged/numbered in my designs so I could use the exact gear that was cut in its exact place, which helped a lot to make things work together ( remember in one of the steps I said the error margin increases as the laser head moves away from the origin? If you use the components that were cut close to each other with each other, you can work around those errors, since the two components close to each, both will have, almost, the same amount of error, which reduces the margin of error between the two by factors of magnitude. in the other hand, if you use 2 components from two opposite sides of the cut medium, there you'll have the largest amount of difference/error )
Step 7: Assembly
IMHO this may be the part that often gets the least amount of attention and causes the most amount of effort, cost and time in the manufacturing and prototyping.
I knew that it's not going to be an easy task to assemble everything and because of that, I had this step in mind from the very first steps, especially during the design process I thought about assembly and designed the components in a way that I could make jigs for them to be able to place and assemble them fast and efficient but also very accurately!
And that's what I did! For the most part, I used the un-used pieces that were cut ( from gears and other elements ) to make jigs that used parts could slide in easily, then I would glue them and clamp them, move them out and start with the next piece! and this worked very nice, for the most part, I hade, however, to build additional jigs for the pieces that were cut in different machines and sometimes for the two far sides of the sheet for the bigger machine ( because of the error margin I talked about ).
Overall this was an easy and smooth process, it took less than a day and there were only a couple of unexpected problems like gluing a bunch of parts together out of order ( leaving a gear or something out! ) or dropping a tiny amount of glue ( by accident ) and causing the whole card to lockout, because one of the gears were glued in place ( in one case, I tried to keep rotating the gears with the hope that it prevents the glue from locking it! but not only it happened finally, but it also spread the tiny amount of glue to the whole system, made the whole thing opaque and at one moment made the whole thing locked! )
Step 8: Sending Them Out!
As I said, I cut a couple more sheets just so I could keep a few for myself and to giveaway later on! But sure enough, turned out the original list my wife had made was not "complete" she decided to add more ( 20% more to be precise! ) so I ended up sending all of them out except one*!
I got some pretty envelopes, put them in and send them out in two batches, they all were delivered intact except a few that were never delivered ( lost ) and one that was delivered with one of the heads unglued ( which the receiver had fixed easily ), Everybody either called or wrote back to tell us how much they loved them and I was really proud of my first time engineering skills designing a first time, unprecedented thing! I was thinking maybe I can commercialize this as a new kind of greeting cards but like many other ideas, it's still there to be done, I moved on to other projects/Ideas, and haven't touch it again! Maybe I'll get back to it again at some point and create a V 2.0!
*Oh! and I forgot to mention! I numbered them! I etch a number at the corner of each card, so they should be pretty collectible! and I kept the serial number #1 for myself!
Runner Up in the
Epilog X Contest