[sdiy] They aren't sawtooths, they're ramps

[David G. Dixon]

> At 11:29 PM 11/4/2009, David G. Dixon wrote:
> >This reminds me, I have now grafted the 2164 Expo onto the Thomas Henry
> >triangle-core VCO-1, and laid out a very tight PCB (2.1" x 5.5") with
> very
> >low parts count which also includes Ray Wilson's sine shaper (to use up
> the
> >extra 13700) and a variable sync.  Simulations suggest that the tracking
> >should be almost perfect, as it should be much easier to obtain excellent
> >tracking with the VCO-1 then with JFET-switched saw-cores.
> 
> How fast is that switching in your simulation?  My best Tri's and Saw's
> aren't all that different regarding tracking. They both switch in ~200 ns.


According to the simulation, switching from the prevailing level (+/-5V) to
ground takes about 50ns.  Then, it takes about 200ns to settle to the low
trigger level, and about 800ns to settle to the high trigger level.
However, the triangle has already reversed direction once the main part of
the switching transient occurs.  Therefore, I'd say that the switching takes
less than 100ns.  This is based on a discrete schmitt trigger made from two
2N3904s, with a 47pF speedup cap (see Scott Stites's website for the VCO-1
schematic).

Also according to the simulation, this circuit will need only a very small
amount of HF trim to track well up to 16kHz (and presumably beyond), with
TL07x everywhere.