MegaDrum for adventurous

[Ken Forgettable]

Wha-do-ya want from life:
<http://www.embeddedartists.com/products/lpcxpresso/lpc1768_xpr.php?PHPSESSID=v6srl9k0kfnvqdnm49lpd0t7o1>
... looks good for 96 inputs & 2 hi-hats ...
<edit> Ooops - only 8 ADCs (what a t.t).

[dm9876]

Personally I think if this project were to swap platform we would do well to use something which would let us get some sort of economy of scale happening. Not to mention, potentially allow people to purchase pre-made modules.

I would propose this one.. its a modular open platform which a couple of small companies manufacture kits (also ARM Cortex-M3 as the one Ken linked to).

http://www.ucapps.de/mbhp.html
http://www.ucapps.de/mbhp_core_stm32.html

and there is an OS already build for it. OK re implementing MegaDrum on this rather than just porting to different hardware is a totally different thing, but should be more maintainable / scalable and could open up so many possibilities

http://www.ucapps.de/mios.html

one benefit down the track I would like to see (from any more advanced hardware) is native MIDI-USB... instead of our current seperate uC to convert. It could save a couple of mS of latency in the chain.. everything counts. Plus its not just a couple of mS, you have so much stuff going on cc messages, multiple note on's triggered at once etc, note offs.. messages surely have to queue up causing more latency. OK so it does not seem to be an issue in practice, but eliminating it can only be good.

[dmitri]

MegaDrum runs on STM32, PIC32, PIC24. None of these are DIY friendly and presently don't give real advantage over Atmega design.

[Ken Forgettable]

... there is an OS already build for it.

It MUST be feasible to layer the definitive TMI over a RTOS in the future; as the actual channel sampling is pretty low spec' - I'm currently using 8Khz x5 (nyquist) which easily models the velocity and positional components on each channel. But I certainly wouldn't try to untangle the aliasing problems of mux'ing up the mux'ers on a partially determinant back line....

However, I'd like to thank GOOGLE search and Yahoo! Babel Fish for all they have done at this time [No simily avelable]

[dm9876]

I'm currently using 8Khz x5 (nyquist) which easily models the velocity and positional components on each channel.

so are you saying that you have coded your own algorithms for sampling piezo's for velocity and position or just that you can see that sampling at this frequency captures the information need? not sure I understand your issues with multiplexing?

Anyway what project are you working on?

[dmitri]

Anyway what project are you working on?

If it is a project and it is not a MegaDrum project then it is off topic here.

[dm9876]

fair enough.. no one would understand the answer anyway