MIDI Organ Console Project (Conn 652)

[John]

P. writes:

I just bought 4 of your [MIDI CPU] boards and would appreciate a little advice on configuration...

I plan to use them to convert an old Conn 652 console to run the Miditzer 260SP virtual theatre organ (after modifying the keyboards to provide 2nd touch contacts).

This then, will have 3 x 61 mote manuals, 2 with 2nd touch, 32 pedals and as many stops as I can reasonably persuade to fit along the existing stop rails plus a number of pistons, some of which would be assigned as generals and some divisional, plus 2 swell and one crescendo pedal.

What I had in mind was to use four boards as follows:
1 pedals / main swell / solo swell / cresendo
2 accomp manual / accomp 2nd touch
3 great manual / great 2nd touch
4 solo manual / pistons

Ideally, I would like to fit bar graphs to show the swell pedal positions but that is not a priority and I think I would need at least one more board for that anyhow. (I may modify the solo manual to provide velocity information to make the pianos more playable but if I do that I'll probably cannibalise an old yamaha keyboard for that...)

Does that setup sound reasonable or should I do it differently?

I'm unfamiliar with "2nd touch". What is it & how does it work?

Also, are the key switches or pedal switches already matrixed? Or will you be adding the matrix diodes, etc.

Please let me know, and I'll do my best to analyze your configuration.

The keyboards are probably matrixed in some fashion, as it currently uses digital keying but I will almost certainly be rewiring them to suit the new boards and fitting the diodes as required.
2nd touch is effectively another set of keyboard contacts operating on a different midi channel which close when additional pressure is applied to one or more keys. It is used to bring in another sound on one or more notes to make it sound as if the player has more than two hands and was originally introduced by Robert Hope Jones when he was building WurliTzer cinema organs. The MidiTzer software I will be running emulates a fairly large 3 manual + pedal WurliTzer known as the 260SP. The effect of 2nd touch can be likened to polyphonic aftertouch in most respects, although it is achieved in a different way.

I have had another look at the capabilities of the boards and changed my mind about the configuration, especially as 2nd touch will add only a very small amount of midi traffic, and I now propose the following configuration:

board 1: accomp manual (61 notes) / accomp 2nd touch (61 notes) (midi channel 1 & 5)

board 2: great manual (61 notes) / great 2nd touch (61 notes) (midi channel 2 & 6)

board 3: solo manual (61 notes - no 2nd touch) / pedalboard (32 notes) / main swell pedal / solo swell pedal / crescendo pedal (3
analogue pots) (midi channel 3 (keyboard), 4 (pedalboard) & 15 (analogues)).

board 4: stops / pistons (this will not accommodate all the stops of the 260SP but has more than enough capacity for the stops I presently have on the console. (midi channel 11 (54 stops) & 14 (25 pistons)). All the stops / pistons will be accessible from a touch screen on top of the console, so adding phyical stop switches to the console can wait until I have more money to spare!

The analogue pots may be better assigned to board 4 as they can generate quite a lot of traffic while playing expressively. What is the limit for the number of matrixed digital inputs that the board can reasonably accomodate with 3 analogue potentiometer inputs hooked up?

You can connect whatever combination of analog & matrix inputs you'd like. You're correct, however--lots of analog input activity on multiple inputs will reduce the matrix scan rate somewhat. But this shouldn't impact the overall functionality. I think you've got a good plan regardless.

If I decide to cannibalise my old velocity sensitive yamaha keyboard and use the guts to replace the solo manual (better control for pianos etc.), board 3 will be used for additional stop switches and the existing analogues. This would be better as it would reduce the midi traffic on a board which is also handling a keyboard.

The 61-key manuals will work well. I suggest arranging the switches in a 8x8 matrix. This will allow both the manual & the 2nd touch to be monitored by a single MIDI CPU, as you have suggested. I'll post an example wiring diagram as a response.

[John]

Here's the best way to wire two of the 61-switch matrices (manual + 2nd touch) to fit a single MIDI CPU.

The position of the switches and diodes can be swapped. But either way, the anodes must connect to the "data" lines, and the cathodes must connect to the "select" lines.

I chose control terminals 16-23 for data, because terminal #19 is unable to generate a select signal (see Table 9-1 in MIDI CPU Hardware Manual). Thus, two 8x8 matrices can be connected.

If this looks good, I can also post the configuration SysEx.

[stefanuk]

John,

You mentioned the sysex for this schematic, does it even need one, and is it published somewhere else?

[John]

Here's the sysex for the diagram above. Of course, the MIDI channels, note numbers, etc can be changed if we want. I chose MIDI channels 2 & 6 for the "great manual".

Update 2011-11-26: fixed config code for select lines so that notes are properly ordered. other posts in thread may not reflect the fix. -John

// Fixed Header
F0 00 01 5D 04 01
// Layer 00h
00

// CT 0-7: Matrix Select for "First Touch"
// Notes C2-C6 (36-96, or 24h-60h) on MIDI Channel 2
00 00 2E 02 5C 00
01 00 2E 02 54 00
02 00 2E 02 4C 00
03 00 2E 02 44 00
04 00 2E 02 3C 00
05 00 2E 02 34 00
06 00 2E 02 2C 00
07 00 2E 02 24 00

// CT 8-15: Matrix Select for "Second Touch"
// Notes C2-C6 (36-96, or 24h-60h) on MIDI Channel 6
08 00 2E 06 5C 00
09 00 2E 06 54 00
0A 00 2E 06 4C 00
0B 00 2E 06 44 00
0C 00 2E 06 3C 00
0D 00 2E 06 34 00
0E 00 2E 06 2C 00
0F 00 2E 06 24 00

// CT 16-23: Matrix Data
10 00 7F 00 00 00
10 01 7F 00 00 00
11 00 7F 00 00 00
11 01 7F 00 00 00
12 00 7F 00 00 00
12 01 7F 00 00 00
13 00 7F 00 00 00
13 01 7F 00 00 00
14 00 7F 00 00 00
14 01 7F 00 00 00
15 00 7F 00 00 00
15 01 7F 00 00 00
16 00 7F 00 00 00
16 01 7F 00 00 00
17 00 7F 00 00 00
17 01 7F 00 00 00

// Fixed Footer
F7

[stefanuk]

OK,

I get the wiring and it will work. Now, I have sysex, is there a place on forums that walks me through the steps of what to actually do with sysex?

[John]

Yes.

http://forum.highlyliquid.com/showthread.php?t=280

Be sure to let me know if you have questions.

[stefanuk]

It seems like two by 61 uses 122 out of 128 inputs, how do I add extra 6 switches ( for combination buttons and such), and maybe sysex for this? I just can't get sysex's myself...

[John]

Originally Posted by stefanuk

It seems like two by 61 uses 122 out of 128 inputs, how do I add extra 6 switches ( for combination buttons and such), and maybe sysex for this? I just can't get sysex's myself...

Hi stefanuk,

You're correct--we could add an additional 6 switches (3 to each matrix).

But there is a limitation--any matrixed switches that share a select line must all send the same type of output. So if you want to send notes with the extra switches, that's fine. In this case, simply add the switches--no configuration change is necessary.

But if you want to send something different, such as CC or program change, this isn't yet possible. (but I think it will be with firmware 1.3+)

Hopefully this answered your question. If so, let's revisit this once the new firmware becomes available--feel free to bump the thread. If not, please let me know.

[stefanuk]

I think this shoul work, with Hauptwerk I was able to use extra notes from the keyboard to operate switches before. Thank you! I will try and see...

[stefanuk]

OK, issue

Old Allen keyboard, first and second touch have a common wire, I can not separate it. Is thee a way to go around it?

The common wire is on the same MIDI note for 1st and 2nd touch, and then there are two separate contacts that I can put a diod on.

In the previous matrix that would mean 9 and 0 are the same wire

Can this be done?

[stefanuk]

what if first four notes of 1st touch are on channel 2, and the so are the first four notes of the 2nd touch?

and then it reverses to 4 and 4 notes on channel 6?

[stefanuk]

Or, can this be done?

[stefanuk]

However the easiest and ideal would be



MIDI channel 2 notes 1-2-3-4... 64

common wire ( in 8 sections of 8 contacts)

MIDI channel 6 notes 1-2-3-4... 64

Need help, am stuck!

[stefanuk]

Also,

Do I need any MIDI jumpers connected for this matrix?

[John]

Originally Posted by stefanuk

OK, issue

Old Allen keyboard, first and second touch have a common wire, I can not separate it. Is thee a way to go around it?

The common wire is on the same MIDI note for 1st and 2nd touch, and then there are two separate contacts that I can put a diod on.

In the previous matrix that would mean 9 and 0 are the same wire

Can this be done?

Joining CT 0 and 9 in the original diagram won't work--the MIDI CPU won't be able to distinguish between keystrokes on the first touch and second touch matrix.

There is a work-around for this problem, but it would require the use of a second MIDI CPU and a different wiring scheme.

Originally Posted by stefanuk

what if first four notes of 1st touch are on channel 2, and the so are the first four notes of the 2nd touch?

and then it reverses to 4 and 4 notes on channel 6?

I'm not sure I understand what you're suggesting here. However: for firmware versions 1.2 and earlier, any select line can be associated with only one output channel. This limitation also is a problem for the diagram you posted.

Originally Posted by stefanuk

However the easiest and ideal would be



MIDI channel 2 notes 1-2-3-4... 64

common wire ( in 8 sections of 8 contacts)

MIDI channel 6 notes 1-2-3-4... 64

Need help, am stuck!

There is a way to solve your problem (let's call it the "shared select line problem") by using two MIDI CPU units. Would you like for me to post the implementation? Let me know.

Originally Posted by stefanuk

Also,

Do I need any MIDI jumpers connected for this matrix?

If you mean for MIDI channel selection, no--the original sysex I posted above uses "hard coded" MIDI channels. The channel jumper is ignored.

Hope this was helpful...happy to continue the conversation. And thanks for your patience...end-of-year has been a bit busy.

[John]

Also, the "shared select line problem" will be solvable using a single MIDI CPU, with firmware version 1.3 (forthcoming).

[stefanuk]

OK, trouble shooting, maybe move to a different post...

Here is what I get

MIDI through test works
Connecting ground to any numbers sends signals.

Sysex seemes to be sent, but no results from the matrix at all....

What do I do?

[stefanuk]

John,

Thank you for all the answers, the new Hauptwerk can deal with mixed signals, so I went ahead and wired similar to original construction, but now I can not get anything to respond to it...

[John]

Before connecting the complicated switch matrix, do the basic tests described here:

http://forum.highlyliquid.com/showthread.php?t=267

[stefanuk]

I did.

Control Terminal Test and MIDI Wiring Test work beatifully. Then I send sysex, and it seems to go through, at least it does not give me any indications of it not going, and the MIDI interface blinks. But no matrix responce.