www.megadrum.info

Hi all,
I would like to share my thoughts about a different approach for positional sensing with you. While I'm intrigued by the possibilities of positional sensing within electronic drums, I have the feeling that Roland's approach (using only one piezo and analyzing the wave length of each hit to determine the position) is not very effective nor precise. Posts on this forum about using the Roland type of drums with Dimitri's experimental positional sensing firmware seem to confirm this idea.

So I thought: why not use the piezo solely to determine the velocity of the hit and use a bunch of membrane switches (or one big membrane 'touch screen') to determine the position of the hit. I did my research and to my knowledge no one has tried to make a DIY device according to this design yet. Somewhat later I discovered that there is a commercial solution, Mandala, that uses a similar approach (see US Patent number US6815602 if you would like to know more about this). In short: we have a proof of concept!

Now, what I would like to try out, whenever I have some free time available (argh!) is to 1) use the megadrum for the piezo part 2) put a couple of membrane switches underneath the drum head 3) hook these up to some device (to be determined) that converts the membrane switches to midi signals. Possibly, for the latter an extended version of the megadrum could be developed that interprets a large number of on/off signals (from the membrane switches) and combines this data with the velocity data (from the piezo) in one midi signal.

Let me know if you think all of this is feasible
Samdrum

UPDATE: for the analysing of the data from the membrane switches, I found a couple of solutions, that are widely used for building midi buttonboxes and midifying organs:

• http://www.ucapps.de/midio128.html
• http://www.doepfer.de/usb64.htm
• http://www.midiboutique.com/

The first option has my current preference, since the specs seem to be in line with our purpose:

• The software captures all 128 inputs within a period of 1 mS. If the status of an input pin has been changed, the program searches for the predefined MIDI event
• The digital input reacts on TTL levels, this means it switches to "0" on a level below 0.5V and switches to "1" on a level above ca. 3V.

For the "switches" I already have some idea in my head (similar to the membrane switches used in DIY ecymbals). The most difficult part will be ensuring that the data from the megadrum is accurately combined (in the pc) with the data from one of the above modules. Here timing is key...

I find your thinking very interesting Samdrum.
It seems like a good plan, but I don't know it it's possible to blend with MD...I hope so
Keep us updated with your thoughts!

Mandala has been discussed before but I can't find the link, I do know that it requires a very firm playing surface rather like the old simmons stuff. The patent is dead in the water because you end up with damaged wrists...

That sounds interesting Ken; I would love to hear more about that!

I have a hunch why the material has to be firm. After each hit it will take some time for the material to "de-bounce" (going back to neutral). This is undesirable because the membrane switch is activated for a too long period (in which another hit could come in at a different area on the drum). If the head is very stiff, the de-bounce would take less time.

It will take some experimenting and creativity to overcome this. Any ideas are more than welcome I was thinking of making my own switch mechanism using a sheet of velostat sandwitched between 2 conductive layers. Or putting a somewhat thick mesh (with large gaps) between 2 conductive layers. Also, I've been looking into "wearable electronics" and there is a wealth of information about this subject available on the web. It is even possible to make the mesh head itself conductive (via conductive paint).