Fast analog sw. as VC resistor
smart at nn.com
Thu Jan 9 01:16:25 CET 1997
> DIY -
> First off thanks to Mark for reviving this idea with such enthusiasm. It
> got me all psyched to do some experimenting...
You're welcome! I didn't notice this note until just now. Thanks for all your
> I went home last night and checked out my ASM-1 VCO. It maxes out at about
> 120kHz with <OK> looking waveforms. When they are square they're alright,
> as you shrink the pulse width they get rather pointy, and as you get the PW
> really low they freak out and oscillate wildly, then they finally lock onto
> +/- V. I connected a wire from the output of the PWM comparator (+/-13.5
> output level) and used that (with a diode) to experiment with a 4053.
Yeah, I used a clipping diode on the comparator output, too.
> that my VCO's use AD712 opamps that have a 17v/us slew rate and this is
> what is both buffering the sawtooth output and actiong as a comparator.
> I wired up the 4053 as an attenuator with 100K resistors. It seems that the
> 4053 can't be set up with Vd = 15v, Vs = 0v, and Ve = -15v... which is
> what I would have liked - is this correct? My data sheet kind of hints at
> it, but doesn't really say. Anyway, after frying a 4053, I set Ve = 0v and
> it worked - this meant that my input signal had to be positive.
I have an RCA data book which has a rather elaborate set of rules for what
voltage can be on what pin. I'll try to figure it out and tell you if +-
15 supplies breaks the rule. For my experiments, I set Vss=Vee=gnd like
you. I didn't try bipolar supplies, but if that won't work, then I'll have
to use another env. gen. design other than the ARP, since it goes negative.
DOH! Oh yeah, I had +/- 12V supplies, not 15V for the experiment. The thing
will eventually have to work on 15 V, because this is what's in the GR300,
and I would like to tap the power off of it, if possible.
> On the
> output I put an RC filter - 1.8K and .01uF. I ran an LFO (level shifted
> into the positive range) into it.
> The output looked great and the attenuation worked very nicely as I turned
> the PW knob on the ASM1. I was impressed. Even with the crappy PWM action
> at this frequency. At lower pulse frequencies it worked even better - too
> low and (obviously) the PWV wave showed on the output.
Wow. That's pretty encouraging. I still haven't tried any time-varying
> I tried it with
> pretty high frequency inputs and got very results (sorry this is so
> subjective - I didn't do any linearity tests).
I think you left a word out...is that "very good results"? That's ok, my
tests were pretty crude.
> The RC network was tuned too low for higher audio frequencies, so, just for
> laughs, I whipped up a 2pole butterworth filter, with a 20kHz cut off for
> the output. This worked well, but it didn't take too much magnification of
> the waveform to see the ripple caused by the pulses on the switch. I think
> you really need either a pulse frequency higher than 100kHz, or a higher
> order filter on the output if you want to use this for audio stuff (like
> the Q pot in a filter).
That's NOT so encouraging.
> I wanted to play around some more and see if I could get a better, faster
> oscillator, so I made a simple triangle osc. around one of those AD712. I
> set it up for 150kHz - unfortunately this is a little too fast, the
> waveform was pretty curvy at the top and bottom. I think a simple, quality
> setup would be using the sawtooth generator from the ASM-1 for your
> reference - as it's very fast.
> It also occurred to me that you could get a log pot reponse (good for the
> ADSR) without too much trouble. A log pot is, afterall just two linear
> sections. So, you attenuate your sawtooth, then run it through an amplifier
> with some zener diodes that will amplify it by 1 when it's below the
> threshold, and by 10 when it's above. Set the threshold right in the middle
> of the wave and presto - you've got a log pot reference!
Wow. How do you do this? Put the zener diodes in the feedback loop? I
haven't heard of that. Sounds great.
> I will be experimenting with a faster VCO and a LM311 comparator tonight
> (unless Mark beats me to it!) - the LM311 is about 6 times faster than the
> Keep up the good work guys!
Looks like you've already passed me up. I won't have time to do more
playing around for a couple of days. Thanks for all the great ideas, Chris!
* Mark Smart *
* Network Technician *
* University Communications Inc. (UCI) *
* smart at medusa.nn.com *
More information about the Synth-diy