Variable ramp again

[Grant Richter]

Hi all,

I got here late, but I was thinking about this approach.
I apologize if this has already been covered.

Lets use the classic integrator (3080), hysteric comparator
circuit. You have two linear current sources. Two solid state
switches switch the two current sources into pin 5 of the 3080.
A flip flop on the square wave output controls the two switches.
You should be able to get the delay in the  whole thing under
500 nanoseconds.

Now you have independent control of the positive and negative going
ramp times. Voltages applied common mode to both current
sources change the frequency. Voltages applied differentially
to the current sources change the waveshape. It seems that
it now becomes a matter of linearity. Correct adjustment
of the gain of the inverter used to generate the differential
signal, should allow the sum of the two signals to remain
very close to 0.

Then it's a matter of the matching of the
current sources. Using a monolithic transistor pair and
a dual monolithic op-amp should make them match
tightly.

An exponential voltage to voltage converter then provides the
frequency control signal.

OK, my $0.02, please feel free to shoot holes in this.