This was discussed somewhere else. I’m learning SMT soldering, and maybe you can join me so we will no longer look like monkeys putting sticks in holes.
I looked for SMT kits and the only thing I found was a sparkfun Simon game which appeared to have outdated documentation and some bad reviews. So I thought that I would design a “toy” circuit – something on which I can practice so no overtly complicated, but still useful. Also, this thing could help me learn about the process of having stuff manufactured – I’d rather blow money on a batch of 50 botched “toys” than 10x more expensive botched “upcoming mutable instruments product”.
Any idea of small musical thingies you’d want to have? The thing should not take more than 1 day to design, and no more than 1 day to write the firmware for. I’d prefer something with an AVR inside because I need practice on TQFP packages.
One example I did to practice SMT board layout: Midi events generator with 16 0-5V inputs

But then so many people are doing these “MIDI CPUs” (highly liquid, Livid, Midibox) boooooring.
Any other idea?
@fcd72, cool idea! but it’ll be battery powered, not MIDI bus powered – I’ve heard the gecko has problems with recent hardware whose MIDI output is too weak.
Sorry, from the requirements: and no more than 1 day to write the firmware for
I’d have thought a USB midi interface would offer no real challenge?
Ever looked inside one? There is not much in there.. And if there is already working firmware, then where is the fun in that for Olivier?
> Sorry, from the requirements: and no more than 1 day to write the firmware for
I saw that. But I imagine few of us have much of an idea what goes into coding and how long it would take?? (Although you did zip your way through the digi effects board firmware pretty quickly it seemed?)
Regarding the USB interface:
I like the arp idea :)
Yeah it can be done in one day. I can add as many options as you want, but what do you want in terms of UI? In terms of board size I want the thing to be about 1/4 the size of a Shruthi digital board.
If this continues like this, the SMT toy will become a Midibox SEQ v5. With 64 knobs, USB and a VCO :)
@rosch. The thing is, what I am looking for is the opposite – not something I would micropolish with 32 options and this and that controllable by CC and extra CV inputs and a debug serial interface, because I know I’d want to push it and implement every possible option to fill the code space. Something that serves one function – a useful one – without needing more than, say, 4 parameters to be set, and that won’t give me the feeling it is incomplete because only 50% of the code space of the AVR is used.
Some more ideas:
I talk about USB hid midi device, ‘cause i’m working on it right now…For the seq if it has to be very small, why not an euclidian one Like this soft one:
http://createdigitalmusic.com/2011/05/shake-it-like-euclid-grooving-patterns-open-source-tool-now-sends-midi-watch/
No many knob or Button are needed, for a great result :)
@ K&D. Hey, maybe there’s something worth doing in the “ear training” help department… Something that plays 2, 3, 5 notes with different difficulties (octave spread, rareness of interval). with 3 or 4 RGB LEDs showing your results over different time spans (last challenge, today, last week, last month)...
Euclid stuff is a good idea! Something that sends MIDI notes on channel 10 or with a few built in sounds?
After all, Korg sells Monotribes with 3 preset drum sounds and they seem to do fine :)
I’ve been messing about with that Euclidean sequencer for a few days now. It’s brilliant! Im glad he turned it into a standalone app, so I can route it to Logic and my hardware toys (Actually I was using it with a Shruthi-1 the other day).
I can’t see how it would be implemented in hardware though!? Least of all in something that takes 2 days to design/code.
While we’re on the subject of odd ball sequencers. I seriously like this:
Otomata
At first it looks like yet another ‘Tenori-on in flash’ rip-off. But it isn’t… It does a pretty cool generative sequencing thing too. Well worth checking out.
Shame it’s flash only though.. But perhaps he’ll build on it yet.
I’d do some serious leg humpage for a hardware version of it! However, I know im venturing well off Oliviers plans for this project now, so i’ll stop harking on about it! ;)
@Luap: One way to do the Euclidian sequencer with only a tempo knob and a MIDI in:
Octave 1 is kick
Octave 2 is snare
Octave 3 is hi-hat
You start a pattern by playing a 2 notes chord. The lower note is the number of elements, the upper note is the size of the pattern (C = 1, C# = 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16). You kill a pattern by playing a single note. I’m pretty sure it would be easier to remember patterns that sound well using the 2 notes that encode them, rather than the 2 numbers.

That does sound doable :) (Even if I do like trippy GUI’s) With the software, you can also offset each pattern against another. I guess you would achieve that by simply offsetting when you hit the required keys?! It seems like one of those things that would make a lot more sense when you try to actually use it, than it does to describe it.
(Ok whats up with the capybara thing? It’s like, staring at me.. As if it thinks I was the one that shaved it’s left leg or something)
Also, anybody here using a Minicommand?
Thinking of extenting the MIDI arpeggiator idea, how about a random chord generator ? There lots of stuff that could be used from the Shruthi, like the ragas for example, or am I wrong ?
Button 1: Play chord
Button 2: Randomize Chord
Encoder / Pot: chord complexity (the higher, the more complex, i.e. more notes
Overall idea: something that you don’t really know what’s coming out, but can be used for inspiration ?
My favourite arpeggiator is probably qmidiarp (Yep, I’m running Linux).
You make a pattern with a string, for example: ‘>>1-0+<00p-2-d1+h1’
>/<=double/half speed
0-9=note in queue
+/- =octave up/down
d/h=double/half note length
p=pause
Maybe one could program that little device by pressing a button, programming the pattern with notes which are mapped somehow, pressing the button again and it’s active.
It would have to sync to midi clock of course.
My suggestion – one annoyance I’ve had in my studio for ages is that Roland JD-800 keyboard sends MIDI active sensing messages all the time. These active sensing messages are useless to me, and I want to filter them out. I looked for standalone little boxes for ages that would do this and nothing else. I don’t think it exists. I looked into building a MIDIbox thing to do it, but it seems like overkill.
Does this fit the desired use case at all?
To summarize, a small box with a MIDI in, a MIDI out, some sort of minimal input device (pair of switches or clickable encoder), powered by batteries (2 x 1.5V or 9V??) some sort of minimal display (LED or 1x8 LCD) and that would run simple one-purpose firmware, upgradable through MIDI.
List of firmwares:
Sort of stripped-down minicommand?
Yeah it can do MIDI monitoring too. I was thinking of a 8x1 LCD like this:

I know I am late to this but a couple projects that would be cool would be these:
simple drum triggers to midi
an uber tiny 1 channel either TB or TR sequencer
something involving a artificial neural network!
that arp sounds sweet though, and especially the euclidean pattern generator!
I’m still inexperienced when it comes to routing SMT boards, and I haven’t really tested how the 3.3V MIDI thing will work (the small, backlit LCD I found from a reliable source is 3.3V, so everything on this board will be at 3.3V to avoid level translation).

@kvitekp: Where do you want the pull-up resistor? There are already two LEDs (G and R) below the MIDI in and out.
There is only one UART on the ATMega168/328, so you have to choose between debugging and MIDI out. The OUT pads at the left and right of the board are connected to the UART TX pin, so you can connect your FTDI cable here. BTW, those pads are here to daisy chain several units – in case you want to put several of them into the same case and build a controller.
Yes, there might be 2 or 3 holes at some point, when I’ll have figured out the housing for this.
There are 9 spare ports: PB4 (accessible through the MISO pin of the ISP programmer), and PC0..7 (which are exposed so that you can connect sensors / pots and use this as a basis for building a MIDI-enabled thingie).
cool, just seen this. some interesting sounding ideas :) Especially a very stable MIDI Clock source. This would however ideally have more than one midi out (maybe?)
God only know how id get on with SMD stuff but worth a shot if not too much.
Sounds like a good project.
Great ideas.
One idea that has plagued me since hearing about it is the “pseudo-algorithmically generated” step-sequencer. Oliver, you mentioned this a long time ago, when first talking about the Shruthi. All it would need is a button/encoder combo, for input and selection.
Maybe you have these preset patterns in memory and you can hybridize or randomize them, by using the encoder to control the amount of the effect.
Euclidean pattern generator would be close enough, though!
Added to the TODO list.
It looks like you're new here. If you want to get involved, click one of these buttons!