Osc reset time

[Fraser, Colin J]

I knocked up a quick prototype oscillator last night on stripboard using the
cap 'flipping' idea.

I used a 4066 switch, half a 4013 flip-flop and half a LM319 comparator.
I used a 5 volt reference for the 'top' of the cap, and set the comparator
to trigger the flip-flop to swap the ends of the cap between the current
source and the 5v ref when the voltage at the current source end of the cap
reached 2.5v.

End result was as expected - a falling sawtooth between 7.5 and 2.5 volts. I
didn't add a diff amp around the cap for tri output, but I put my scope
probes across the cap and there it was...

Adding up worst case response times for the comparator, flip-flop and
switch, I estimate the 'reset' time in this configuration to be about 300ns.
Oscillator sync isn't so straightforward as a normal ramp osc. I could use a
sync pulse to clock the other half of the 4013 to open a fet which would
drain the cap quickly at the current source end, and use the comparator
output to reset the flip-flop and turn off this fet when the limit is
reached - need to test this but in theory it should be fine.
Also, I can add another 'sync' input that triggers the cap swap flip-flop
directly. This will invert the position of the sawtooth wave around it's
midpoint, and also reverse the direction of the tri wave (I think...).
Need to test what this sounds like to see if it's worthwhile.
I still haven't figured out how I can add precise linear detune...

I'll add a proper expo converter with 1v/oct cv input at the weekend and see
how it performs.


Colin f