If you’d asked me a few weeks ago whether I would ever do this again, I’d have laughed (and then maybe cried a little). But as I finished off the last few units, I started thinking about how much I’d learned from the experience. I was surprised to find myself getting excited about giving it another try and seeing if I could make things run more smoothly. I’ve even, heaven help me, started designing a completely new product!

In the meantime, I hope others can also benefit from some lessons I learned while turning my hobby into a business (or at least have a laugh at my expense).

The problem is supply, not demand

Though in hindsight it seems hopelessly naïve, I thought that taking preorders was the perfect business model: it removes the need for any sort of external investment, while simultaneously demonstrating that demand exists for your product at your price point. My faith in preorders reflected a belief that selling things would be the hard part of starting a business, but in my experience that wasn’t the case: selling was easy, it was actually making things that was hard.

Murphy was right

It’s not that I thought manufacturing would be easy, it’s just that I believed I had anticipated the potential problems and addressed them reasonably well. I believed that, unlike sales, manufacturing was something over which I had direct control. And, foolishly, I believed that things would mostly go as I had planned. I was very, very wrong about all of this.
If you’ll forgive a little self-indulgent venting, here’s a selected list of problems that arose: the enclosures were delivered several weeks late; the assembled circuit boards were delivered several weeks late; the fastener supplier sent the wrong screws; my original Chinese arcade parts supplier fell through; it appeared as if my replacement Chinese supplier ripped me off, but it turned out that my bank had messed up the wire transfer; even after admitting that it was their mistake, my bank refused to return my money for several weeks; because of the delay I had to switch to a much-more-expensive local arcade parts supplier; the local supplier took weeks longer than promised; the stickers were misprinted; the reprinted stickers were misprinted; and postage rates increased dramatically, an expense that I had to swallow.
If I hadn’t taken preorders, these delays would have been frustrating, to be sure, but I wouldn’t have had customers waiting for them to be resolved, so the pressure would’ve been much lower. In several cases, being able to wait would’ve allowed me to save a significant amount of money.

The plural of “quick” is “slow”

Part of my manufacturing plan was that I made an effort to contract out all the big tasks, leaving myself just a few “quick little things” to do for each unit before shipping them. What I soon found was that a few “quick little things” soon combine, Voltron-like. Unfortunately, their combined form is not an awesome robot, but rather a much larger task. As a result, I found myself taking weeks to finish assembly, rather than the days that I had assumed it would take.

Time is inversely proportional to space

The less space you have for a task, the more time it will take to finish. I assembled all of the Pianocades in my apartment. Every spare bit of space was filled with parts, with boxes literally stacked to the ceilings. When I needed something from a box on the bottom (or the top or the middle, for that matter), everything needed to be rearranged. That takes time. Furthermore, all of the clutter also meant that I could only work on a couple of units at a time, which meant that I couldn’t take full advantage of assembly-line techniques, which also slowed everything down.

Think small

“Small” means a few things here: physical size (see the previous paragraph), complexity (see the entire rest of this post), and also number of units.

I originally wanted to do a Kickstarter campaign. I had dreams of far exceeding my goals, selling hundreds of synthesizers, and it all being awesome. My friend and business advisor Henry Faber urged me to take my own preorders so I could keep things small, work out the kinks, and grow at my own pace. I only listened because I wasn’t given a choice: Kickstarter doesn’t allow projects based outside the United States, and I couldn’t make an American partner work. Boy, was I lucky. Doing around 100 units was hard; I shudder to think about what it would’ve been like to do more.

My customers are awesome

Not all the lessons were so harsh: every time I announced a new delay I got multiple messages from my customers. They weren’t complaining, though: they were voicing their support. I couldn’t believe how nice everyone was about it! People were genuinely enthusiastic about supporting a one-man boutique business. Now, I’m not saying that I will rely on the goodwill of my customers in the future, but it was still awfully nice.

My friends are awesome

When I had fallen way behind, I had several friends who essentially insisted on helping me assemble Pianocades, and who asked for nothing in return. Nick, Shaun, Patrick, and Oren: thank you so much!

Do not start your first business at the same time that you have your first baby (or My wife is awesome)

My babies

This one should have been a no-brainer. In my defence, I had planned on shipping before my son was born, but delays in the launch meant that everything kind of happened at once. It would’ve been wise to have delayed the launch, but I was not wise. (On the other hand, if you wait for the perfect time to do something terrifying, it won’t ever happen.)

I can now tell you from experience that being up all night with a baby does not mix well with assembling electronics all day. Thankfully, I have a very supportive wife who never got too angry at me for spending time working on the Pianocade instead of helping raise my son. On the bright side, we miraculously managed to avoid getting spit-up on any Pianocades!

Entrepreneurship is just another set of problems to solve

I worried that I would hate running a business, and, much to my surprise, I didn’t (although I did hate aspects of it); In the end, it’s just an interesting new set of problems to solve, and for me, problem solving is part of the fun.

I could write endlessly about the experience, but I think those are some of the most important lessons I took out of this. If you’re interested, I’m happy to discuss more in the comments, or you can get in touch privately.

Photo by Jay Shuster

I would’ve loved to have run a Kickstarter campaign to sell the Pianocade, but unfortunately Kickstarter restricts it’s projects to those based in the United States. There are international alternatives, but I decided instead to do it the old-fashioned way, so I’m taking preorders myself. Order one now, and tell all your friends!

The Pianocade is 100% open source hardware, firmware, and software. All the code, schematics, and PCB files will be released when the product ships (if not before). In the meantime, I’m happy to answer any questions either here or via email at info@pianocade.com.

In the meantime, I was interviewed about the “maker movement” recently on The Agenda With Steve Paikin. As always with interviews, there are things I wish I had said or said differently, but overall I think it went well. (I don’t know for sure, because I refuse to watch myself on video.)

About the experience: Steve Paikin is incredibly affable, his staff are all lovely, and the day after the piece aired a stranger on the street said “Hey, you were on The Agenda last night! Good job!” which was pretty neat.

Keep reading for the full breakdown of how I did it, more pictures, and source files for the 3D models.

Printing

This was the easy part: I made the 3D models (OpenSCAD for the cufflinks, Blender for the crowbar), uploaded them to Shapeways, and had them printed in their Frosted Ultra Detail material. If you’re interested in printing your own or modifying the files, here are the cufflinks (Shapeways, Thingiverse) and the crowbar (Shapeways, Thingiverse).

Plastic 3D print

Casting

In addition to being a skilled jewellery-maker, my friend Norah professionally solders circuit boards for satellites. She is a total maker hero, and patiently indulges my requests for metallurgy and soldering help and advice. Norah pointed me in the direction of HL Casting, one of several professional casting businesses here in Toronto. The people at HL were a pleasure to work with. For about $20 each they made rubber molds of my originals. These molds are then injected with wax, which is then used to perform lost-wax casting to make the silver copies.

From 3D print to mold to finished piece

Finishing

The real work begins after the silver has been cast. At the point the pieces are whitish, have a piece sticking out from where the silver was poured (called a “sprue”), and show marks from the seam in the mold.

Unfinished cast, with sprue visible on left cufflink

Norah and I removed the sprues using a jeweller’s saw (which cuts through silver like a knife through hot butter) and filed down the marks from the mold, and polished them with a combination of tumbling in a polishing agent and detailing by hand. All told, it was several hours with of labour (including Norah teaching me what to do) spread out over three or four afternoons.

Why not just print in silver?

For the most part I’ve been happy with Shapeways’ sterling silver prints. The main place it falls short for me is pricing ($30 flat + $20/cm^3). If you’re making very small pieces, or multiple copies of the same piece, it’s much more economical to use a reusable mold. There are also other benefits to doing one’s own casting: it sidesteps Shapeways’ design restrictions for silver. That means much higher detail, more control, and multi-piece items become financially feasible. A mold can also be used for other materials, like gold, if you’re so inclined.
The downside, of course, is that it takes a lot more time and effort.
Would I do it again? Probably. I’d say the choice between casting it myself and having Shapeways do it really depends on the piece itself and my mood. Both ways have their merits, and I’m happy to have both options available to me.

My main role was in designing, manufacturing, and building small noisemakers to give out to the crowd. There were four different colour-coded notes, and during one performance the audience was prompted with coloured cards to play along.I was surprised and delighted by how popular the boards were; we gave out all 150 of them in a matter of minutes, and people were asking for them long after they were all gone. It seems there’s an untapped thirst for simple noisemakers; that might be something I have to follow up on in the future…

The circuit is just a simple 555 astable oscillator, with different resistors on each colour to create a different note. They were assembled in two marathon soldering sessions at Site 3, where I was joined by several student volunteers from Ryerson’s New Media. My eternal gratitude goes out to them! You can take a little peek behind-the-scenes in this video:

Click here to view the embedded video.

Click here to view the embedded video.

Very briefly, kiwi drive is a “holonomic” method of moving a robot, which basically means it can move in any direction without having to turn first. It accomplishes this using three omniwheels mounted at the points of an equilateral triangle. With some vector math, you can find rotational speeds for each wheen such that the robot will move in whatever direction you like. I won’t go into the theory too much here, but here are some diagrams and here’s some math (via).

In the video I’m controlling the robot with a Wii Nunchuck. That’s because my 14-year-old cousin was hanging out at the shop when I built it; he saw the controller in my parts bucket and said “you HAVE to connect the nunchuck to the robot!” So I did. I used Tod E. Kurt’s breakout board and example code.

The robot is dead simple. It essentially consists of three different parts — a microcontroller, omniwheels, and servos — and it took about five minutes to put together. Keep reading to learn how it works and download the code.

Parts

• 3 continuous rotation servos (you can use other types of motors, but this was simplest for me; I used these)
• 3 omniwheels (I used these, but using these or these or any omniwheels should work
• Arduino (or other microcontroller)
• Good power supply (if you don’t use a good one, you may get power problems or brownout resets)
• Some sort of platform (I used a scrap of acrylic)
• Double-sided foam tape (or servo brackets)
• Superglue (or screws)

Building it

1. Attach your servo horns to your omniwheels using superglue or screws. Be sure that they’re centred!
2. On your platform, mark three lines that meet in the centre of the platform and are each 120 degrees from each other. Use a protractor!
3. Find the servo positions by lining up the shaft with each of the guidelines. Then, attach the servos to your platform using double-sided tape or servo brackets.
4. Connect the servos: one wire goes to ground, one wire goes to the power supply, and one wire goes to an Arduino pin (determined in code).

And it’s built!

Code

Download some example code! This inclues three Arduino sketches:
KiwiTest1.xml — runs a test pattern, moving the robot back and forth in twelve different directions
KiwiWii.xml — the robot controlled by the Wii nunchuck. Normally it’s controlled by the analogue stick, but if you hold down one of the buttons it switches to being controlled by the accelerometer
KiwiDrive — Doesn’t do anything on its own, but contains the functions used in the other two: give it an angle and a speed, and it drives the robot in that direction. Adding rotation is (for now) left as an exercise for the reader (because I don’t need it, and am too lazy to do it myself). Hint: see this site.

I wasn’t really planning on posting about them (at least not yet), but since they’re getting some coverage under my Shapeways username, I figured I ought to at least mention it. You can buy a pair here. The pictures show them printed in unpolished sterling silver. Stainless steel is much cheaper, but I’m not sure how much detail will be retained in the print.

I’m really happy with how they turned out and I already have a bunch more 3D printed projects in the works (some of which involve the video game references that I know you love), so stay tuned!

Click here to view the embedded video.

Leigh made the plushie and I made the electronics.

For more information, including source code, visit the project page.

Click here to view the embedded video.