Fast analog sw. as VC resistor

Mark Smart smart at
Thu Jan 16 03:02:14 CET 1997

I did some more piddling around with these 4053 switching circuits last
week. I'll try to accurately remember what I did. This weekend I spent
soldering a DB50 connector to the main board of my GR-300 and cutting traces
(I've passed the point of no return!) on it to allow me to send signals off
the board and back on. With this I can experiment with circuits on a
breadboard connected to the GR-300, and also connect to my external
hardware, once I have tried it all.

Brad, I'm anxious to see your circuit, but I haven't been able to get any
of the software I have available to accurately read any of the MIME images.
It's very frustrating. Did you end up using a switched pot like mine, or is
it something different? I can't tell. If you could put the image up on your
web site (actually maybe it's there already, I didn't look yet) that would
make it easier for me to get to it.

Okay, 4053 STUFF:


I hooked up the circuit with a new 4053, and the linearity behaved very
well compared to my first circuit. Conclusion: my nonlinearity problems
were caused by using a fried 4053. 

Chris, I think I figured out why you were not able to get a 0.1 uF cap on
the output of a voltage divider to completely filter the signal...because I
didn't fully report how I made my voltage divider! It was like this:

     / 100K
     |     /
1/3  o     / 100K
4053  \____|__   ________________> output
     o     |         |
     |     /         |
     |     / 100K    |
     |_____|         |
     |              ---
     /              --- 0.1 uF  
     / 100K          |
     |               |
     |               |
    ---             ---
     -               -

In this setup, the voltage divider is only sweeping from +4V to +8V, not as
big of a swing as if you had just hooked the VCP to +12V and GND directly.
When you do, the 0.1 cap IS inadequate to get rid of the weirdness on the
output voltage. Sorry for my incomplete description causing confusion.


I was able to do this ok without toasting anything, at least not right
away! I hooked Vee (pin 7) to -12V, Vss (pin 8), to GND, and Vdd (pin 16)
to +12V. Then I ran my function generator's 100KHz triangle wave through an
LM339 comparator, and the output into the clock of one of the 4053's
switches. Brad, I tried your suggestion of reducing the pullup resistor on
the LM339's output. It helped the up-transition of the comparator to be
more vertical, but the down-transition became sloppy. With my original
3.3k resistor on the output, it has the opposite problem. I'm going to try
again later with a faster comparator chip like the LM319 and see how that
works. The slanted side always causes the voltage divider output to jump
very suddenly at the end...very annoying.

I tried several configurations of an output voltage divider using the
bipolar supply: between GND and +12V through 100K resistors like above,
between -12V and GND through 100K's and between -12V and +12 V through
100K's. These all seemed to work fine. However, when I tried between +12V
and -12V WITHOUT the other 100K resistors, I toasted my 4053! I'm not sure
what's causing this: too much current passing through the switch or what.

So Chris, I'm trying to think of why you toasted your 4053 when you tried
bipolar supplies and came up with a few theories:

      a. You did what I just described!
      b. Your clock signal was referenced to -V instead of to GND. In the
         data book they're pretty adamant about the fact that the clock sig
         for the switch still be between +V and GND even if the signals
         you're switching are bipolar. The LM339 did this for me so I
         didn't have to use a diode like I did on my earlier circuit.
      c. Maybe you reversed Vee and Vss when hooking it up? This is also   
         a severe no-no according to the data book.

I dunno. Anyway, I'm now satisfied that this VC pot can be used to control
my ARP envelope generator circuit since it can handle negative voltages ok.


Thanks to everyone for the ideas on generating a logarithmic sawtooth for
the comparator to produce log pots!

Chris, I tried out your breakpoint generator using a 1458 op-amp. It KINDA
WORKED! The curve on the output almost looked the way it should, but with 
a burp at the bottom: (keep in mind that I was running a triangle wave into
it because I didn't have access to a sawtooth generator) I had to slow the
triangle wave down to like 1KHz so the 1458 could handle it.
I'll try to draw the output:

                      __--      --__
                  __--              --__
      ____-----|\-                      --__

Something like that. I'm guessing the discontinuity is caused by the sudden 
change in the gain of the amplifier when the diode starts conducting. Maybe
the circuit could be modified so that the output at a gain of ten is the
same as the output at a gain of 1 at the transition point, like by level
shifting the signal so that the transition occurs when the input voltage is
zero. How? I dunno. Also, I drew the downgoing side of the wave looking
like it was all the same slope on purpose...that's how it looked. I have no
idea why. Perhaps I misunderstood your circuit description. Or maybe 1458's
just suck.

I liked the other idea of charging a capacitor with a resistor to the
+supply to generate a logarithmic sawtooth wave. Another cool thing to try.


Filter: Rick Jansen's Format Moog-style ladder filter with temperature
regulation. Maybe use VC pot for resonance if I can get it to work on an
audio signal.

Filter CV converter: as I mentioned before, I have to convert the GR300's
CV, which is proportional to the the guitar string's period, to a 1V/Octave
control voltage to drive a filter or anything else like an external synth.
Optimally I would like to have it be stable and accurate enough that you
can resonate the filters and play chords! I found a circuit which does both
the 1/x and log conversion necessary in Electronotes Application Note #114.
This circuit is part of a pitch detector design which operates almost
exactly like the GR300, but without the fancy adaptive filters on the
input. I thought I would use Rick Jansen's temperature-stabilized 3046 idea
on this converter also, to keep it from drifting. The circuit uses two
matched pairs, so I would have to have two 3046's for each string. I have
decided after pricing Analog Devices analog divider chips that this is a
better alternative. Those suckers are expensive.

I'm also going to try out circuits for converting the GR300's
varying-amplitude sawtooth waves into square waves, and octave dividing
these for cool bass sounds. I was playing the GR300 through an octave
divider pedal before and really like the sound. This stuff's pretty easy.

Other pipe dreams: a monophonic bass voice which follows the lowest string
played (like on an ARP Omni, but on a fretboard). OR, a monophonic lead
voice which follows the highest string played (like on an ARP Quadra).
I've never seen either of these things on any commercial guitar
synthesizer, so I want to try them out. I have basic circuits designed for

My, but I do ramble on! Thanks to everyone for the ideas, and Brad I wanna
see your ADSR circuit.

*     Mark Smart                               *
*     Network Technician                       *
*     University Communications Inc. (UCI)     *
*     smart at                      *

More information about the Synth-diy mailing list