by chalcolithic on 11/9/23, 10:43 AM with 23 comments
by IAmPym on 11/10/23, 6:03 PM
You have some software experience, this is good. Assuming you have embedded software experience (which is essential for what you are doing and not at all like web programming) you have a foothold to finish this project without going completely insane.
Hardware is a completely different beast. With software everything happens in your head and when you get stuck you can usually think your way out of it. With hardware, you really do need to learn to do things meticulously and step by step. There are many things that can go wrong. You will learn a lot on this journey.
Don't be afraid to ask for help. Finding a hardware community will be essential to do this as you learn the skills you need. You're going to make a lot of mistakes, best to go in embracing it
A good place to start is to join an open source community for MIDI. Two off the top of my head is http://www.ucapps.de/ for MIDI specific hardware and https://www.electro-smith.com/daisy
A large amount of the work you will find yourself doing at the beginning is just figuring out where and who to ask questions about things like the keybed and other hardware issues. None of this is terribly 'hard' but none of it is easy. Most of it comes from experience.
It is almost always best to use someone else's product when you are designing your first. Roger Linn (https://www.rogerlinndesign.com/) gave me some fantastic advice a couple years ago when I was trying to design switch caps: "Do you want to be a company that designs switch caps or one that designs synthesizers?" and that stuck. I don't want to design switch caps, so I bought them off the shelf, contracted someone to design my own, and moved on.
So to that end... just buy a couple keybeds from Fatar or someone else, or just grab a synth you already have or buy one off ebay and harvest the keybed from that (often times much cheaper than buying direct!)
Don't be afraid to spend money on tools. If you find something difficult, like soldering, I have not once regretted spending money on better tools. They grow with you and save you hundreds if not thousands of hours of unnecessary frustration. When you know you need a tool, find a way to get it.
Maybe I should make a blog post about this... hmm
Hope that helps!
by strangecasts on 11/9/23, 2:00 PM
[1] https://www.thingiverse.com/thing:3760403
[2] https://gael.doepfer.eu/index.php/en/item/manual-fatar-61tp9
by 52-6F-62 on 11/10/23, 2:06 PM
For hardware I’d assume you’d either Frankenstein a prototype out of old parts or 3D print something. Quick google search even turned this up
https://hackaday.com/2015/07/12/3d-printed-modular-keyboard/
https://www.instructables.com/3D-Printed-1-Octave-Piano/?amp...
by brudgers on 11/10/23, 6:50 PM
Keybeds are a bit commodified. See http://www.fatar.com/
Making a case is low tech manufacturing.
And everything else - displays, silicon pads, circuit boards, etc - is clearly commodified.
—-
Relatively easy versus manufacturing anything that requires more than light manufacturing where all components can be jobbed out.
I mean we are talking single sided PCB’s with a handful of switches, pots, and encoders plus a microcontroller and display.
by sneak on 11/10/23, 12:13 PM
Hardware is a different beast from software. I would attempt to build a prototype first before trying to design for production; you need to walk before you can run.
A prototype will also help you refine the concept via playtesting.
Everything is doable! Don’t let anyone tell you differently. If you have a dream, realize it, even if it’s just a one-off.
Unlike software, if your goal is mass production, you will design at least one that you throw away and redo from scratch.
by b20000 on 11/10/23, 6:30 PM
by eternityforest on 11/10/23, 10:35 AM
The electronics and code is a 1 day to 1 week kind of thing to get started.
You will want a custom PCB, but if this is at all boutique-ish, you can probably afford the $5 for an off the shelf ESP32 or RP2040 module with USB Bluetooth built in.
I would guess the PCB fab costs will only be $30-$40 even in single device quantity for a decent sized board, less for a small one, it's insanely cheap now.
If there's any hard parts in the code it will be in whatever fancy customization settings stuff you need.
If there is no fancy UI and it's just a fixed mapping, then software should be trivial.
If this is low volume enough to use 3D printing, I wouldn't completely write off being able to make the keyboard from scratch.
I suspect but can't say for sure, that starting with only stuff a SW engineer would have, and assuming unlimited free time, and already having all the skills, a prototype would cost well under $1000, maybe even $500, with the main expensive part being a 3D printer.
If I were doing one, I would probably start by prototyping a single key in a few variants, along with the associated electronics on a breadboard.
I think I'd estimate about $160 of work and $40 of parts(Since I already have the tools) to try a few ideas over a day or so, just for some preliminary research, but a real product engineering firm could be "If you need to ask you can't afford it" kind of money....
In volumes too large to just 3D print, there's a lot more to think about, you'll want someone with molding experience, since the design rules are rather different, and the cost of failure is way higher when you're actually having to make molds.
Oh, and whatever you do: don't use a MicroUSB connector!! Please! I have seen way too much audio stuff with a delicate flimsy MicroUSB on it! USB-C or printer USB is worth it!
Also don't put too much weight on online electronics advice: A lot of it is geared towards people who want to do more analog stuff. The skills may have a lot of overlap, but the gear is a bit simpler for this kind of relatively basic digital work.
Just ignore any of those mile long essential parts and tools lists, you probably only need a quarter of that stuff.
I have no idea about the marketing side or the HR side of this though. It seems like boutique musical gear is kind of a big deal though, so I would assume it doesn't have to be ultra low cost?
I'm happy to answer any further questions, most of my background is working on ultralow budget projects for companies without much hardware experience.
Business and marketing and all that is hard, making small quantities is something you could probably do in a month or two if you were really motivated.
It really is amazing the difference in difficulty between making one or two of something and actually going to production!
Also I should probably ask the obvious question: does this need pressure sensitivity, and does it need aftertouch? If not, you can probably find mechanical keyboards keyswitches that work.
If you need sensitivity but not aftertouch, you can probably just use pairs of keys at different positions, and detect the time interval between them.
If you need both, I think I'd try Velostat rubber as the sensing element.
If you need it to be weighted and feel like a real piano, it's probably going to be way harder.