[sdiy] More VCO fun

[Ian Fritz]

At 10:15 PM 12/7/2009, David G. Dixon wrote:
> > [off-list]
>
>Actually, it wasn't, but that's OK!
>
>
> > I'm assuming your circuit is very similar to Ian's, as I'm really
> > commenting on that (I don't think I ever saw a link to your circuit).
> > Both the switch itself and the control are not ideal for what you want
> > to achieve.  The JFET used is a depletion transistor, which means it
> > is switched on when V_GS becomes zero (or indeed very slightly
> > positive, but not so much that the G-S-diode opens).  After reset, the
> > comparator is held at -12V (since both pin 3 and 6 are tied to this).
> > This is only good if you are using a transistor with a high cutoff
> > voltage that actually needs such a high V_GS to be switched off.  The
> > one you are actually using has much lower cutoff and should not be
> > driven so deep into depletion.
> >
> > During reset the gate capacitance will be discharged through R23-R24
> > to ground only.  Again, if you have a transistor with a high cutoff
> > voltage this is probably going to work out, otherwise the transistor
> > may never actually switch fully on until quite some time has passed
> > (the gate RC is about 10-15ns here and is getting larger as zero V_GS
> > is approached).  That is if there wasn't the "timing" cap, which
> > increases the effective RC by about 3 times.  The simulation should
> > show this... anyway, for a JFET with low cutoff you need to come much
> > closer to zero V_GS to really switch them on and I suspect you never
> > really do that before the comparator calls it a day and pulls down
> > again.
> >
> > It would be much better to use a comparator that has a push-pull
> > output stage or follow up with one.  Alternatively, if you want to
> > stay with that circuit feed some ~mA current into the output of the
> > LM319 via a resistor from +12V and put a Schottky diode from output to
> > ground.  This way the comparator will switch faster, the JFET will
> > open much better and you should be able to reset in quite a bit under
> > 1µs.

Since you ar commenting on my circuit:

1 -- What you call "my" circuit is actually Terry's circuit with some newer 
components.  Terry's circuit is fast and accurate and has been used in many 
commercial modules.

2 -- My version's switching time is well under a usec, in agreement with 
what Terry has measured for his own version.  It tracks to <0.1% up to 
30kHz.  Perhaps that isn't good enough for you?

3 -- The 2N4391 is *not* a low cutoff device.  The pinchoff is speced at 4 
to10V, and the units I have are close to 10V.  The comparator output is not 
-12V, it's closer to -11V, so the switch is *not* in deep depletion.

4 -- This is a really old design.  You might want to look at a more recent 
offering (posted on the e-music forum), which uses a MOSFET to switch in ~ 
200 nS.

Thanks for your interest.

Ian