I will do my absolute best to describe everything in a non-technical way

April 30, 2019 8:43 AM

Electrical engineer Samson March made a smartwatch from scratch and illustrated it with a conversational walk through his process. (Be sure to click the "Load 27 more images" button to see the whole thing.). He posted about it on /r/diy.
posted by a snickering nuthatch (18 comments total)

According to the comments, cost of materials was around $88 USD, which with a little markup would still be cheaper than an iWatch. It would be kind of amazing if he took his design and Kickstarted a DIY hobbyist kit with prebuilt circuit boards, etc. and sold it as a fun weekend project for the teen "maker" market.
posted by Strange Interlude at 9:10 AM on April 30, 2019 [4 favorites]

Way beyond my own abilities, but I love it when people who are skilled and able to do such stuff put out things like this, because honestly, it takes a lot of the mystery away about how things are made. And it's a GORGEOUS watch, to boot!
posted by xingcat at 9:28 AM on April 30, 2019

It's a neat end-to-end project but the maker does state that to be commercially viable, it would need to cost $200-$250. That's a hard sell when Apple Watches (which do a lot more) seem to be on sale perpetually at $250, Google Wear watches are usually below $200, and things like Pebbles are less than $50.
posted by meowzilla at 9:57 AM on April 30, 2019

I paid $210 for my fitbit so that's not way out price range for a smart watch.
posted by octothorpe at 10:30 AM on April 30, 2019

Are any of these, commercial or the DIY one, water resistant?
posted by amtho at 10:32 AM on April 30, 2019

Pretty much all the commercial smartwatches are water resistant to some degree. The better ones are rated for swimming.

I don't think this DIY one is - I didn't find any reference to water, seals, glue, or rubber on the page.
posted by meowzilla at 11:04 AM on April 30, 2019

I just did a similar project with a friend's kid. It was based on a ESP32, a custom board, an OLED display, a micro USB-rechargeable LiPo battery, and a 3D printed case. I helped him design and assemble it, he wrote the firmware and apps from scratch. At 14. Kids these days. But it was a fun project!
posted by xedrik at 11:23 AM on April 30, 2019 [2 favorites]

Cool, I did one of these back in the day. I wanted Leela's this-thing-on-my-arm from Futurama. The hardest part was getting something that fit my forearm, I ended up using an archery arm guard. Couple of years later, it was easier to just strap an iPhone to my arm, heh. My name is Ozymandias, king of Kings....
posted by BeeDo at 11:32 AM on April 30, 2019 [1 favorite]

This is very cool. Reminds me a bit of "Mirror Mirror", that mirror which can also display notifications, messages, and the like.

Tangent: I'd very much like to own (or build) a CAD router. Things I've discovered:

• CAD software itself is either hugely expensive (SolidWorks or its contemporaries) or completely cloud-based, including any and all files you create (Fusion360, which is "free" for hobbyists)...I hate cloud-only shit
• CAD software requires a mind-shift and is complicated and often frustrating. Programs like SketchUp, which seemed to take some of the mystery out of CAD, are not "parametric" and cannot reliably be used to produce G-code or other "hardware ready" output
• There really isn't a great central resource for DIY plans which can be vetted before buying, as far as I can tell. It's a bit like buying a theme for WordPress. Often there's a "fancy" demo, and...maybe you'll get something decent, or maybe you'll get a flaming pile of poo
• Often plans for CNC devices require other CNC devices to make pieces for it, a bit of a catch-22
• At the end of the day, you'd have to want to build your own for the experience, as it's not going to save you a ton of money. There are Chinese manufacturers who will supply you with CNC machines with better resolution, repeatability and rigidity than what you could build yourself, for the cost (or less!) of buying the raw pieces you'd need for your home build

It's the last point which is frustrating to me at this point. Sort of like 3-D printing. You think "how cool, I can make some neat shit" and then the cost of making the thing is so much more than buying it. Sure, journey vs. destination, but it just seems odd I can buy a finished machine, painted and with all sorts of goodies for less than a pile of extrusions and some motors in a box.
posted by maxwelton at 11:40 AM on April 30, 2019 [1 favorite]

Maxwelton, a few points from a fellow enthusiast:
•Fusion 360 is free for hobbyist use and can save locally. It is daunting to start, but they have a great series of video tutorials that really are quite good. I volunteer with our FIRST robotics team, and we have taken so many kids from "never used AutoCAD" to designing robot parts and sending them to the ShopBot in just a week or two.

•A lot of DIY CNC plans are... yeah. But I have built several of FolgerTech's kits, and they are well documented and put together, and the support is great. Their Fortis kit is a great place to start. Their kits are complete, but you'll need basic tools, and usually minor soldering skills.

•If you want to build a from-scratch CNC device and need parts made to do so, your local maker-space would probably help, or often if you ask nicely on related message boards. Most 3D printing / DIY CNC hobbyists love to geek out about the hobby, and will gladly help other people get started. I only have a smallish (5x5x8) CNC router at this point, but I may be able to machine some parts, if they're small enough.

•To your last point, the thing I like about building my own is that they are incredibly hackable and customizable. I have taken my basic Folger FT-5 printer and enlarged its build area. I've added features to the firmware. Added an extruder cooler and a finer extruder (0.3mm instead of 0.5mm). Modified the bed heater. Lots of tweaks.

DIY is not for everyone and is certainly not the easiest way to get started, but it can be very rewarding and you sure do learn a lot about your device, and are able to fix just about anything that goes wrong. And, it's fun to start small, learn how it all works, and then you've got like 90% of the parts you need to rebuild it into something larger.
posted by xedrik at 12:11 PM on April 30, 2019 [3 favorites]

As far as CAD goes, I've personally been using OpenSCAD for my projects. It's got a much different approach than most CAD tools in that you describe your object in terms of unions/differences of basic shapes, define functions for repeated patterns, etc. I find it much easier to use than mouse-based CAD where I'm constantly dragging something the wrong way or fiddling with two lines to get them to connect, but I know I'm in the minority there.

But it is licensed under the GPL and exports to a large number of formats so I've used it on 3D printers and laser cutters, and I know it can be used for CNC as well.
posted by fader at 12:33 PM on April 30, 2019 [1 favorite]

And oh yeah, the way I got started with OpenSCAD was through this excellent blog post walking through everything from making simple shapes through combining them to make a 3D printable chess piece.
posted by fader at 12:39 PM on April 30, 2019 [1 favorite]

If you happen to already have a 3d printer, the Mostly Printed CNC is a decent project. At least I think it is. I have to finish mine.
posted by TheHuntForBlueMonday at 12:43 PM on April 30, 2019

I've been trying my hand at FreeCAD, which, for being free, seems pretty impressive, but I am a rank amateur and I often find myself staring at the screen going "why WHY is that unconstrained?" Which is more me than the program. I'm working my way through trying to model a complex, relatively organic shape (lots of parts which have curves in all three planes) by drawing X, Y and Z schematics for all possible edges using construction shapes and then for each part importing those as guides for corners and the like...have no idea if that's the best way. Maybe I'll try 360 and see if that's more intuitive.

Thanks for the links and ideas! (I'm someone who likes to "only buy once" but I think that has to be the wrong approach in this field. Spending a few hundred on a machine I know is not going to do what I want, ultimately, would still get me making stuff and teach me a bunch, which would allow me to know ultimately what my needs are.)
posted by maxwelton at 1:00 PM on April 30, 2019

More generally, I love projects like this. Artisanal electronics -- whether modded or custom build -- are amazing. People who ask the question "Android or iOS" and think "Screw it, I got this. They won't make what I want anyhow." They are the very antithesis of modern consumer electronics.

In that vein, here are a few recent favourites of mine: The Ono-Sendai Cyberspace 7 (reddit). The Kosagi Novena which in it's first iteration was intended to be bound into a book. This lovingly rebuilt thrift store cassette player (reddit) straight to us from the 19A0s.
posted by TheHuntForBlueMonday at 1:05 PM on April 30, 2019 [1 favorite]

Meanwhile, someone else made a watch with a 6502 and a Nokia-standard LCD, prompted by the fact that the 6502 is still being manufactured, and now is available in much smaller packages than the ones in the 80s. (Well, it's not strictly a watch, running a machine-code monitor and BASIC interpreter, but that's a software issue.)
posted by acb at 3:58 PM on April 30, 2019 [3 favorites]

Are any of these, commercial or the DIY one, water resistant?

The designer was on one of the electronics subreddits and was asked this. The answer was a definite no, as still in dev stage. He had plans on potting the electronics which would help.

(...and while I'm here +1 for OpenSCAD. It's awesome and fun.)
posted by pompomtom at 5:34 PM on April 30, 2019 [1 favorite]

This is SO relevant to my interests. Thanks, Jpfed!
posted by Horkus at 7:36 PM on April 30, 2019 [1 favorite]