Conference napalm::commusic_v1

    Edd - you have a hardware problem.  (Since I just spent heap big time
    on phone with Roland re: MKS70 problem [see 2050], I can say this with
    certainty.)  IF the 30 is like the 70, there's no "software reset"
    capability, so I doubt there's any RAM to reload.
    
    The MKS30 is a 6 note beastie, and cycles thru each "osc card set" for
    every subsequent note played. To verify that it's always a certain
    note, power down/back up, then play a slow C scale, noting which is
    'sticky'.  Once you find the sticker, veryify by playing 6 more notes. 
    If it's consistent, it's definitely hardware.
    
    BTW - if you *do* call Roland, ask for "Mike in Service".  Anyone else
    I've talked to is a bozo ... and according to a service rep friend of
    mine that I trust implicitly, he's the only one out there who knows
    what's happening.  Good luck.
    
-b

Edd expect that if it isn't under warranty that they'll send you to Daddy's as 
the local authrized repair dealer. They only service gear that's still covered
by waranty.

I've been there....

dbii

    Brad: You've got the symptoms practically right. The first note sticks,
    then the 7th, then the 13th...
    
    Some tests last night showed that both DCOs are ALWAYS injecting some
    sort of audible signal into the audio chain. That's the 'clicking'
    I hear before I hit the first note. I can play with the DCO range
    or waveform and get that noise to predictably change...
    
    It seems that one *voice* doesn't want to release...
    
    Edd the Bummed

    ...just got off the phone with Jeff at Union Music. (BTW - he's 
    recovering from a car accident.)
    
       Union has factory trained techs doing the repair work on the
    premises...
    
    Edd

I'll ship this text off with the unit...
    
                         *   *   *
    
The following symptoms occur BEFORE a single note-on is presented to
the unit via MIDI...


1. Immediately upon power-up a steady state (non-fluctuating in pitch)
   tone appears at the audio outputs and the headphone output. The tone 
   varies in pitch across power cycles, but does not change within the
   realm of one cycle.

2. Within approximately 1 second after power-up, the LED will display the 
   currently selected MIDI channel. As soon as the MIDI channel is displayed
   a white/pink noise is heard at the audio outputs. This noise ends as
   soon as the display changes to the default patch selection (1-1).

3. Once the machine has completed its' power-up routine and defaulted to
   patch 1-1, a semi-rhythmic 'ticking' can be heard. The tempo of this
   'ticking' varies. It is directly related to the state of the 2 DCO's
   selected range. (This can be verified by changing either parameter 21 or 23.)
   This sound is also related to the currently selected waveform for each
   DCO. (Changing parameter 22 or 24 will cause the 'ticking' to change
   to a buzz. In the case of DCO-2, selecting option 4 will cause a noise
   component to be present in the audio output. This shouldn't be a surprise,
   option 4 is a 'noise' waveform.) If the sawtooth or square waveform is
   selected, changing parameters 21 or 23 will cause an expected change in
   pitch directly related to the selected range.

   Changing patches will cause the sound to change also, sometimes it's
   a slow tick, sometimes a low hum, etc., but playing with the DCO ranges
   and waveforms always causes the noise to change in a predictable fashion,
   i.e., changing the DCO range from 16' to 8' will cause the ticking to
   go twice as fast or the buzz to raise one octave.

4. The malfunction manifests itself in both oscillators. This may not be 
   readily apparent if the value of parameter 18 (source mix) is extremely
   low or high. A value of 50, which balances the 2 DCOs equally, will reveal
   the sound caused by each DCO.

The following symptoms manifest after the first note is played...

1. The first note to be played over MIDI will continue to sound, at a reduced
   amplitude, even after the note-off message has been received. Subsequent
   notes, played 1 at a time, appear to function correctly but the original
   note continues to sound.

2. The note will remain active across patch changes. Naturally, the timbre
   will change, but the note continues to sound.

   Don't be fooled into thinking the note has stopped if you select a
   patch with no sustain and a very quick decay/release. Change patches
   again to one with sustain and the note will be there.

3. The original "stuck" note can be silenced by playing a chord/cluster of
   at least six notes. The original note will stop, however, one of the six
   notes in the chord/cluster will now become stuck. While I haven't been
   able to examine the serial MIDI data stream, my gut reaction is that the
   last note in the serial stream is the one to get stuck.

   Being 6 note polyphonic, this *seems* (to my untrained eyes) to indicate
   the problem lies in the circuitry that generates the 1st of the 6
   voices. This circuit seems to always be active, changing only when the
   last note priority voice allocation scheme causes it to sound a new
   note.

Other info....

   I've eliminated cables and the keyboard controller as a source of the
   problem by substituting known good devices. 

   I also observed that the component labeled "D2" on the smaller of the
   2 printed circuit boards (not the master board) gets untouchably hot
   after a short period of operation, while the similar looking component
   "D1" stays cool. This could very well be normal (Why would I look if
   something wasn't broke?) but I thought it would be worth mentioning.

    FROM:  Roland
           Product Repair
    
    TO:    Ed Cote
    
    Dear Mr. Cote
    
    Upon receipt of your unit and review of the extensive (and quite
    complete letter accompanying the unit, we have come to the deter-
    mination that your unit is broken.
    
    Thank you for your interest in Roland products.
    
    Sincerely,
    Mr. Fix It
    

    Why does -.1 seem prophetic?
    
    Edd

    Aw - they'll do you right.  I just spent � hour on the phone the other
    day buttering up the owner for DECMS. 

    You *did* mention that you're from DEC, right?  (Or do they know you by
    sme - er, sight?) 

-b

    Yeah, Union knows both my face and my employer. They've always treated
    me good in the past.
    
    I just get the willies when I have to put my faith in someone else to 
    do what I know I can't. Being dependant on somebody else is not a
    character trait I've developed...
    
    Edd

    
    	Ive been inside the beast. There are effectively 6 channels
    	of sound circuitry that include some custom not-too-many-pins
    	chips. One per channel. Might be Solid State Music chips. 
    	Sorta doubt it, cause they were USA, Roland isnt.
    
    	This may be as easy as a bias resistor for the VCA is broken.
    
    	Or some other VCA problem, but then, that is probably inside
    	that custom baby....tsk tsk.
    
	Goodum Luckums.
    
    	(The MKS-70 is a monster compared to MKS30)
    
    	ron
    

    Edd,
    
            I've had a -similar- problem with my Matrix 1000. When I would
    first turn it on, there would be a note playing....although its
    relative amplitude to the notes that you'd play was small. It was
    always the last note you played that stuck. After about 10 minutes of
    warm up, the problem went away. I finaly had the thing sent out. It
    was not under warranty by the time it developed this personality
    trait. Since I got it at Wurlies, I had Eddie Fritz send it to the
    shop (eddie raves of this shop in Connecticut somewhere). It came
    back.."NO PROBLEM FOUND", a very bent rack ear, and tons of dirt
    marks all over it. Nice place Eddie. I took it home,,,same problem.
    I took it apart. I found that when I fanned some air over the board,
    the stuck note was VERY loud. As load as regular notes. I traced thed
    problem down to two thermistors that are placed right near the
    oscillators. I presume they are part of some sort of temperature
    stability circuit. 
           If I touched one of the thermistors, the note would be real
    loud. If I touched BOTH thermistors, the note would be soft. My
    conclusion is a missmatch in the thermistors. My sulotion, do
    nothing. I can live with letting the thing warm up for 10 minutes
    before I use it. 
    
        ....just maybe your problem is similar....???
    
    
           Frank
    

    Wouldst that mine should be so obvious!!
    
    This is a little more than a stuck note, it's a stuck *voice*. The
    actually note sounding can be changed, but *some* note will stick.
    
    ...wish I had a service manual. My board blew up on me once. (I mean
    *smoke* comin' from the speakers!) With a manual, a DVM and some help
    from this conference I located the offending 5W resistor and had myself
    back in business for $0.79.
    
    Edd
    

    ...just spoke with Steve at Union. He believes he's got it narrowed
    down to an *in stock* $15 chip.
    
    He also thanked me for my diag's. Said it saved me some good $$ as
    he had a good place to start looking instead of a tag that simply
    said "broke".
    
    Edd

My MKS-7 has started acting funny.  It's the two Juno channels, they sound like
they are heterodyning, or being modulated by some audio signal.  The bass and
drums work fine.  It's a pain since I like the sounds and the thing probably
isn't worth having fixed.  I plan on poking around inside it this weekend. My
warrantee has expired, of course.

	Eirikur

    RE .14
    
    I bought an MKS-7 from LaSalle in Boston a few months back, and then
    found a problem with the Juno voices when I got it home and let it warm
    up for a while. Seemed like the filter on one of the voices got stuck
    open after about 20 minutes, so every sixth note was very bright
    sounding. LaSalle fixed it for me since it was still under a limited
    (demo) warrantee. You might check with them, I believe they did the
    work in house and didn't just send it to Roland. 
    
    Rob

    Well, one custom IC, 1 transistor, 1 week and $65 later, I'm a happy
    camper...:^)
    
    Kudos to Steve at Union.
    
    Edd