Finally had some time to finish my MicroDexed, but alas, had no case. All was solved with a trip to the local dollar store. Being somewhat bookish, I found the perfect case! In honour of @C0d3man’s wonderful work, I replicated his red polka-dot tablecloth (minus the cloth) for the photos attached! Still have some wiring cleanup to do as I don’t have any of those tiny jumpers so I had to use wires instead.
The MicroDexed is working well in my test setup. Looking forward to connecting it to a full size keyboard and studio monitors sometime soon. A fun build! Thanks C0d3man!
The postman with my PCB, instead, is still swimming thru the cold oceans to get here… I hope he will send me a postcard when it arrives in New Zealand…
Fortunately, I’m working on the case for my new Zynthian, so it’s really not a problem… Let’s consider those €€ as a donation for you… or for the postman’s widow…
Mine only had to travel across the big pond, which it took almost a month. I guess the postman isn’t a good swimmer. Or, of course there’s the possibility that he’s a music fan and kept it for himself. Keep the faith and hope it arrives soon!
More for the postman’s widow. You will get a PCB from the next charge for free!
(Next charge is currently not in production… I just need some time to check the new layout before producing a lot of expensive srap)
@wolfpaw98 between @C0d3man and me there is just land. Doing some little math, I think i could cover that distance by foot without sweating in half the time we waited…
@C0d3man a question: do you think that a teensy 3.6 running microDexed have processing power left to scan a 61 keys velocity-sensitive keyboard? 'cause the case for my new Zyntiah is this:
Unfortunately its main board is beyond repair, but the keybed, front panel, and controls, are perfect… if there is some CPU time left, it could run uDexed and act as a MIDI class device for zynthian… else I’ll need two CPUs…
This is really a difficult question. The answer is: maybe. Currently 16 voices are working. For some algorithms there is not much CPU time left. But perhaps you don’t need to scan the keys every audio-block-cycle… other solution: just use only 15 voices of polyphony
I am thinking of an algorithm which does this automaticly: You can play as much tones until there is only a defined limit of CPU time for a cycle left…
BTW: I was thinking to do the same with a Fatar keyboard, but currently there is no time left for this project.
My best performance in poliphony is the two fingers version of chopsticks… go figure
A 1kHz scan frequency is enough. Most keyboard scan well over 2 kHz with custom chips… and this reminds me as I’m writing that… I have a box of EP2C5T144C8N FPGAs and I’m not afraid to use them…
Hmmm, with 128 bytes per buffer you have a calculation frequency of about 345 Hz. So you may get a problem here. The best working buffer size is 64 bytes or about 689 Hz
But why not using a simple Arduino/PIC/… with a matrix->MIDI engine and using MIDI for the Teensy (perhaps not the complex current loop variant, but the 5V serial one)?
Than’s the “else I’ll need two CPUs” part, @C0d3man
But really, I think I’ll go with the FPGA… too much time has passed since I played with VHDL… moreover, with a 200MHz main clock and 90 I/O pins, I can scan the keybed and all the front panel buttons and leds faster and easier than any similar uP/uC setup
Unless you want a copy of the system for your keyboard (or as a part of Zynthian accessories), in that case I could use something more “hobbyist friendly” than a TQFP-144 package and the dreaded Altera software suite
That’s really powerful. You can add several controllers…
Thanks - currently there is not much time for this project. I think there will be a lot of software for the Arduino to map a Painomatrix to MIDI and this would be the way I will try in the next ??? months.
I released a new version of MicroDexed. Currently not more features. Only some bugs are fixed and I replaced the MIDI engine with the one used in the MIDI library.
I also added a HEX file (in the hex-folder) for easy testing the software without my environment of libraries. I have done the same for MicroMDAEPiano.
The new commit will be at codeberg in the next 8 hours.
Not as many as you could think, IMHO.
At least, not for those “vanilla” Arduinos. Maybe 16MHz are not enough to scan 122 switches with enough frequency to get 127 levels of velocity. The best implementation I’ve found was on the teensy (hence my previous question).
As a note: the main CPU of Roland D5 runs at 12MHz, and it has two gate arrays to scan key matrix… not just the keybed, but front panel also…
The stl files are for the case, a top, a top with cooling slots, a top with cooling slots and holes for screws.There was a message saying new users could only have two attachments so I guess I will need another post.
I still need to solder in the pins for the music board and make a final check on the connections. I’ll let you know soon. Chaos has necessitated a short break to get some order in my work space.
Still have some wiring cleanup to do as I don’t have any of those tiny jumpers so I had to use wires instead.
