AVR Freaks

Wow. Great feedback. So it looks like generally they're considered a "good idea". Think I'll dip my toe in the water and see how I get on. Apart from anything else I've got 30..40 pots I want to wire up to 16 channel analog multiplexers and then into AVR ADC pins. 

Hi folks,

So firstly a huge thank you to everyone who sent me a PM about helping me to actually produce a PCB - you are very generous!

I'll just explain a bit more about where I am with this and maybe you'll see why I might not be quite ready to go for the full PCB solution as (maybe mistakenly?) it feels a bit too "final"

As noted in a previous thread:

https://www.avrfreaks.net/forum/...

This is all about designing a synthesizer. I haven't quite finished this "soft" design yet but this gives a taste of the kind of "front panel" I want this thing to have:

Now a lot of modern synths have very little more than a fairly capable GLCD, a handful of buttons and a small number (even just one) encoders. Then to program you dive into menus (either clicking buttons or click/rotating the encoder etc). But if you make a UI like that you lose the immediacy of instant gratification of twiddling one or more controls at the same time so it's much more difficult to "try out" various settings while getting instant feedback of how that sounds.

So I want to go "old school" on this. An oft repeated/requested feature in the synth loving community is "one function knobs". So instead of making the same pot/encoder adjust 10 different things according to which menu chosen "mode" you happen to be in right now people are looking for a resurrection of the 80's/90's synth world where the basic synths (mainly because most of what was in them was analogue rather than digital) had a knob that did one and only one thing - perhaps the filter cut off or filter resonance or something and that's basically because it was the pot that formed part of the analogue filter circuity. It wasn't the case that it's position was being digitized by an ADC and then the control software was using this to program a cutoff/register value in some other device (which is kind of what goes on inside modern synths).

So, long story short ("too late"!) while it is going to be a "digital synth" I want something with an old-world analogue feel to the user interface. To that end I mainly discounted the use of encoders (which usually are tied to a display so it can feedback to the user what position the control is currently in). I do plan to have a small display and encoder for "housekeeping" but I want the main work to be done by a whole raft of pots and the knobs themselves will give instant user feedback as to what position they are in. See madly coloured diagram above and you can pick any one of those controls and instantly see how far it has been rotated between limits.

So to just test the pots and see what they "feel" like I have just wedged them into a breadboard for now:

But that was basically just to test one example of the 7mm shaft and one example of the 12mm shaft - which lead me to conclude that the 12mm is the better of the two.

So now I have bags of these things so I want to lay out 40 (4 rows of 10) to see what the whole UI will feel like but it's very very unlikely to be the final solution - this is just another step in the prototyping but I want to try stuff like seeing how close/ar apart to place them both to try and make a fairly "compact" layout but not so close that you bump another control as your fat fingers are trying to adjust another knob. So I don't even know how far apart I want them to be in the grid yet.

I also don't know if I want to stick with raw pots. I like "pocket operator" synthesizers a lot. They have an "industrial look":

but I also have a bunch of matching knobs...

OK those are on encoders not pots but they can be used and I have bags of these knobs anyway.

But if I put the knobs on then the pots would need to be positioned further apart to ensure there was still "finger room" to grab hold of the controls.

So I want a solid base (not flimsy like when you just push the pins of a pot into a breadboard) but something I can "try stuff out with" (assuming my desoldering skills will permit it) to try positioning them at various distances apart to see how it looks, how it feels and to actually read the values and feed them into code.

With that in mind (and again because I don't want to have to worry about "raw chips" at this stage) I have ten of these:

which are 16 way CD74HC4067 analog multiplexers so that I can easily attach 40+ pots to just one or two ADC pins.

Oh and there's another thing. A lot of modern devices...

(I have two of those already) Have some kind of "button interface" which is used (a) for playing through a piano keyboard, (b) entering steps into a "step sequencer" and (c) "finger drumming".

I have a distant plan to include some kind of keys like this in my design too. Once again I have bags full of tact switches so those could be used for "step sequencing" (like the 16 buttons across the middle of the Liven XFM above). But you can't really "play" tact switches - they won't work either as piano keys or finger drum pads.

So like the IK Uno (bottom one above) capacitive pads are an interesting option - that can certainly work for "piano keys" and it'll even work for finger drumming.

A "better" solution for finger drumming though (which can also be used as piano keys and even touch/velocity sensitive piano keys) are the kind of pads in the Novation Circuit Tracks (top one). You could (very expensively!) create such a number of pads with either force sensing resistors or piezo elements used as input not output (the latter is used a lot in electronic drum kits!) but there is a "cheap and cheerful" solution that involves copper pads separate by velostat and with a foam square on top. You can get proper velocity sensitivity this way (it's kind of what is under the hood in a Circuit Tracks) but it's the most mechanically complex solution as you effectively want to make a "sandwich" of copper/velostat/copper/foam and make sure it's all held in place. For the base layer of copper than can, of course, be a large pad on a PCB but you are then going to probably need something like a 3D printed plastic enclosure to hold the other elements of the "sandwich" that makes each piano/drum pad in place - that bit may be beyond me.

But whether I opt to add capacitive pads (most likely) or velostat pads or simply just tact switches (and again you could make 3D printed plastic "keys" to go on top?) all that almost certainly needs a "PCB backbone" both to give it rigid structure but also to form part of the circuit (whether the pads are capacitive or part of a veolstat/foam sandwich).

So anyway, yes, I do see a full PCB layout for this in the future (assuming early prototyping experiments succeed) but in the short term I guess I am just using a board as a "solid backbone" onto which to mount stuff (but also connect it electrically togeth). Hence the question about whether these solderable breadboards (which are possibly better than just vero/stripboard) are any good ;-)

Because SuperBooth (where synth suppliers gather to show their new stuff) is on right now in Germany I just read about:

https://www.synthux.academy/simple

On the surface this looks "clever" because they have designed a PCB with a lot of "sites" into which you can optionally plug pots, switches, jacks to layout your own front panel design. Putting any device at any position.

But then you see the price. A kit with 20 pots, 10 switches, 10 jacks and a Teensy 4 (and the "clever" PCB) costs €289 (OK, short term discount to €199). Well I know a Teensy 4 is $25 and the pots/switches/jacks are about $20-$25 from Aliexpress so that is going to have to be one hell of a PCB for the money!!

Quite clever though.

But that's the thing - I want to get AWAY from the unreliability of spring clipped contacts in a breadboard. I expect buttons/knobs/switches to be heavily tweaked and they will simply come loose from a breadboard before very long.

As JLCPCB is so cheap it is kind of tempting to have a crack at it. Just trying to weigh up things between EasyEDA and KiCAD.

The great thing about webapps IMAO is that I don't have to install anything on my PC and I don't have "data synchronisation issues"

The no-install thing is nice especially when working on a corporate machine where I don't (usually) have admin rights (as it happens I applied for and got rights yesterday but it is for corporate safety so that not "any old thing" can get installed)

The web storage thing is nice because I can do changes on one machine, leave that, switch to another machine (and for some things that can even mean my phone) and pick up exactly where I left off. The alternative is the horrors of things like OneDrive, Dropbox etc or maybe using Github or some other centralized RVC system to check things in and out.

In the focussed little world where I play (synths etc) there is an increasing move to online apps and project storage. Things like Roland ZenBeats (a DAW but one that happens in you browser), also Bandlab Cakewalk (another DAW). I also like that there are online synths like this:

https://notes.ameo.design/fm.html

So I don't need an "app" on Windows machine, a Linux machine, my phone. I can just "play" with such a toy whatever device I happen to be viewing the interwebs with at the time.

To be honest when I read a bout KiCAD and EasyEDA (the two main programs suggested/supporting JLCPCB who appear to have the cheapest of PCB services - and even assembly / component sourcing options - I would say KiCAD though potentially more complex looks like the "better fit" for the way I like to do things. But EasyEDA is a nice idea in that it is cloud hosted. But what I don't get is why, if I have signed into it from one machine it cannot set a cookie (even if it's a limited lifespan) to say "keep logged in" for a day or two. Instead, every time you return you have to log in then you have to play the Captcha game which is irritating beyond belief. Presumably they are trying to prevent robots scouring their component catalog/stock levels or something ?

Ah I wonder if the repeated Captcha was because of the VPN I'm connected to most of the time? It makes my machine look like it's in Germany. I should try it from my personal laptop perhaps. 

... rats, Practical Wireless but not Practical Electronics :-(