[sdiy] Hi frequency VC clock

[chris]

To be honest, I never really understood why people resorted to use a
4046 as a stand-alone oscillator in the first place.
That's not what it has been designed to do. Ok, it *can* be used like
that, but in the intended PLL usage, the absolute frequency isn't
determined by the osc intself but some input frequency, usually divided. 
The osc only needs to "go somewhat faster with a higher CV", everything
else is fixed by the feedback loop and comparator.

In a way it's similar to a DCO wave shaper circuit - it doesn't have to
be precise, but just needs to create a saw from a digital square with a
precisely given frquency. If you chose to make a stand-alone osc from
such a sloppy wave shaper by adding some feedback, you wouldn't get far
either.

Chris



On Wed, 21 Jun 2017 11:03:02 +0100 Tom Wiltshire <tom at electricdruid.net>
wrote:

> They're all "good" in the sense that they do what they say on the data sheets. It's just that the data sheets vary between manufacturers, and also leave out many significant pieces of information. Notice, for example, that the datasheet doesn't actually provide a equation relating the component values and the VCO frequency. They focus much more on the linearity, but not the actual frequency. Makes sense for a PLL, perhaps, but not so good for us.
> I did find that Texas Instruments chips were more linear than NXP ones. This was discovered by testing a couple of each and writing down a series of data points and then getting the computer to generate a line of best fit.
> 
> I was looking for something that could provide a couple of octaves of exponential frequency modulation at 8MHz for top octave generation. There's an old thread on here "4046 VCO overclocking a PIC" if you want to find it in the archives. Some manufacturers versions of the 74HC4046 claim to be able to go to 10MHz, so those were the ones I chose. I was using a simplified variant of the Thomas Henry exponential control. What I found was that the actual frequency range is extremely variable between individual chips, and is even worse between manufacturers. For something that needed to be tuned to a musical pitch, this was hopeless.
> 
> I still think that what I was trying to do should really be possible, but I wasted quite a bit of time on it and got nowhere, so eventually I got discouraged and gave up. I might go back to it at some point.
> 
> Tom
> 
> On 21 Jun 2017, at 00:03, mark verbos <markverbos at gmail.com> wrote:
> 
> > I’m using a Texas Instruments one. It’s SO16. I hope I can make these work, because I bought a full reel of them!
> > Which brand are the “good ones”?
> > 
> > Mark
> 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy