HJ2743 at denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de
Mon Jan 8 22:02:00 CET 1996
I'm back from holidays, and I'd like to tell You about my adventures
in this time.
First, I got my Morphing Programmer working.
The idea is that of a simple programmer for a modular system,
to store not all (;->), but at least *some* of its parameters, just
to enable some live performance tricks (well, sort of live -
sometimes I play "live", i.e. without overdubs, to a tape at home).
I choose to have 3 "programs" with 6 parameters each.
This might look small at first glance, but believe me You can
make very dramatic changes with 6 parameters on a Modular
where everything is voltage controllable.
Now the point of *my* programmer is that it "morphs". This is
not just a crossfade, nor just a lag or portamento on each parameter,
but it's really a linear interpolation between the parameters. I used
a similar technique as in my Interpolating Scanner, i.e. different triangle
shaped window funktions that overlap, only that now the outer windows
are trapezoid and just the inner window is a triangle. I also used
precision voltage references for good temperature stability of the
I chose to have different funktions for the 6 rows (parameters):
2 parameters have linear pots, one parameter has log pots, one
parameter has bipolar (+/-) pots. The remaining two parameters are
VCAs - You *can* do crossfades between sounds with those, if You like,
but they are more intended to work on envelopes, LFOs and so on.
At the moment, I have patched a setup with a tremolo wurlitzer sound
morphing thru a buzzy lead to a deep self oscillating filter S&H - not
very innovative, but quite impressive to show the capabilities of
New toy! (purists skip this section!)
Last week, my wife and I were on a trip to London. Spending (only)
half a day at Charing Cross Road, I dream of mine finally came true:
I bought my first d*i*g*i*t*a*l synth since loooong. The only one I said
I would, if only it would be affordable. Well, UKP 845,- for a brand new
Yamaha VL-7 *is* affordable ! So maybe I will build less new analogue
modules in the next future :-(
(Still waiting for the VL to arrive - I didn't it carry with me but had it
For the VCAs in the Morphing Programmer, I tried a technique proposed
by Serge T. in an old EN issue. He points out that the linearisation diodes
of LM13600's are designed to work with *currents* instead of voltages,
and the usual 500 Ohm to 1k resistors from the inputs to GND ruins the
performance of the diodes. Larger resistors to a negative voltage would
bring the input more to the current control side and should give a SNR
improovement of several dBs.
So instead the usual configuration (15k to +15V, 2*470R + 100R trimmer
to GND) I used this one: 10k to +15V, 2* 18k + 4k7 trimmer to -15V.
It definetly works. I haven't measured the actual noise figure, but I can
feed a +/- 10V triangle into this circuit with a 15k input resistor without
apparent distortion of the waveform. So maybe someone who has the time
and equipment to actually measure SNR might try this circuit.
Serge proposes 3...4 dB improovement with current sources, so I expect
a little less with enlarged resistors. But the strong point is that You
need any additional components, just different values and a different
Anybody else tried this?
More information about the Synth-diy