Making A $2000 Synth For $99

what's up i'm edward and for a long time i really wanted the juno 106. the only problem was it was super expensive so i decided to make my own um yeah to start off i chose the teensy 4.1 for this project because of the high price to performance ratio and also because of the tnt audio library which makes the programming a lot easier it also worked with arduino which was really helpful for me as a beginner i first soldered together the audio adapter board and the tnc 3.2 which i was using at the time and that just gave me a really easy platform to test my code on so there's four oscillators sub noise pulse and saw where sub is just pulse with an octave down the volume is controlled by this mixer and then they go through the high and low pass filters which are then modulated by the envelope the signal is then duplicated and sent into the flangers and is ultimately output as a stereo signal that goes to the sgtl through the headphone jack and then into your ears once i had the software done i had to start working on the hardware in this case it was pretty simple i had midi in and out then a bunch of sliders rotary potentiometers and buttons as well as the multiplexers to read them in i also had the leds the tnc the audio shield and the audio inputs and outputs one thing that i ran into is that tc only has 48 pins 12 of which are used for the audio board and six of which are used for power and ground so that only leaves about 30 for io additionally i didn't know that you could use ics to power multiple leds such as the lm3914 so i wasted 9 pins on just leds i settled on the idea of using multiplexers to read in my inputs a multiplexer is a combination of logical gates that lets you read multiple inputs while only using one data pin to the teensy and log base 2n pins for control you can think of it as one of those toy train switchers based on the configuration of the control pins a specific input line is connected to the output here's an example if we wanted to read in the input at index 9 we would configure the control pins to spell out 9 in binary so 1 0 0 1 this means pin s0 would be high pin s1 would be low s2 would be low and s3 would be high next i started designing the pcb and eagle this part was really hard as i had pretty much zero hardware knowledge going into it like i didn't even know how an led worked the process was really tedious but i got through by watching a ton of youtube videos and pretty much just copying elements from a bunch of other people's designs one thing that surprised me was that 90 of the time was just spent researching components and reading data sheets this is because for every component you add you need to know how it works find the symbol which is what it looks like on a schematic the footprint which is the shape it occupies on the pcb and the 3d model which lets you model the project in 3d you then combine all of these into eagle to make something called a library i did this for every single part i used which added a ton of time and combined with the fact that there's no way to test if your circuit actually works or not led to a lot of stress once i had everything down i had to route it i tried to write it by hand a few times but as it was my first time i ended up painting myself into a corner a lot and having to restart i ended up using just auto router for everything which turned out to be a horrible idea after i finished the pcb i ordered it on jlcpcb the sponsor of this video it was honestly a very cool experience it takes about three days for them to build it and then a week to ship it to you and you can see the progress as the manufacturing process goes i bought the parts on digikey which is actually really dope especially for canadians they have a policy if you order over a hundred dollars in cad the shipping is free and if you order less than a hundred the shipping is a flat rate of eight dollars and both options have tariffs included this is way better than other companies one time i ordered eighty dollars of parts and had to pay 40 shipping and 50 customs next soldering once i had all the parts i soldered it together i started with the smd components so that the board would be flat and then moved on to the through-hole parts i think my soldering really improved through this project here's my first ever solder for example something i noticed that's kind of crazy is the amount of work that goes into designing these components for example they designed these potentiometers so you can bend the legs down so that it holds itself up and leaves your hands free to solder once i had everything soldered i went back to software next debouncing so there's this problem that happens when you actuate a mechanical switch and if you played minecraft before you might know what it is it's called contact bounce or chatter basically when you press down on the button a combination of mechanical vibration and electromagnetic interference caused it to bounce up and down for a short period of time during this time the reading is unstable and it can be a big problem if you don't deal with it thus deep bouncing there's a really cool and short algorithm i found to deal with it this is also the first time i've ever actually used binary in programming and it was really awesome basically this function is called every n milliseconds and uses digital logic to create a history of the reading states kind of like a graph next problems i ran into a ton of problems while designing this the most obvious clearances i put the key switches way too close together so you can't actually put keycaps on them also multiplexers and shorts i'm not really sure what's at fault here but there's a lot of behavior i don't understand with the multiplexers for example if you pull every 200 milliseconds or another relatively long interval it will cause a weird delay sound in the synth even though it shouldn't if you pull with no delay the multiflexion won't be able to react fast enough to the teensy to the change in the control pins which means that the inputs read consecutively will often return the same value if you pull at a reasonable interval of 2 to 15 milliseconds these issues don't happen but some readings are still tied to each other for example the volume of the noise the noise oscillator is that input five while the frequency of the low pass filter is at input seven when you play a note and you move the slider seven down there they both change like they're tied together additionally there were button shorts the buttons are pretty much all shorted together but here here's which ones are connected um i literally have no idea why this happens and i still can't fix it as i'm making this video the funny thing is i i had enough butt pins to run the buttons directly to the teensy which would be way easier and also cheaper because i need one less multiplexer so i totally over complicated this the leds do look pretty cool though so here's as you can see there's kind of just like a totally random pattern to what each button does to the point where it's like kind of like a puzzle like like from skylanders or something [Music] [Applause] [Music] don't forget to hit that like and follow button guys see you next time

Channel: Edward Wang

Views: 247,506

Keywords:

Id: hBFbo3VF1TQ

Length: 8min 31sec (511 seconds)

Published: Mon Mar 07 2022