Conference napalm::commusic_v1

    I don't remember the article, but I can assure you that a Merge
    box isn't simple, if you need the fully-general case to work.
    If you can live with sending MIDI information from only one unit
    at a time you can build the much simpler A-B switch.
    
    However, if both boxes will be sending note data simultaneously,
    and/or system exclusive data, you will need the fully-general merger.
    Merging must be done on a message basis, as opposed to a bit- or
    byte-basis, since the receiver expects to see messages.  This means
    that your merge box will have to buffer a message from each input,
    and send it to the output when the message is complete.  This can
    require up to 4096 bytes of message buffer per input if you are
    using DX7 or TX7 system exclusive messages.
    
    Another problem is MIDI clock.  You want just one source of clock,
    and you want all of your sequencers to be driven by that clock.
    If one of the clock paths goes through the merger then the merger
    will have to pass on clock messages without excessive buffering.
    
    If you do build a MIDI merger, please let us know.  The problems
    are great enough that your solutions should be very interesting.
        John Sauter

    If you want to build an A/B MIDI box so you can toggle between
    the DX and the Drumslave into the sequencer, I have an article on 
    how to do it in Electronic Musician from last fall.
    
    Dave

    There was some note activity a while back regarding a mail order
    company (Xerbitron ?? or some such silly name) that sells a cheap
    MIDI thru / A-B switcher box in kit form. You can buy plans, circuit
    board or a full kit. I ordered some cheapo software and a catalog
    from them and they seem OK. I'll check for correct name and address
    tonight.
    
    I think the comments about true merge devices are unfortunately
    true. I haven't seen anything for less than a couple of hundred.
    
    Rob

    Thanks for the input thus far.  I hadn't thought about the message
    ambiguity problem that a thru box must take care of.  Yea, I would
    also like to take the MIDI clock from the TR505 and merge that into
    the sequencer along with the DX and the Drumslave.  That does pose
    a problem that only a *real* merge box can solve.  It is not a simple
    logical OR due to collision potential (both bitwise and messagewise).
    
    Looks like the only alternative is to build a switch box between
    the DX and the Drumslave, and live with the sequencer's not_
    as_user_accessable clock.  I have plans for a switch box out of
    keyboard from a few months back.  It doesn't say whether or not
    you have to power down all instruments before you throw the switch
    or not tho.  Do you????  I was under the impression that it is not
    good to break/make MIDI connections while the equipment was up.
    
    I plan on building a multiple thru box as per the MIDI spec.  Just
    add a few more thru circuits in parallel to the output of the opto...
    
    Keep the comments rolling - I'm relatively new to MIDI use...
    
    Todd the_getting_midicultured.
    

    As far as I know there is no problem making and breaking MIDI
    connections with equipment powered on.  If there are messages being
    passed you will garbage the messages, of course, but that shouldn't
    cause anything worse than "stuck notes".
    
    Is it possible for your TR505 to *take* MIDI timing?  If so, you
    could run a MIDI cable from the sequencer's MIDI OUT to the TR505's
    MIDI IN.
        John Sauter

    Yes, the TR505 does take MIDI clock in, but changing tempo on the
    sequencer isn't as easy (user interface wise) as it is to turn the
    tempo dial on the TR505.  However, for the price of a MIDI merge
    box, it isn't worth the extra feature.  To get the tempo I want
    in a hurry, I can take the TR505 out of MIDI sync mode, play a
    pattern in the same time signature as the one I wish to record, look
    at the BPM, dial the BPM into the sequencer, and put the TR505 back
    into sync mode.  Voila!  I just saved $200...
    
    Thanks for the info...
    
    Todd.
    

    re:xerbitron - see note 513.  It would cost about $40 for their
    MIDI switcher (SK-4) not including power supply, solder, wire or case, but 
    including pc-board, components and manual.  Pc-board is $11.  Manual
    is $4.

    Power supply?  The MIDI switcher in Keyboard is totally passive.
    It consists of nothing more than a DPDT switch and two female MIDI
    jacks.  Hmmmmmm...
    
    Todd.
    

    The Xerbitron product mentioned in .7 is both a switcher and a thru
    box. You get two MIDI inputs (with a switch selecter) and four outputs.
    I think you probably need some power to run the circuits that do
    the split into four outs.
    
    Rob

    
    
    
    Yes, simple midi switchers can be made with DTDT switches - I use
    a couple of them. There is no problem disconnecting midi lines with
    the power on - the reason the MMA gives for their crazy 5 ma current
    loop standard is "to allow that".
    
    A midi through multiplier box is pretty simple to construct. The
    midi spec actually gives a circuit example of this. If you have
    the maintenance manual for any midi device you can quickly piece
    together a simple schematic. All you need is some phototransistors,
    op amps and 220 ohm resistors. After finding the optocouplers, the
    hardest part is making the holes in the box for the din connectors
    and ripping off a suitable power supply.
    
    A midi merge box is, as John Sauter said, a different kettle of
    fish. You have to do it by message, not byte or bit. You really
    need a microprocessor to do it with enough memory to buffer whatever
    level you'd like it to work to (ie. if you want system private messages
    this could be a problem - they are LONGGGGGGGGG for many applications).
    
    One of my pet gripes at the midi spec is that it doesn't insist
    that when a device receives messages for channels it isn't the least
    bit interested in, the device should "pass the messages through"
    to its midi out port. This would mean that each midi device would
    effectively merge incomming channel information for which it had
    no interest with its own outgoing midi information. (I'd accept
    restricting this to note/velocity/after-touch, etc. information
    only, or some subset which did not include system specific messages.)
    
    Of course, you couldn't merge different clocks, and you'd want to
    be able to disable pass-through too (I suppose).
    
    Wouldn't that be great? Note that the Chroma Polaris does exactly
    this. But then again, it has an 80186 to burn ...
    
    Robert
        

    I hear ya regarding MIDI thru.  When I first attacked MIDI for the
    first time using my sequencer, drum machine, and synth, I was under
    the impression that a device's MIDI_THRU = MIDI_IN + MIDI_OUT. 
    Boy was I disappointed when I discovered that MIDI_THRU was just
    a clone of whatever comes in the MIDI_IN and contains no MIDI info
    for the passing device...
    
    Todd.
    

    Some MIDI devices support a "mix" option, that combines the stuff
    coming in the IN with the stuff generated locally, if you enable
    the mix option.  Some others have an output that can be switched
    to be either an OUT or a THRU.
    
    len.
    

    
    
    
    Yup, sure would be nice if all MIDI devices had the mix option.
    Maybe if ever there is a MIDI V2 such behaviour will be specified.
    Of course, maybe by then some of the other oddities of MIDI will
    be rationalized ;-) .
    
    Robert
    
    

    Well, this seems like as good a place as any to ask my question.
    I have my Atari-st driving my dx-100, tx-7 and tr-505. The midi
    cable runs from the atari "out" to the dx-100 "in". From there, it runs
    from the dx-100 "thru" to the tx-7 "in". The problem is that the stuff
    from the atari only comes out the "thru" port of the dx-100, but the
    dx-100 keys only come out the "out" port. I would like them both to
    be on the same port. I assume this is the classic problem that sparked 
    the creation of midi-merge devices?? Is a midi-merge box or switcher
    the only soltuion??   
    
    Thanks for any info,
    Jim

    Jim, I,m not quite sure it's what you need and it depends on what
    you are running on the Atari. 
    I use Master Tracks Pro on a Mac and it can act as a merger by adding
    together what comes in through the IN port and what is being played
    by the sequencer. You could probably then connect the Dx100 to the
    Atari, then from the Atari to the IN of the DX100 and THROUGH to
    the TX7. If you were able to turn local control OFF on the DX100,
    you would then effectively get what comes from the DX100 OUT port,
    going to the ATARI, being mixed with whatever the sequence is playing,
    then going to the IN of the DX100 where notes would play. The same
    notes would also appear at the THRO and go to the TX7.
    
    Confused...???
    
    Ken
    

    I used a topology just as -.1 described for a couple years, only
    with my DX21 as the controller. The only "gotcha" was the lack of
    a "local off" feature on the DX, so I used up polyphony at twice
    the normal rate whilst I was sequencing lines to be played by the
    DX. Playback wasn't a problem...
    
    -.1's solution will work, but may not be optimum unless you have
    the "local off" feature on the DX100. If you do, your problems are
    solved...
    
    Edd

There's an article in this month's Electronic Musician about one-chip midi
processors using I think a 68709 or something like that.  Anyway, the chip is
an 8-bit micro with an on-chip RAM, EPROM and i/o ports.  The little bugger even
has a built-in routine to burn it's own EPROM from another PROM chip.  All you
need is a cross-assembler on your development system and a prom blaster, which
the author of the EM article says you can build yourself.  He doesn't build
anything with it this month, but he has wild ideas about cheap midi filters,
mergers, lighting controllers, etc, etc, ad nauseum.

Sounds like something to watch...

    There's another one called the 6805 which has much of the same
    functionality and I've been using them for about 3 years. Definitely
    the way to go instead of a lot of chips. So far for my MIDI set
    up I have three of them, one as a MIDI controller that converted
    an old analog kbd to a midi kbd, another forms the guts of my 8*8
    MIDI MATRIX and a third is in a box that will ultimately send program
    changes to all my SGUs (still under development). The only problem
    with the 6805 is the lack of non volatile (standby) memory so you
    have to add this externally if you need it. There is a good development
    system for it which includes cross assembler, simulator and Eprom
    blower. It runs on the PC and its available from a company called
    TEC somewhere in the U.S. (I forget where but could find out if
    anyones interested).
    
    Ken