Scan matrix - How far can I take it?

[machinate]

Hi yall. I'm slooowly purchasing the parts I'll need for my project, which, in short, will be a big old bunch of buttons with some extra bits in there. Roughly 144 to 160 buttons AND about 24 encoders.

So I'm looking into scan matrices, in order to keep the number of dinx4s down for the buttons, along the lines of this:

http://ucapps.de/mios/sm_example2_v1.zip - I like the sound of "super-fast!"  8)

But then I read the text here:

http://ucapps.de/mbhp/mbhp_scan_matrix1.pdf

-saying that an example of more than 64 buttons will come at a later point in time. I remember popping my head in on this forum back then - has the development continued, or did it stop back then? "Additional examples, e.g. for 1024 buttons scanned by a single core module, are planned once this driver has been tested in a real environment." - I assume that some bit of testing will have happened since that time?

Will it be feasible to establish 3 such matrices in a project at this point? On a scale of 1-10, how much of a code genius would you have to be? From a laymans POV the idea that you have this really fast 64 input matrix would mean really good scanning of larger arrays as well, right?

My only other alternative I guess would be a two core and 8 dinx4 system, which would be too much, I fear.

Cheers, and thank you in advance for any help you might be able to throw in my direction - I plan to do a thoroughly documented project to share with you all!

I also have many thoughts on future projects involving even larger button rigs, so this is of very high importance to me.

Andreas - Machinate

[TK.]

has the development continued, or did it stop back then?

I haven't continued, because nobody gave me feedback if the example works in real life.

I assume that some bit of testing will have happened since that time?

This would also be interesting for me ;-)

I only tried it very shortly with the old keyboard of a friend. I won't be able to test larger matrices, therefore my hope is, that somebody with the appr. hardware and assembly skills enhances and example for a larger number of inputs.

(Note that I by myself have absolutely no advantage of this project, therefore it's really low priority for me to support variations)

Additional info: from the technical point of view, scanning 1024 buttons organized in a matrix is no problem, and it will work with very low latency (< 1 mS, so faster than a single MIDI event could be transmitted)

Best Regards, Thorsten.

[machinate]

Hey man, that's strangely encouraging :-) I'm getting code-fiends to help me - my common sense tells me that duplicating the code and making a few alterations should do the trick - again, I'm glad I've got people helping me;)

So for testing purposes how big should we do this matrix? Would 160 do it?

I really want to help out all I can, if anything by soldering up a ton of buttons :) Who knows, I may get excited and do 320 buttons in one go, hehe ;)

Andreas

[TK.]

So for testing purposes how big should we do this matrix? Would 160 do it?

it depends on your endurance while soldering so many buttons and diodes ;-)

Best Regards, Thorsten.

[machinate]

hehe, well, I'll need the buttons either way, so I kinda have to try it ;)