|
Since around 2020, I've been building some electronic Uilleann (kinda sounds like the beginning of Illinois) bagpipes. This is bringing together a whole bunch of disciplines that I'm only 20% competent in, and I've been asking for lots of advice from HCH, so I thought y'all might like to see the whole project in one place. Uilleann pipes are pretty strongly associated with Irish music. They're played sitting down, and are sort of a more chill instrument than the Scottish Highland bagpipes most people think of when they think of bagpipes. You can also make the chanter (the pipe that makes the melody) stop playing notes. Check out this video of Catherine Ashcroft playing Táimse im' Chodladh to hear some really cool Uillleann piping. There are quite a lot of electronic bagpipes, both commercial and "open source", that other people have already made. What makes mine different, aside from me being the one to bring it from concept to final product, is that mine is going to be built from the ground up as Uileann pipes, with sensors to detect how far off the knee the chanter is, the ability to bend/glissando notes, and the eventual inclusion of "register" pipes to play chords: you can hear these chords in the second half of that video. My pipes are also going to use FM synthesis to generate tones, instead of relying on MIDI, which was not really designed with this sort of instrument in mind. Lastly, I'm going to provide at least two different intonations, so a player can flip a switch and be out of tune with every other instrument if they want. I can get into extreme detail on any of these points, depending on what people would like to know. But for now, here are some photos!  This is the Fagerstrom Technopipes. This guy did something similar to me, but clearly is better at physical design than I am. The original technopipe was a Scottish highland pipe, so the hole spacing is the same as the Scottish chanter, but this is the Uilleann version so it's fingered differently and has different tones. It also features a little switch to detect when it's on your knee, since that's important. The Fagerstrom guy is really cool, and is clearly a huge bagpipe nerd like I am. If I can make something half as good as this, I'm going to be incredibly pumped.  Here's what I have right now. This is actually playable! It's using screws for the capacitive touch sensors, and has a little display to tell you what's going on. In this photo it's in setup mode, where you can set the volume of the chanter and the drones, and change the patch of each. The pitch bending works, and it's using a Time-Of-Flight sensor on the breadboard to measure how far off the knee it is. Right now that's just a binary "on/off" reading, but later I'm hoping to use the distance to apply a low pass filter that changes the tone of what's being played for an expressiveness control.  I recently got a 3D printer, and a few weeks ago spent a few days creating a printable model of an actual chanter as measured by https://pipers.ie/. This is going to serve as the basis for what eventually becomes the "nice" version of the chanter, without a rat's nest of wires coming out the top. I typically only get to work on this project on weekends, and it's very slow-going, so if you're interested, bookmarking this thread will get you a slow trickle of updates. I'm guessing it's going to take at least an additional year before I have something better looking or sounding than the current thing, but who knows. If people have suggestions as I stumble through various aspects of this project, I would love your input. I'm professional level at the microcontroller programming, experienced amateur level at the electronics, and borderline incompetent at anything physical, like designing a button that doesn't jam or how wires under stress might break. cruft fucked around with this message at 17:23 on Jun 26, 2022 |
#
?
Apr 23, 2022 17:59
|
|
|
|
| # ? May 5, 2024 20:43 |
|
|
This weekend's work is going to be designing a PCB to house the chips I need in the body: the touch sensor (AT42QT2120) and the time-of-flight sensor (VL6180X). Both of these chips use an I2C bus, just like the Nunchuk on a Nintendo Wii. I found a bunch of eBay listings for a pretty small VL6180X, and I'm hoping when it arrives next week it will be small enough that I won't have to make the 3D printable model significantly larger in order to house it. Otherwise, I'm going to have to design my own tiny breakout for just the chip, and then run some wiring to the touch sensor board I'm definitely going to have to make. The Fagerstrom design has a long, slender PCB that runs the entire length of the instrument. It uses standoff screws to space to the PVC housing, and then has little screws going into those that you touch. It's a really nice design, but it would dictate the exact finger hole spacing, which I don't think I want to do. What I'm thinking instead is that I'll make a little board that goes at the top of the instrument somewhere, and then 3D print some sort of wiring harness to take individual wires to the position under the holes of the instrument. Then I can use screws to connect to those holes. This will make it a lot easier to tweak hole spacing without needing to order a whole new instrument, and will be easier to wire up than my current one, where I had to fish 8 wires through the PVC and run them out the top. If I do this right, the only thing coming out of the top will be the 4 wires of an I2C bus (probably a premade Qwiic or Stemma wire). But today I'm just focusing on designing the PCB for the AT42QT2120. I'm going to go with the SOIC-20 package, which I'm confident I can solder. What I'm not confident about is PCB design utilities. I know Eagle is supposed to be really good and KiCad is supposed to be not bad, but the most promising option for me right now looks like Fritzing, which is probably making people groan to read. Fritzing is right up my alley, though: I've actually designed a PCB with it before that seemed like it could actually work. And the board I need is not going to be complicated at all. It just needs to be small. This morning I'm going to attempt to turn a generic SOIC-20 part into the AT42QT2120 and then lay out the board I need.
|
|
#
?
Apr 23, 2022 18:12
|
|
|
cruft posted:snip Yup, you got me with this one. Kicad is pretty nice for home gaming. Also, I play GHB and some 440 small pipes, both quite terribly. A single tiny chanter somehow seems wrong for how complicated looking Uileann pipes look. I am interested in this project!
|
|
#
?
Apr 24, 2022 04:32
|
|
|
Update: Fritzing sucks. Crashed twice and now it won't open the file I saved with a full day's work. So now I'm learning KiCad. I think I managed to make a resistor and wire it to the QT2120!
|
|
#
?
Apr 24, 2022 19:52
|
|
|
Commodore_64 posted:Yup, you got me with this one. Kicad is pretty nice for home gaming. Also, I play GHB and some 440 small pipes, both quite terribly. A single tiny chanter somehow seems wrong for how complicated looking Uileann pipes look. I am interested in this project! I guess I should lay out how the parts all fit together!
Because the microprocessor is capable of playing exact octaves, I'm detuning the top and bottom note of the drones by a tiny amount. This prevents the emitted waveforms from lining up exactly with each other, and makes a sort of slow pulsing that is very similar to what a chorus effect does. The drones sound all the time, or they don't sound at all. Currently you have to enable or disable them in the setup menu. If it turns out you need to do this mid-tune (which I have never seen), I can add a row of buttons somewhere to do it. The register keys are typically crudely depressed with the outside of the player's right palm: the keys are pretty big. My plan here is to 3d print something equivalent and just hook them up to some buttons, probably using an I2C GPIO chip. I'm approaching this whole project similar to how someone 300 years ago would have done it: building the chanter first, then drones (called a "half set"), then the registers (called a "full set"). Right now I have code for the chanter and drones, and stubs for the registers. Once I have version 3 of the chanter built, I'll look into making them sound better, and then I'm going to have to spend a lot more time practicing! I'm at an Irish dance competition right now trying to understand how to use KiCad. I've got a schematic completed, I think. Next up: turning that into a PCB. 
|
|
#
?
Apr 30, 2022 19:35
|
|
|
I was doing pretty well laying out the PCB, getting used to how KiCad works (at least, getting used to my initial misunderstanding of how KiCad works) when it occurred to me that maybe I should check my dimensions before going any further. Ignoring the actual measured inner bore dimensions, since nobody is going to blow air through this, the inner bore of the chanter goes from 13.2mm at the bottom to 7.1mm at the top (where this needs to go). Resizing the PCB layout to 8mm, I get this:  ... welp. That white rectangle is how wide the board can be. The outer diameter at the top of the chanter is actually 14.8mm, before all the "decorative" rings and stuff are added. The current 3d model doesn't include a big ol' long thing that looks like a steel tube which is placed over the top and connects to the bag. So to look "legit" I can add quite a bit of height and probably diameter. Rotating the main chip on its side, I can probably get by with a 14mm wide PCB. So I'll just proceed with that constraint and see how tall the PCB is going to have to be, then redesign the chanter model to hold it. There's a lot of room up at the top that I can probably do something sneaky with. I could get the PCB narrower by using a TSSOP-20 package instead of this SO-20 package, but I think I'm probably pushing my boundaries enough just by committing to surface mount soldering, so this is going to have to do.
|
|
#
?
Apr 30, 2022 22:43
|
|
|
I don't know how anything works, but could you stick the PCB somewhere else on the instrument? Inside the bellows or strapped onto the bellows arm or something? Maybe that gets to be alot of wires or the buttons have to go on the PCB?
|
|
#
?
May 1, 2022 01:20
|
|
|
Kaiser Schnitzel posted:I don't know how anything works, but could you stick the PCB somewhere else on the instrument? Inside the bellows or strapped onto the bellows arm or something? Maybe that gets to be alot of wires or the buttons have to go on the PCB? Yeah, and I might have to do that. If I did, I'd need to run 12 wires instead of 4, though. There are 8 touch pads on the instrument, plus I still need the 4 to get to the proximity sensor at the bottom. The advantage to having this PCB inside is that I can run just the four wire bus from the microcontroller. The shorter run to the touch pads may also mean it's more responsive or something. Capacitance is weird and I don't quite understand it but in general with electronics, shorter wires is better.
|
|
#
?
May 1, 2022 01:34
|
|
|
Things are looking good from the PCB angle. With lots of help from the HCH electronics thread, I am closing in on a workable circuit board for the thing. It's going to be 12.7mm wide, the smallest the shop I'm using will build it. That's fine. Here's the third revision: I need to make some more changes, but this will be the basic size of the thing. In related news, the time-of-flight sensors I ordered from China arrived today, and they are fabulous! With only very minor modifications to the 3d model, I'm going to be able to slide this thing in to a slot in the instrument, no problem. They came in a pack of four, so I'll solder pins on one and drop it into a breadboard for testing.  It's time for me to start thinking about how to physically place the electronics! The sensor needs to face down and out the bore to the bottom: this is how it measures how close it is to your leg. I'm thinking I'll make a detachable part that can be screwed into the main instrument. The sensor will slide into a slot from the side and I'll need some way to lock it in: probably just cyanoacrylate glue, which works great on PCB and the plastic I'm printing with. Then the four wires will come out into a connector that can plug and unplug from the rest of the chanter. For the keys, I've been thinking it would be nice to make a printable strip with pre-drilled holes for the screws. This would let me make a handy channel for running wires down to the screw holes, and could also provide a way to run that 4-wire bus down to the bottom piece. The top of this strip will need to hold my custom-designed PCB, which will have a connector for the Arduino that will live... somewhere. I know this text is kind of vague so I'll try to model something up in the next few weeks to help visualize it. But the main thing is that so far everything seems totally workable, and I am super pumped about not having to design a PCB for this proximity sensor, because it mounts with a Ball Grid Array, and soldering those by hand is a real pain (it involves putting everything in the oven).
|
|
#
?
May 3, 2022 00:03
|
|
|
It's been a heck of a month, and since this is a hobby project, I haven't made much progress. In the last month, my Internet morse code site has started to take off, so I've focused a bit more on that. Part of the project is a Seeeduino XIAO based morse code "keyer". This actually has some crossover with the pipes project, since it's sort of a musical instrument, and needs capacative touch sensors. I made this:  Things I've learned from this:
I should really just pull the trigger on the PCBs, but it's scary to do something for the first time, and I want to quintuple-check the board before paying for a build.
|
|
#
?
May 30, 2022 01:56
|
|
|
cruft posted:It's been a heck of a month, and since this is a hobby project, I haven't made much progress. Oshpark 2 layer standard?? I’d say pull the trigger sooner rather than later so you can a) find the inevitable flipper footprint or misnamed/unintentionally joined on the schematic nets b) just get developing c) figure out what you like and dislike about your physical layout
|
|
#
?
May 30, 2022 03:47
|
|
|
Commodore_64 posted:Oshpark 2 layer standard?? I’d say pull the trigger sooner rather than later so you can a) find the inevitable flipper footprint or misnamed/unintentionally joined on the schematic nets b) just get developing c) figure out what you like and dislike about your physical layout Yeah, there's merit to this argument for sure. But the experience with the SAMD21 QTouch stuff has got me spooked now. I need to sort out how to get more than two fingers pressing at once before I move forward. So I'm going to focus on fixing this Morse code keyer. I'm now on a trip to an Irish dance competition so no progress for another week. 
|
|
#
?
Jun 2, 2022 21:50
|
|
|
I'm realizing I need to get organized about this, so I'm setting milestones in my gitea instance.
That first one is going to be kind of a sprint, but I'm tired of not having anything playable, and I promised a friend I'd send her something she could practice on, so I need to get busy. Unfortunately, in order to do this, I'm probably going to have to spend tonight and part of tomorrow fixing bugs in the morse code repeater, so I don't kill off the audience with showstoppers while I let it languish in order to focus on the pipes. I might also need to print out a couple keyers, even with the super-crowded case design I have right now. Hopefully I can get that done this weekend before I have to leave for next week's business trip. During the trip I'll fix the one remaining problem on the PCB and put in the order for it, and with any luck I can work on it next weekend. I'll use OSH Park because I have 0 other recommendations from anybody. I guess I need to order some chips, too. Digispark has been great to me so far so I'll stick with them. My hope is that by breaking this down into smaller steps, I can get the satisfaction of something people can try out, and keep myself motivated. It's certainly been a motivator for the morse code stuff.
|
|
#
?
Jun 10, 2022 21:18
|
|
|
Oshpark is great, NextPCB okay, JLCPCB is good. If you need something weird stackup wise I've given CircuitHub a try at work. Slick interface and good prices on assembly, but I've not yet got my first order back so the quality is????? You should still use Oshpark.
|
|
#
?
Jun 10, 2022 23:05
|
|
|
It's going slllooowwww. It's summer, so there are a bunch of vacations, and I guess I don't want to be indoors soldering stuff on the weekends. Anyway, I knocked out a morse code keyer. It looks like this:  I'm going to mail it to a person in California to do some testing on it with real usage (my morse code is awful) and see what I hear back. Things I've learned since last time:
I'm learning a lot for the bagpipes by doing this, but it feels like a distraction. I actually have an interest base for the morse code thing, though, so I'm going to keep working on it until I have a saleable design, then maybe I can convince somebody to make and sell them. I'm probably going to have to push all my milestones back by a month on the bagpipe. Maybe two months.
|
|
#
?
Jun 26, 2022 01:06
|
|
|
cruft posted:Things I've learned since last time: I forget, did you look into using something like pre wired JST XH cables and headers on the boards? https://www.amazon.com/SoloGood-Sil...ps%2C146&sr=8-5 https://www.amazon.com/Connector-Pr...ps%2C254&sr=8-8 https://www.amazon.com/XLX-10PCS-Ca...s%2C160&sr=8-15 https://www.digikey.com/en/products/detail/jst-sales-america-inc/S6B-XH-SM4-TB-LF-SN/1651066 https://www.digikey.com/en/products/detail/jst-sales-america-inc/S14B-XH-A-LF-SN/1651020 Heck, you might look for something similar in premade ribbon cables / flat flexible pcb cables, the ones made of polyimide for the copper substrate.
|
|
#
?
Jun 26, 2022 02:59
|
|
|
Commodore_64 posted:I forget, did you look into using something like pre wired JST XH cables and headers on the boards? This is a great idea that I'll probably go with. The board can just be something to solder the XIAO onto, and some JST headers. And I'll feel a lot less nervous about designing and ordering such a board for my very first PCB than this QT2120 breakout which may or may not need resistors. Yeah. This is a great idea, thanks! E: now that I'm thinking about it, I should try putting some resistors on the capacitive touch pins. The code has a field for what size resistor. Maybe that will help multitouch.
|
|
#
?
Jun 26, 2022 03:10
|
|
|
Wow, the thread is still open. A lot of life stuff has come up and prevented me from pushing this forward. I guess the thing about hobbies is they're the first to get cut when time is short. But I'm still working on it, and when cruft jr heads down to college, I imagine I'll suddenly have a lot more time to work on it. I did get a morse code doodad designed and manufactured. I can now whip those out and mail them to the four or five people who want one, and now I will never again have to make a PCB for the first time. The friend I said I'd send one to gave up on me. She ordered an actual half set (chanter and drones) and reminded me that the actual instrument I'm copying is this unwieldy mess of wood and brass, so maybe I should focus a little less on making a nice package and a little more on making something that exists. Not long after, my brother-in-law, who makes things for a living, convinced me that I was being dumb by trying to design something that could be mass-produced. At most I'll make a dozen. If I find a dozen people who genuinely want one of these, it will be among my most successful hobby projects ever. The MPR121 touch sensor is out of production  but I can't get the poorly-documented ATmel junk to see more than one finger at a time, so I just bought some new old stock MPR121 boards from Adafruit and if these things are super popular I'll figure out the ATmel stuff at that point.I haven't gotten audio working on anything other than the Adafruit feather trellis that locks up whenever I attach the I2C time-of-flight sensor. And I'm not even sure I was using the kinda not-made-to-be-dynamically-reconfigured PJC audio library correctly: I may have just been pushing out sawtooth waves and not true FM synthesis. So both of those are going away, and I'll just output a sine wave initially. Hopefully by... July? I'll have a working instrument again, that plays sine waves, can bend notes, and has two octaves and supports a few settings. Next step will be to get that system integrated into my 3D model of an actual chanter. I'll probably re-introduce the i2c laser time-of-flight sensor and i2c OLED display at this point since I've already coded all that up and made both central to the user interface, and they'll want a place to live in the chanter.
|
|
#
?
Apr 10, 2023 21:28
|
|
|
Going back to "screw it, do whatever works" mode was a good idea. Here's where that's led me:
Yesterday was spent struggling to make all the weird Teensy software actually work: I found workarounds to reliably upload a program to it, and got it to run a demo program with the audio synthesis library. Yay! Today I spent the whole time discovering that the VL6180X hates me. It only works with my ARM M0 boards, the Teensy just refuses to play. I had it doing some neat tricks with putting the instrument into setup mode when it was more than 5cm away from your knee, at which point the buttons become inputs... yeah. Writing it out like that, seems like a recipe for disaster during a live event. I think I'll add another touch input for setup mode. Or maybe I'll make it a switch. I've got a bunch of momentary pushbuttons that I can probably work into the design well enough for the knee sensor. I'm not sure if they're going to require too much force... time will tell. Easy enough to swap in a different style pushbutton. So at this point I have my codebase working again! The display indicates that everything is initializing properly. Hopefully tomorrow I can hook up the old craptastic chanter and get it making sounds again. And then I'll upload a video of me attempting to play Wild Mountain Thyme, with audio. Feeling pretty hopeful. It's nice to be making progress again.
|
|
#
?
Apr 18, 2023 04:21
|
|
|
It lives! muahahahahahahaha e: inline kills sound. Have a direct link to a video to me flailing around on the thing: https://imgur.com/I4duSPD I just couldn't deal with the--let's be charitable here--quirks of the Teensy, and realized last night in bed that now that I've given up on the TOF sensor, I can use the NeoTrellis, which is a lot easier to deal with. So I am! Upsides: doesn't make me want to throw the computer across the room Downsides: 12bit stereo output instead of 16bit, no longer being sold But the good news is that if I decide to make another one of these, I can just use an Adafruit Feather M4 Express. All I actually need is the two 12-bit DAC outputs and an I2C bus. Anyway. It's working. Next step: making my 3D printed chanter work with this setup. That will mean finding a way to mount the MPR121 and probably the display onto the instrument. I also need to think about how I'm going to wire the thing up. I'm thinking about a couple of options:
That last one would be the simplest. I'll probably do that. Gotta keep reminding myself I only need to make one for version 1. In case you skipped that wall of text looking for a photo or something: https://imgur.com/I4duSPD cruft fucked around with this message at 03:24 on Apr 19, 2023 |
|
#
?
Apr 19, 2023 03:18
|
|
|
|
| # ? May 5, 2024 20:43 |
|
|
I gave up.
So, yeah, that's it. I'm out. Here's what I wound up with:
I ordered a penny chanter kit from David Daye. When it gets here, I'm going to immediately try to get a 3d printable chanter designed, that I can swap in. David has published a pantload of design documents online, I want to carry that legacy forward if I can. At first I felt like opening this thread cursed me, but really, I just hadn't yet realized that I'd already accomplished what I had set out to do. If anybody's interested in trying to recreate my work and move it forward, you can PM me or email me (my address is all over the linked git repos)!
|
|
#
?
Dec 1, 2023 22:59
|
|
