[sdiy] LFO VC Skew?

David G Dixon dixon at mail.ubc.ca
Sun Feb 23 04:10:55 CET 2020


Thanks, Ben!

I just took a walk and reflected on this circuit (which is also acting a
little funny in simulation), and I have come to a tentative conclusion: The
problem is probably because I'm using a TL074 op-amp as a comparator in the
tri-square, and it is probably a little too slow to switch on the fly-back
in saw/ramp mode.  No one would ever use an op-amp comparator in an actual
VCO circuit (at least, I never would).  Because the comparator is a little
slow, the waveform is allowed to go slightly higher than the threshold, and
this slows it down.

Maybe I'll switch it out for a 311 and give that a try.  What I do know is
that the issue has absolutely nothing to do with the 2164 circuit, which is
performing admirably.



-----Original Message-----
From: Ben Stuyts [mailto:ben at stuyts.nl] 
Sent: Saturday, February 22, 2020 4:17 PM
To: David G Dixon
Cc: SDIY List
Subject: Re: [sdiy] LFO VC Skew?

Hi David,

Hope you're doing bettr now. Take care of yourself!

Thank you for taking us through your  experiments, I enjoyed your writings
and seeing the results on youtube.

Ben


> On 22 Feb 2020, at 21:58, David G Dixon <dixon at mail.ubc.ca> wrote:
> 
> Hey SDIY Team!
> 
> So, after a night of hospitality at Vancouver General Hospital's Emergency
> Department to treat my extreme nausea and dehydration, I was able today to
> make all the planned changes to my stupid morphing LFO.  Here's what I
did:
> 
> 1) I rearranged, and kludged a couple of diodes to, each of the control
> opamps to limit the output of those opamps to 5V.  This prevented loss of
> the signal at the extremes of CV.
> 
> 2) I changed resistors on the VCA linearizing opamps to mimic a 3k fixed
> resistor into a 100k pot (5V into 1M), instead of the 1k resistor I was
> mimicking (5V into 3M).  This softened the transition to pure saw or ramp.
> 
> 3) I decreased the integrator capacitor from 332 to 681 to speed up the
LFO.
> 
> After making all these changes, I think I'm done, and here is my official
> assessment:
> 
> While this might be great for morphing an LFO, it is probably not terribly
> useful for morphing a VCO.  The problem is that, when the waveform
> approaches the saw or ramp, it "snaps in" to that, and in the process gets
> slightly bigger and droops by about a whole tone.  This has nothing to do
> with the VCA circuit per se, I don't believe.  It is just the nature of
the
> beast.  I looked at my existing morphing LFOs which just have pots, and
they
> did a bit of the same thing -- not as extreme, but enough to render them
> out-of-tune if they were VCOs.  I could probably play with all the values
> and the gains and the cap and make it a little better, but I don't believe
> that this circuit will ever be able to transition from saw to tri to ramp
> without sudden changes in frequency at the extremes.  Which is really too
> bad, because every other aspect of its performance is perfect.  The saw
and
> ramp are really clean and sharp.
> 
> One way around this (partly) would be to put in an even larger faux fixed
> resistor (decrease the 1M resistors to something even smaller, like 499k).
> That would make the transition to saw or ramp less hard, but eventually it
> won't reach saw or ramp, and if it doesn't get far enough (i.e., if the
> flyback part of the saw doesn't get vertical enough) then there is little
> timbral change (not enough overtones) and it becomes a bit useless.
> 
> I guess the question is, is there a compromise where the endpoints are
> "sawey" or "rampey" enough, but not so hard that they pull the waveform
into
> a lower frequency?  Maybe I'll have one more go, just to see where that
> critical transition is.  Maybe I'll be scientific about it and compare the
> endpoint frequencies with the triangle (midpoint) frequency as a function
of
> the fixed-resistor-mimicking resistor value.
> 
> 
> 
> -----Original Message-----
> From: Synth-diy [mailto:synth-diy-bounces at synth-diy.org] On Behalf Of
David
> G Dixon
> Sent: Thursday, February 20, 2020 8:11 PM
> To: 'Jason Proctor'
> Cc: 'SDIY List'
> Subject: Re: [sdiy] LFO VC Skew?
> 
> So, after I made that video, my gastritis/duodenitis/pancreatitis flared
up,
> and I got extremely nauseous, so all I could do was sit in my easy chair
and
> stare at the TV.
> 
> I felt a little bit better today, but still crappy.  In any case, I have
> added four diodes to the control circuit in my simulation, and can state
> with confidence that this will eliminate the loss of signal at the
extremes,
> even if the CV signal is too big.
> 
> My plan is to kludge these diodes onto my board (which will be a
challenge,
> because I made the layout rather tight) and confirm, and make a new video.
> Hopefully, I'll feel well enough to do that tomorrow.
> 
> 
> 
> -----Original Message-----
> From: David G Dixon [mailto:dixon at mail.ubc.ca] 
> Sent: Wednesday, February 19, 2020 6:26 PM
> To: 'David G Dixon'; 'Jason Proctor'
> Cc: 'SDIY List'
> Subject: RE: [sdiy] LFO VC Skew?
> 
> 
> I think this thing is basically good to go, but I do need to add a little
> something more to limit the behavior to the points where the pot would be
> either fully CW or fully CCW.  In the video, it is obvious that if the CV
> signal is too large or Morph pot isn't well centred, then the morphing is
> allowed to go past the extreme points, and the waveform disappears.  I'll
be
> messing around with this tonight.
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy




More information about the Synth-diy mailing list