Conference napalm::commusic_v1

    The way your keyboard PROBABLY works is shown below:
    	
    	
    	
    	-------------------VISIBLE KEY-------------------o pivot
    
    	|	|
    		|     O contact #1
    		|/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/o    coil spring
    		                                             
    		      O contact #2
                                                                
    The spring is gold-plated, as are the two contacts.  What the keyboard
    actually uses to detect velocity is the time difference from when
    the coil spring breaks contact with contact #1 and makes contact
    with contact#2. Long time--> low velocity, short time--> high velocity.
    	
    If contact #2 were worn or bent downwards, the time it takes to
    connect to contact#2 would increase- and the velocity would look
    "low".  Sometimes there is an adjustment screw (or screw + shims)
    to set the height of contact #2.  You want to decrease the distance
    between contact#2 and contact#1 (but only in the middle of the
    keyboard.)	
    
    (The above is how Ensoniq ESQ/Mirage and Seil DK-80 keyboards work- it
    probably applies to your KX as well.  If it's not in warranty, you may
    want to open it up and take a look.) 
                                         

    That's very interesting.  I don't imagine that many keyboard users
    are even aware of the phenomenon.  It didn't even occur to me until
    you mentioned it.
    
    The DX (and the KX) does work as Bill described.  I disassembled
    my keyboard once just for the grins.  I also managed to mangle some
    of the moving metal contact strips in the process, so I proceeded
    to aquaint myself with the mechanism.   (Send me mail if you want
    caveats on removing the keys).
    
    I can't imagine how you would go about calibrating the contacts again,
    but you could certainly close them up, as Bill suggests, to restore
    some zing to your keyboard but you might sacrifice some of the
    lower velocities.  Then again, we don't need no steenkin' dynamics,
    since the fact that rock and roll is always audible over the road
    noise made it what it is today.
    
    This is just like adjusting the action on that old Rhodes.
    
    Steph                          
    

re .1

I agree. The Kurzweil K250 uses leaf-spring contacts, but the principal
is the same. The metal becomes fatigued over time and the space between
the upper and lower contacts determines the ability to generate the
high velocities. The bottom contacts of the heavily used keys will
gradually bend downwards, causing the symtoms you describe.

dave

    'stonehands' ?? - how about 'well-developed' ?
    
    What's a MIDIMAX? some kinda MIDI event recorder/display device?
    
    I bought my KX88 new in mid-86.. going on two years of solid (see
    top line) playing.. still routinely see velocities (in Performer's
    Edit window) of 120+ without consciously pounding. But what's true
    for a KX76 is true for a KX88.
    
    Both the Emax rack and the Kurzweil rack allow a programmable keyboard
    velocity curve. It might be possible to compensate in this manner.
    Also, Performer allows creation of a velocity curve on 
    sections of the 'keyboard'.
    
    Thanks for the tip/warning !!!
    
    karl    

RE: .5

    Ok.  Well-developed.  You must be pre-sales.  {smirk}

    MIDIMAX is a piece of Atari software written by Steph Bailey that
    collects MIDI note info (velocity range, pressure range, etc) and
    displays the information (in summary form) back to you on the monitor.

    The ESQs also allow "programmable" velocity curves (soft,medium,hard)
    as do the Yword TXes (0-7 attenuation), but I would rather not have to
    deal with this. 

-b

    One crude quick way to calibrate the keyboard is to adjust the width
    between the contacts using a caliper. Granted, this doesn't take
    everything into consideration, but it's cheaper and quicker than
    trying to concoct some kind of calibrated key presser. You can probably
    get some cheap calipers for under $20, and results will most likely
    be satisfactory, plus you won't have to have the thing powered up
    while strewn all over your workbench.
    
    						John.

    probably not.
    
    because the tension on the contact must be taken into account.

    The higher the tension (read BEND), the further the key can
    be depressed WITHOUT the wire leaving the at-rest (up) contact.

    Messing with these can be difficult. We used J-wire gold plated
    contact wires for ARP keyboards. Getting repeatable TIME of FLIGHT
    (ie, velocity) is possible. Proper setup is very key in this. not
    just the distance between the contacts.
    
    The Kurz Midi-board uses conductive rubber that conducts between
    two contacts. Gives poly aftertouch. BUT ALSO Ive heard that the
    rubber loses its 'twang' (DEC STD 004Q...no,seriously) after use. 
    
    
    rr    

    re .8:
    
    I don't understand. Does the spring bend in flight or something?
    I can see how the spring tension might effect the hit displacement,
    but I don't see how it effects the velocity. Can you explain this
    a little more for us ignorant folks?
    
    						John.

    re .8 and .9 - I don't know how to explain it any more clearly than
    Ron did - if the spring is pressing "too" hard against the key, then
    it will "follow" the key further (i.e., its "resting" or natural
    position when not displaced by the key is further along the key's
    path of movement).  The net effect is the actual distance over which the
    velocity is measured will be shorter than the distance assumed in
    the computation (i.e., the distance which is divided by the
    time to get the velocity), making the velocity appear higher.  I.e.,
    because the "top" switch opens later than it should, or the "bottom"
    switch closes earlier than it should, the time that's measured is
    shorter than it should be, making the key appear to be moving faster.
                     
    len.
    

    thanks len.
    
    I wonder what the new ensonic kbds use for switching and
    aftertouch. Alll that clicking you hear sorta points to
    a slightly different mechanical switching....
    
    rr
    

    
>    re .8 and .9 - I don't know how to explain it any more clearly than
>    Ron did - if the spring is pressing "too" hard against the key, then
>    it will "follow" the key further (i.e., its "resting" or natural
>    position when not displaced by the key is further along the key's
>    path of movement).

    In the diagram in a previous note, the spring is perpendicular to
    the key's path of movement. By this do you mean that the spring
    is under compression ( seems reasonable to me, rather than being
    under tension ), and that the spring has some lateral displacement
    dependent on this? Are you saying this causes a nonlinear function
    of spring displacement vs. key displacement. If I understand the
    diagram correctly, the resting position and the depressed position
    are not used in the velocity translation. The are two active points
    where the spring breaks with the upper contact and then makes with
    the lower contact, somewhere in the middle of the key travel. By the
    diagram, it looked as though the spring was much longer in comparison
    to the distance between the contacts. I expect some nonlinearity
    with a spring under compression, but I was thinking that this would
    not be a significant cause. I was thinking that the contacts themselves
    are likely to be more malleable than the springs and therefore subject
    to greater deformation. I was also thinking that the springs would
    follow a fairly close approximation of a linear function in relation
    to the displacement over the contact width only.
        
>    The net effect is the actual distance over which the
>    velocity is measured will be shorter than the distance assumed in
>    the computation (i.e., the distance which is divided by the
>    time to get the velocity), making the velocity appear higher.  I.e.,
>    because the "top" switch opens later than it should, or the "bottom"
>    switch closes earlier than it should, the time that's measured is
>    shorter than it should be, making the key appear to be moving faster.
    
    This explains how movement of the contacts effects the velocity,
    which I understand. What I don't understand is how the spring, over
    the distance between the two contacts only, changes. Perhaps if
    you described how the spring is fastened on either end. I am really
    interested in this ( as if you hadn't guessed ), as I may have the
    same kind of problem someday. I am confused as to why I can't repair
    it without special tools to a reasonable approximation of flat velocity
    response. I'd open mine up, but it's still under warranty. Thanks
    in advance for anything you can tell me. I don't think I'm being
    dense, I just don't understand how the spring itself effects velocity.
    
    						John.
    

    
    I think the deal is that the spring,  its compression and or lateral
    displacement really has nothing to do with the timing involved.  Fur-
    thermore,  I think alot of people are under the impression that  the
    spring is some tightly coiled piece of wire,  and that just  is  not
    the case.  In general,  the springs used inside most synth keyboards
    is a thin bar of semi-flexible metal that is always under some  kind
    of compression to keep it in place.
    
    I think the point that they are trying to make is  that  the  actual
    contact,  which is usually a very thin piece of  metal  (like  thick
    tin foil) has a tendency over  the  months  to  bend  slightly  away
    from the key itself,  in  a  downward  direction.  The  distance  is
    slight, but the time is so short during key depression that it makes
    just enough difference to impair the velocity sensitivity.
    
    _sjz

RE: .13

    In terms of Yword boards, Zach is correct.

    I think there's probably a (not so) simple timing circuit that measures
    the delta between (top switch open) and (bottom switch close) events.
    There is a single piece of contact that causes this switching, like so: 

	    FRONT            SIDE

	    ----               o	top switch base contact
	     ::            ---------	switch throw contact
	    ====           o		bottom switch base contact

    Since the bases are vertically offset (ie, the bottom base is closer to
    the front of the keyboard than is the top base), the switch throw
    contact can become bent.  Granted, the bend may be slight, but we're
    not measuring much here, either. In any event, this will eventually
    result in a deeper key depression being required both in order to get
    the top "switch" open (trigger) and the bottom "switch" closed. 

    I'm no hardware heavy, so I could be all wet from a practical
    standpoint.  This is simply the way that I best understand it. 

-b

    Well, I'll be darned. That offset explains everything. One picture
    is worth a thousand woids.
    
    OK, let's try again, how about a caliper on the key itself, rather
    than on the contacts. You would also need an ohmmeter to check when
    the contacts are touching. It would be a two pass operation. First
    record all the displacements for all the contacts. Then remove the
    keys. Triangulate the adjustment to specific contacts and adjust
    accordingly. You might have to study up on your geometry.
    
    						John.

    Uhhhh.... the Ensoniq keyboards do NOT have a flat piece of metal
    for the moving contact.  It's a gold-plated metal coil spring, with a
    fibrous core stuffed inside, and it's under tension, not compression.
    The fibrous stuff may be elastic of some sort (to support the gold
    coil)
    
    Yes, tension.  This generates a wiping action on every open and
    close, because the spring "wraps around" the nonmoving contact
    significantly.  When the key is in the UP position, the moving contact
    is shaped like a checkmark (concave upwards).  When down, the moving
    contact is shaped like a checkmark, but concave DOWN.
    
    (and only one end of the spring is attached to the moving key- the
    end near the contacts.  The other end is fixed to the keyframe.)
    
    I *have* seen keyboards that used a "finger" of springy metal to
    make/break switch contacts, but they were all non-velocity-sensing.
    No reason why they couldn't be made like .13, though.  And .13 does
    explain well how as the metal finger-spring got tired it would
    make the keys less sensitive to velocity.
             
    -------------------
    
    Time for some bent-tip needlenose pliers, eh?  Wait a bit- needlenoses
    will probably make a kinked mess of those finger contacts.  Try taking
    a popsickle stick and cutting a slot crosswise in the end, like: 
    	
    		 __________________________________  __
    		/				   ||  \
    	       (	     			   --   )
    		\______________________________________/
    
    to make a tool to make gentle adjustments more easily (and with
    less chance of putting a kink in one of the finger contacts).  The
    Pianette our lead keyboardist has has such a tuning tool (made from
    fiberboard, I think) in a little pouch next to the reeds.  Factory
    original!
    	
    Enjoy...
    	
    	-Bill
         

    Well, I finally got my return phone call from Yamaha.

    They said that the KX76 has the same velocity restrictions as the old
    DX7s.  They also said that the KX88 could transmit greater velocity
    values because the keyboards were different (even though the internal
    circuitry was the same).  These same restrictions apply to all non-88
    KX type boards (KX5,1,etc). 

    Of course, this is pure poppycock.  I wonder if these guys really know
    what's going on?  I get a different story from every support person
    that I deal with - from ANY company.  Sigh.

-b

    Gee, going back thru all these old notes is fun.  Lots of loose ends.
    Anyway, I figured I'd go ahead and post the conclusion to my 'saga'. 

    There is no problem with the keyboard, and Yamaha's tech support is
    brain-dead.  The keyboard has no trouble outputting 127 velocities
    consistently - AS LONG AS THE STAND IT SITS ON IS STABLE. 

    I have a USS 3-tier A-fram stand with 54" struts, and the thing shakes
    a little bit, but not enough - so I thought - to affect performance.
    WRONG.  If I have the KX sitting on a scissor-stand (X stand) or on the
    edge of a concrete stage, I have absolutely no trouble getting 127 out
    of it consistently.  Turns out that the heavily used keys are in the
    middle of the 54" strut, right where it tends to bend/give the most. 

    Amazing stuff, these MIDI toys.

-b