[sdiy] Searching for a VC Clock schematic ......

[Magnus Danielson]

From: "Rykhaard D.A.M.I.A.N." <rykhaard at gmail.com>
Subject: Re: [sdiy] Searching for a VC Clock schematic ......
Date: Mon, 26 Dec 2005 10:33:24 -0500
Message-ID: <b6b288a70512260733j17fd07eaoc64c4c8abefddea9 at mail.gmail.com>

> Hey Magnus. :)

Ho Ho Ho Ryk,

> > Well Mag - Harry's gotten me wondering now - with the high speed
> > > possibilities. :)  I haven't paid a lot of attention to the 4046 over my
> > > building years, as there hadn't been any easily available to me, to play
> > with.  But I now DO think that my 1 supplier that stocks CMOS DOES have them
> > in stock.  Soooooo. ;)  Time to start looking again, for schematics of stuff
> > to do with them. :)
> >
> > If you toss a TL074 ontop of your 4046, you can get a triangle-wave (as BJ
> > demonstrated) and you also got the square-wave output. This is a pretty
> > compact triangle/square oscillator which can then be used to create sawtooth
> > and sine fairly quickly. You can then get PMed-PWM and non-PMed-PWM as you
> > feel comfortable.
> 
> 
> Here is one place, where I'm going to be stopping myself from using Quad
> opamps:  LFOs.  Due to the comparator being one of them, in an LFO.  The
> interruption that it CAN seem to cause in other opamps within the quad, is
> quite noticeable. :(

And you have tossed in a bypass capacitor?

> In this case, it'd be acting as a buffer - but I still have concern about
> it's possibly effecting the other opamps, with it's buffering the output
> from a Schmidt Inverter - who'll ve having fairly sharp edges / quick
> changes in voltage.  :-/

If you don't load it too hefty you will not be in such a bad situation. Also,
these are not sitting in the oscillator loop, which is much more trickier since
that would change the state of the oscillator.

> With it's main purpose being a Clock, I could still use the Triangle as an
> additional LFO of sorts.  It'll almost always be up in the 10,000's and
> higher of hertz, due to the 4024 dividers that they'll be running. :)
> FM effects could be used from it, as well. :)

Well, I was only trying to give you crazy ideas. You do know that with a little
of fiddeling you could generate lower-rate triangles from the lower rate
squarewaves and the triangle you start with?

> > Well here - I was going to put the 3080 in the loop around the 40106 that
> > charges / discharges the cap, working with the squarewave output only for
> > driving other CMOS devices.  So the output of the 40106, would go into the +
> > input of the 3080 (through a 100k/220r divider), and the output of the 3080
> > through a diode (to block the negative output) would be connected to the
> > Grounded Cap, and the input of the 40106.
> >
> > Why the diode? Just connect the CA3080 between the same supply lines as
> > your 40106 and you don't have a problem. Also, the 40106 already has such a
> > diode anyway, two diodes going to each supply line actually.
> 
> 
> Hooking up the 3080 between +V/Ground hadn't occurred to me at first. :)
> With this though - wouldn't I have to setup 1/2 voltage divider for the
> reference point to the - Input of the 3080?  (Here's where I'm stepping
> beyond my known theory - as I don't ever build anything that runs on a
> 'Single Supply'.  :-/

Yes, it can be good to have such a reference and especially in this case.
The only real thing that happend was that we shifted the voltages upwards.

> Or - would I just have the 3080 negative input, tied directly to Ground?

Not in this case.

> I haven't looked at the 40106's internals in my CMOS databook yet.  It HAS
> it's own protection from -V's, built in??  That would be very cool, if so.
> :D

It does, it also protects towards V+. But, if you are running them on the same
supplies, this will not be a point of worry anyway.

> > (I realized tonight, whilst out socializing - that I'd have to bring the
> > > 3080 output back up, about 500 times, after it's 500 times reduction,
> > for > the 3080's safety.  So that'll require an opamp between the 3080 and
> > the > Cap/40106 input.  There goes chip count up +1 for each clock.)
> >
> > ? Me no comprendo (OK, there's about all my spanish (except popular
> > swear-words then))
> 
> 
> Haha.  That's okay. :D  I know how to say 'Hi', and that's about it. :D
> 
> I'd figured on using the stand input resistor divider to the 3080.  100k
> resistor between input and +input, with a 220r resistor to Ground, from the
> +input to Ground. That drops the Input voltage level by 454.5454etc.
> So - as with some other circuits that are voltage controlled - I figured
> that the output would have to be amplified 454 times, to come back to the
> original voltage level coming into it. :)  Hence - an opamp being required
> at the output, before the cap. :)
> 
> Noooo?  Am I missing something? :D

Oooooh yes, first of all, you scale by the choice of capacitor, since the
output current (and not voltage) of the 3080 will be scaled down, but then
again, you can scale the output current *alot* on the 3080 by the Iabc input,
so it is more a matter of considering the maximum current and ensuring the
right range. Beleive me, you do *not* want to toss in an op-amp there, it will
only help to confuse the issue. You can, but you have to modify the world a bit
after that, but I would not do it as the first attempt.

To some degree one could even question the damping, it may even be better to
have it go full-blown into the input, since then you would to some degree make
the difference in high and low voltage of the 40106 output be dampend by the
non-linearness of the input. You do not care about being very linear there,
since it sees two levels only and a stead voltage.

> > Ok.  So from what you've said - my idea SHOULD work. :)  All V inputs to
> > the 3080, would then go through an Expo. first and then into the 3080 - and
> > possibly - I could get 1V/octave speed control??  :)
> >
> > Yes? Why not? The CA3080 datasheet even demonstrates an oscillator with
> > the range of 1:1000000 and what you are trying to do is basically a variant.
> > But I don't think you have heard my message fully: DROP THAT DIODE
> 
> 
> Haha!  :D  Understood - now. ;)

Good.

> (at least mentally). Ah well, there will be one there anyway. You can fully
> > resolve that issue by having the 3080 and the 40106 running on the same
> > supplies, then the protection diodes will not bias up at all.
> 
> 
> That's right!  I never even thought, of the forward voltage drop. :O
> But here - I could use an opamp based rectifier, that takes care of the
> voltage drop, as WELL as amplifiying the output.  That would take care of
> both - even though I apparently don't have to worry about the diode, as the
> 40106 IS protected.  No?  :)

Why do you care about voltage drop? Hook the power-lines up, and the 40106
will invert current before you reach the ends. Scale the current with Iabc and
select the capacitor to make the frequency range reasnoble. Should work well
enought.

> No, you can drive your 4046 on an expo, it has a current input! ;O)
> > Pin 12 is really a current input, which directly hits a current-mirror at
> > Vcc. So, your garden varity expo-circuit would fit quite nicely into that
> > pin.
> 
> 
> As Harry'd said though - not accurate enough, for 1V/octave.  Still though -
> with my wishing to keep the circuit as simple as heck, I can toss away the
> '1V/Octave' requirement in exchange for simple circuitry. :)

Hehe... ;O)

> If you have problem reaching higher frequencies with your 40106 you are
> > probably running into the 40106 output current limit. Reduce the capacitor
> > value to get around that. Oh, you *do* want to isolate your core from your
> > load so use one of the inverters as buffer.
> 
> 
> Right.  That would drop my Clock count from 6 to 3, but that's ok.  I should
> have enough clocks with 3, anyways.  As well as external input.
> To keep circuitry down - I could go with 2 clocks, and a buffer each for 2
> external inputs.  Hmmmm.  (Still developing this thing, all over the place.
> :) )

Know the feeling. ;O)

> > It also reminded me - that I can just run the 3080 off of +15/Ground!
> > > Thereby - no negative voltages for the CMOS. :D  (Unless I ran it off of
> > +/- > 7.5V which I don't wish to bother with. :P )
> >
> > Exactly! This is what I have been trying to hint. What you could do is to
> > run both the 40106 and the CA3080 on +/- 5V or something. That will work too
> > and you get a balanced waveform directly. I do recommend this!
> 
> 
> I didn't want to go this route, as there'd be extra requirements for
> dropping the power supply voltages to both of them, with LM329's (for +/-
> 6.9V) or 5V regulators.  (A fair bit of wasted current, from the +/- supply,
> heating up the voltage regulators.  :-/  )

This thing shouldn't run that hot anyway.

> I'm going to prototype (TODAY, FINALLY!  Haha.) around the +15/Ground PS for
> them. :)

OK. Good luck!

> Here's to hoping that you had a wonderful day, on Christmas and received
> lots of wonderful stuff. :) :)

One book. Non-DIY I might add. Ah well. ;O)

We have snow again, that is always something. ;O)

Cheers,
Magnus