ilmenator

MIDIbox Matrix - MIDI Matrix with (up to) 56 input and 56 output ports

35 posts in this topic

Because my Waldorf Midibay became too small for connecting all of my MIDI equipment, I went ahead and designed my own MIDI matrix / programmable patchbay. The modular MIDIbox Matrix has up to 56 input and 56 output ports and provides a smart solution for the fact that not all MIDI equipment is usually located in the same rack or corner of the studio space.

 

MIDIbox Matrix features break-out boxes (BOBs) that give access to 4 MIDI in and 4 MIDI out ports at a time. These are connected via 9-pin serial cables to the main unit and are designed to be conveniently placed in the back of your rack, e.g. attached to the inner sidewall of a rack enclosure with screws or velcro tape. This means that instead of running 8 MIDI cables to the MIDIbox Matrix in order to connect four synths, you only need a single, cheap serial cable from your side rack to the center point of your matrix.

 

Two BOBs each connect to a single I/O board, which does all the level shifting and signal refreshing. Up to seven I/O boards can be attached to the heart of the MIDIbox Matrix, an FPGA-based switching and routing logic that is controlled via a core board such as the STM32 or the LPC17.

 

As the system is highly modular, it is possible to start with a low port count (4) and then increase the number of available ports by simply adding I/O boards and BOBs as needed. Fully loaded, i.e. offering 56 input and 56 output MIDI ports, the MIDIbox Matrix consists of

  • 1x FPGA board
  • 7x I/O boards
  • 14x BOBs

plus a core board and the PCB holding the user interface.

 

This is how a BOB looks like in my setup: post-3442-0-73779800-1402587749_thumb.jp

 

Software-wise, this is almost ready - everything works as expected, just the PANIC button has to be implemented in the next weeks, and some code cleanup will be necessary. You can find more documentation in the Wiki if you are interested.

 

Kind regards, ilmenator

2 people like this

Share this post


Link to post
Share on other sites

Very cool project!

Just not there with enough legacy synths to need it yet, but good to know, when G.A.S. strikes again! :)

 

Many greets!

Peter

Share this post


Link to post
Share on other sites

wow. I'll be watching this closely 

Share this post


Link to post
Share on other sites

hi ilmenator,

 

have read the mb_matrix-wiki. looks like a really interesting project. is it designed for the "CoreEP2C8" ? i suppose you can program the fpga-daughterboard with your fpga-mainboard ? eventhough i like your modular concept, i would go for something like 2 I/O-Boards and 4 BOBs (BlitterObjects ? :D). would it be possible to connect them directly to the stm32-core ? are you planning to upload your source into the mios32-repository ?

and yes as a altium-user i must admit, that i really like your kicad drawings!

keep up the great work!

 

mOnO

Share this post


Link to post
Share on other sites
eventhough i like your modular concept, i would go for something like 2 I/O-Boards and 4 BOBs (BlitterObjects ? :D).
That's a design decision based on my specific "studio" setup and the costs involved in fabricating small boards in larger quantities (as opposed to large boards in smaller quantities). Also, with this specific implementation it is possible to use cheap serial cables that are available everywhere. Serial cable plugs can easily be secured against falling off their sockets, which is nice when some of your gear is mounted in racks with wheels.

 

would it be possible to connect them directly to the stm32-core ?
Absolutely yes. A BOB (Break-Out Box) can be connected to the new STM32F4 core board via J11E. It is a nice alternative to the I/O board that you can order together with the core board from SmashTV (and you only need one for all four I/O ports). I hope to find some time this weekend to design a little adapter board (from 10pin IDC connector to 9pin DSUB) with level shifting for the MIDI out signals. I am not sure it is a good idea to keep the MIDI signals at the 3.3V level if distances are getting larger. At a 5V signal level I have tested 25m by chaining the five serial cables I have right now, and this worked perfectly. I don't even think that's the limit, but again this was at 5V signal level, which is what is specified in the MIDI standard if I remember correctly.

 

are you planning to upload your source into the mios32-repository ?
Yes, the MIOS sources will eventually go there, but there is very little magic involved - it is mostly based on the available modules for accessing the SD card, the SCS, and the shift register handling. Btw, I am developing this on an "old" STM32F1 core, and in terms of performance I guess that even a PIC core would do.

Share this post


Link to post
Share on other sites

Hmm, this is relevant to my interests :D

I have enough hardware synths to warrant getting some of these as I kind of sort of should get a MIDI patch bay at any rate. The old Mac G5 tower runs some old-school Mac PPC software like SoundDiver, Galaxy and such, the PC runs MIDIquest and separate editors for the XV-5080 and a few more. The i7 MBP runs Cntrl, the editor for the Radias, MIDIbox Studio and what not. Something like this would help me. This is even before I think about the iPad and those editors and things...

I'd be happy to help out testing this. Well done!

Edited by jojjelito

Share this post


Link to post
Share on other sites

I was pretty sure this would be something for you, too... :tongue:

 

I'll get back to you once I have received the next version of the FPGA boards which will provide a means for MIDI activity monitoring on all in and out ports.

Share this post


Link to post
Share on other sites

I hope to find some time this weekend to design a little adapter board (from 10pin IDC connector to 9pin DSUB) with level shifting for the MIDI out signals.

 

Here you go:

 

post-3442-0-27219400-1402732520_thumb.jp post-3442-0-54642200-1402732516_thumb.jp

 

This adapter board connects to J11E of the STM32F4 discovery board. With this in place, you can use the BOB as a regular four port MIDI I/O extension with the discovery core board.

Share this post


Link to post
Share on other sites

thanks ilmenator for your explanation. my understanding is, that i could connect only one BOB to J11E of the stm32core, which will give me 4 midi-ports (in/out). they work already with 5v in my case. dont need the smashtv-version, as i could easily do (and want) my very own. if i would need more than the 4 midi-ports, lets say 16, i would need the fpga-board with i/o-boards, so im obviously not getting around having one fpga-board (btw is it the "CoreEP2C8" ?) and probably 2 i/o-boards + 4 bobs, right ?

 

im fully aware, that your studio-setup requires 56 midi-ports. thats absolutely fine to me, but i was asking for a version for my setup, where 16 midi-ports should be enough for a start, as i could connect even more, if i wanted too...

 

mOnO

Share this post


Link to post
Share on other sites

I think there is a misunderstanding here. The MIDIbox Matrix routing functionality is not running on the STM32 core, but on the FPGA. It is a stand-alone project and not a hardware extension to other projects. The STM32 core in this project only controls the routing and provides a user interface as well as program storage. So, even if you only wanted a 4-port router for this project you'd still needed the FPGA board. (In which case it would make more sense to go for a software-based router running on the STM32 core itself).

 

If you want more MIDI ports in your other applications like NG, SEQ or whatnot, you need the IIC board.

 

However, if you only want to spare the space for two of SmashTVs MIDI I/O boards, or if your MIDI ports should be physically away from your core board, then the BOB in combination with the above adapter board is also an option.

Share this post


Link to post
Share on other sites

yes i got it now! this is why i think, your project is really great! i was asking for a 16-port-version (16 midi-ins/16 midi-outs), in which case i would need the fpga-board, 2 i/o-boards and 4 bobs. i think that this would fit nicely in my rack! if i would decide, that i need more than 16 ports, i would need another unit. i believe 16-ports will be the maximum for a 19-inch-device. had something like that in mind for quite a while now, but never had a real thought of it. i will keep watching here for any updates anyway. is the fpga-daughterboard the "CoreEP2C8" ? they really have a lot of interesting stuff there btw. are you planing to publish any schematics, apart from the sources for the stm32 and the fpga ?

Share this post


Link to post
Share on other sites

If you want all MIDI ports on the back of your rack enclosure then you might want to think about using SmashTV's MIDI I/O board for that, as the BOB has MIDI sockets on two opposite sides.

 

Take a look in the Wiki and at the BOB section and you can already see what kind of documentation will go there. Schematics will be available, and I am planning to provide PCBs to those interested - together with the BOM it should then be very easy to build your own MIDIbox Matrix. I will also provide the design files for the acrylic enclosure of the BOBs. I'll reveal the FPGA daughterboard details once the MIDI activity display has been tested on the next FPGA board iteration.

Share this post


Link to post
Share on other sites

thank you for your information, i have to be a bit more patient for the moment, as its still in work, but it definetely looks like a very useful project for me already...:-)

 

keep up the good work!

 

mOnO

Edited by mono

Share this post


Link to post
Share on other sites

A little update: I have received the second version prototype PCB of the FPGA board, the "heart" of the MIDI matrix. A quick test revealed that the communication works as expected, and I get feedback from the FPGA board to the core about input and output activity of the MIDI ports. Next on the list is an activity display board, essentially an 8x8 Duo LED matrix board.

 

The first seven rows will indicate MIDI port activity, the last row will display "virtual port" activity, e.g. activity of the mergers. This board might also be used by other projects that require some form of Duo LED matrix. It is located anywhere in the DIN/DOUT SR chain and connects to J8/J9 of the core board.

 

post-3442-0-55468700-1403438191_thumb.jp post-3442-0-91957700-1403438181_thumb.jp

Share this post


Link to post
Share on other sites

...excellent news! the duo-led-matrix-board looks quit neat. just one question: is the copper rest ring for the duo-leds big enough ? on the picture (bottom side) it looks like its too small and then it gets diffcult to solder. usually i add 30 mil to the hole-size, which gives me a 15mil copper rest-ring around the hole. with that i get good solder results. not sure how to do that in kicad, but im quite sure, that you can setup the pad-sizes somewhere. also i like the sil-header format. 10-pin is what i use for my projects as well. :D

 

keep up the good work anyway !

 

mOnO

Share this post


Link to post
Share on other sites
is the copper rest ring for the duo-leds big enough ?

It's the same as on my and it works ok there :smile:

Share this post


Link to post
Share on other sites

ok thanks for the information. it was a only a thought from me...

 

mOnO

Share this post


Link to post
Share on other sites

No problem, it's always good to have a second pair of eyes looking at the details! If I go for this option (not sure yet, the alternative is an RGB LED matrix module that consumes a little less front panel space) I might revise the pads.

Share this post


Link to post
Share on other sites

I have read through the wiki page, excellent documentation. Thanks for sharing! And nice to meet you BOB ;)

 

Count me in for one of the FPGA boards, and some BOBs as well.

Share this post


Link to post
Share on other sites

So after all I went with the RGB LED Matrix module to indicate MIDI activity on the 56 in-ports (red) and 56 out-ports (green) in the upper 7 rows. The lowest row displays MIDI merger activity, the two inputs of Merger 1 on the first two LEDs in red, and its output on the third in green. The same goes for Merger B on the following three LEDs. The last two LEDs in the last row are not involved for now. Also, I'll have to think of what to do with the blue color...

 

post-3442-0-71090100-1405546821_thumb.jp

 

In the picture you can see that a MIDI signal is coming in on port 1, which is then forwarded by the Matrix to ports 8 and 49. It is also forwarded to input B of Merger 1 (red LED in the lowest row), which then passes it on to its output (green LED next to it), which is connected to MIDI out port 56 (green LED in the lower right corner).

 

I have the RGB LED Matrix module working on breadboard for now, proper PCBs are on their way and should arrive next week or so. Apart from a little code cleanup, the Config page in the menu, and a Panic function, everything I wanted in this router is functional by now. :happy:

Edited by ilmenator

Share this post


Link to post
Share on other sites

Just a few more impressions of the system so far - the UI is still on veroboard, but the proper PCB should arrive this week.


post-3442-0-53210800-1406022745_thumb.jp post-3442-0-67485700-1406022751_thumb.jp post-3442-0-40705000-1406022757_thumb.jp

 

The I/O boards are stacked and just low enough to fit in a 2U rack case. You can see two versions, on the left the first prototype v1.0 (2x), on the right (3x) v1.1 with angled headers and a slightly different layout (but identical functionality). The MIDI activity display fits in a 2U frontpanel, as well as the UI (not the one on veroboard shown here).

Share this post


Link to post
Share on other sites

I'm following this project, really cool ,and usefull, i am actually thinking about buyong a midi patchbay maybe it will be better to wait yours ! :rofl:

 

Let us know about the project !

 

Have fun ! :smile:

Share this post


Link to post
Share on other sites

I got some extra pieces from Ilmenator as I have a boatload of MIDI stuff to control if I patch it all in. There's a package from Reichelt arriving hopefully tomorrow so I can get cracking building this. Hopefully I can document what I'm doing and we could achieve some repeatability.

 

I'll test this with both some interesting MIDI clock setup as well as mixing polypressure from my old Ensoniqs with lots of cc:s and arpeggiator data flowing out from my K5000s. If it handles all that, some interesting routing and refrains from messing up any Sysex I'll be a very happy camper :smile:

Edited by jojjelito

Share this post


Link to post
Share on other sites

The current status is documented in the Wiki, and jojjelito has a set of boards and FPGA core to do some beta testing.

 

I am currently relocating to another country, hence the extended period of silence and lack of recent updates. It will probably take me until after Christmas until I have my workshop up and running again. However, the unit is basically working the way I intended, and jojjelito volunteered to do some torturing tests.

 

Best, ilmenator

Share this post


Link to post
Share on other sites

Wow this is great, superb work!! I'm building a new studio/workshop in my garage now and this machine would be the perfect data centre. I'm lookinig forward to build one myself.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now