FW: [sdiy] MIDI sync experiments...more thoughts

[Tim Heffield]

Thinking about this further, it would appear that there is no guarantee that
the divider chips will start counting with the 24 ppqn pulse that coincides
with the rising edge of the first beat in the measure.  It seems that the
dividers can start counting on any one of the 24 ppqn pulses within the
first quarter note.  According to my calculations at 120 bpm the width of
the 24 ppqn pulses is ~21 ms and this may explain the 42 mS increments of
the time delta when you repeatedly start/stop the DR-5.  I would think that
the start/stop control from the MIDI-SYNC PIC would take care of this but it
doesn't seem to.

Tim

----Original Message-----
From: owner-synth-diy at dropmix.xs4all.nl
[mailto:owner-synth-diy at dropmix.xs4all.nl]On Behalf Of Tim Heffield
Sent: Thursday, September 08, 2005 10:39 PM
To: synth-diy at dropmix.xs4all.nl
Subject: [sdiy] MIDI sync experiments


I have been experimenting with the Blacet MIDI-SYNC chip in preparation for
building a MIDI to analog CV sequencer.  I have observed the following
things if you are considering implementing a similar system.  I built up the
circuit noted in the Blacet docs with the CD4018 and CD4520 dividers and
connected it to my Boss DR-5 Rhythm Section.  I tapped into the LED beat
driver circuit on the DR-5 for a reference clock and compared the edges of
this clock with various points on the MIDI-SYNC circuit.  Here are the
results:

1.) The 24 ppqn coming out of the PIC MIDI-SYNC chip (pin 6) lags the DR-5
clock by ~440 uS.

2.) With a 120 bpm rhythm sequence playing on the DR-5 I noticed that the
edges of the pulses coming out of the 4520 divider do not match the DR-5
clock edges.  In fact, if you repeatedly press the DR-5 start/stop button,
the delta between the positive going DR-5 clock and the CD4520 negative
going (it's inverted) pulse can vary from 0 to over 200 mS.  It seems to
step in 42mS increments, the overall delta changes with different bpm
settings.

I expected that if I am playing a 4/4 rhythm sequence and use the 4 bpm
output on the 4520 divider (pin 4) that I would get a synced pulse for every
beat of the DR-5 LED.  Am I missing something here?  Is this because
pressing the start/stop button coincides with different 24 ppqn clocks and
the whole process starts in different places each time?  BTW I have tried
this same experiment using Cakewalk to drive the MIDI-SYNC and using the
audio metronome from the computer as a reference beat clock with similar
results.

Any suggestions?

Tim