DIY MIDI keyboard

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi - I did this once (in the pre-midi days for an analogue mono synth I was constructing) because I had an old piano.. First I put contacts under the keys (piano still worked and keyboard felt great) then, to reduce the size of the instrument and to be able to play at night through headphones without waking my parents (I was 14- this was about 35 years ago!), I chopped the keyboard assembly away from the piano.. It ruined the whole thing! -- Good luck, anyway! (I now use a Yamaha PDP as my MKB - it cost £250 second hand, and Roland EP's with weighted KB's can be got cheap S/H).
 
Regarding sensors - I have found capacitive sensing to be the cleanest and simplest method of detecting key movement / velocity / pressure etc.. one suface needs to be at 0V, then the other (moving) part is connected to the scanning pulse (via diode) .. the capacitance formed will change as the key is moved.. returning the signal into a pair of comparators will give a timing difference as the capacitance increases - XORing the comparator outputs gives a pulse proportional to the the capacitance - a typical system (matrix wired) requires 16 comparators and 8 XORs, and diode matrix with charge limiting resistors (to charge the capacitors and act as keyboard 'section' address) - The only difficulty I have had with this method is that, even with extremely high value resistors (1M) the capacitance change is small compared to 'background' capacitance, and fast processing is required to get good resolution from the signal. Also, some 'dummy' contacts are required to give a reference TC as the capacitor dielectric changes with the environment (humidity etc) - also, GOOD ESD PROTECTION is required on the inputs!
 
Havoc said:

You would not be watching all the keys all the time. Just a scenario, probably not the best but the basic thing I was thinking about. Each key as a memory location set to 0 to start. You start scanning the keys. Once you detect a key being depressed, you add 1 to the memory location and continue scanning. Now while you continue to depress that key, the cpu keeps scanning and adding to the memory corresponding to that key. Once it hits the bottom switch, the note on command is send, together with the velocity info. and scanning continues. Once the bottom key opens again (you released the key), the note off command is send and the memory cleared.


It seems like this would be an awful lot of stuff for a midi controller to be doing, if the goal for this project was to make a keyboard that functions just like any other midi keyboard. Maybe, I don't quite get how the tech works, but where would you be doing these timing and memory operations ? I could definately see this in the realm of possibility, but it seems like you would have to design your own control circuitry, maybe a PIC or fpga, to translate the signals from all these switched to something that a normal midi controller could understand. The piezo method might be simpler.

-Dave
 
piezos.

take the entire piano action, and replace the strings with piezo bugs under rubber blocks. That'll give you the impulses across all of the notes as notes & velocity. From then on, well, it's sort of like a drum module. People DIY tham. RISC microprocessors and stuff. Not my field.

Or... buy a bunch of drum modules and control a computer sampler module for each range of notes driven by sets of piezos. Divide it up one module per octave. Lots and lots of cables, lots of drum modules, a bar with loads of piezos running behind the hammers of a gutted piano action. Rather heath robinson, but I bet it would sound and feel like the business.

After all, pianos are percussion instruments, sort of like automatic xylophones. The hammers give it the feel.
 
Another option, instead of relying on a mechanical design, use small electromagnets to provide a controlled "braking force" for each key.

It would be a somewhat insane and extremely difficult system to implement (compared to piano keys hitting pressure sensors), but hey – you could make the keys harder or softer at the touch of a button! Even make them feel like an old-fashioned mechanical organ! :devilr:
 
M-Audio Hammer-Action is what you need

Hey

I play an M-Audio Pro 88 hammer-action keyboard. It weighs more than 20 kg, so carrying it around isn't as easy as making a jam sandwich, but it's better than lugging a piano.

The hammer-action systems are great because they have mini hammer systems inside them. I think they use small cam wheels and a kind of flying hammer, but I haven't opened mine up yet to check.

With the way it does splits, I also use it as my organ, having assigned the nine drawbars to their equivalents in the soft synth. Further, it has piles of other controllers on board, so I also mix the various instruments I'm splitting from the keyboard. I never even open the lid of my laptop during a gig.

I still need a whole car:
1) flatten passenger seat and lay the keyboard on it (with seatbelt!)
2) laptop on passenger side floor.
3) gear bag (sound card, plugs, pedals etc) on back drivers-side floor
4) amp on back seat, drivers side.
5) keyboard stand and seat in the boot (trunk)

plus whatever else I'll be playing: trumpet, percussion, bass and other detritis in the boot too.

So the solution to your problem is to not want a piano. Mine is the most portable and lightest piano-alike solution I have ever used, and it plays better than most of the ridiculously-priced "pro" systems I've been given when playing at festivals and so on. (Kurzweil and their evil accomplices).

Other people make hammer-action systems, but I like M-Audio's feel the best, especially the way it integrates with the velocity curves.

Sorry, dude. That's the way it is... and I don't even have my own car; I have to borrow one for every show. :(
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.