Archive of the former Yahoo!Groups mailing list: MOTM

I'd thought I'd rephrase my questions, and leave out the silly _2001: A
Space Odyssey_ jokes, in hope of getting a better response.

I would like to build an edge detecting circuit in order to derive a
trigger signal from the voltage changes of a stepped waveform. I plan on
using it with the CV output of a UEG in LOOP ONLY mode with the SLOPE
switch set to STEP. So it will generate a signal when the voltage changes
in either direction.

I'm not an EE, but I have a copy of Horowitz & Hill. Based on that text, I
have pieced together the following ideas for a circuit:

Take the CV output of the UEG, run it through a voltage follower (as not to
draw too much current), to a differentiator. Figure 4.51 on page 224 looks
like a cap (C) going into the inverting input of an op-amp, with a resistor
(R) going from the op-amp's output to the inverting input, and the
non-inverting input going to ground, such that:

Vout = -RC dVin/dt

Then it says this circuit generally has "problems with noise and
instabilities at high frequencies because of the op-amp's high gain and,
internal phase shifts". It shows a new circuit, with high-frequency
rolloff, in Figure 4.52. Yet it doesn't give a formula in which to
calculate the component values. Does anyone know how to get this circuit
in Figure 4.52 so that it will work?? Any other ideas??

Even if I can get the differentiator to work, I would end up with both
negative and positive voltages. So after this differentiatior, I would
need an "absolute value circuit" or "active full wave rectifier", I found
two on page 222.

After that, I need to get this positive voltage up to approximately 5
volts, with enough current available to be used as a trigger. Is there a
comparator circuit that anyone would like to recommend?? I'm thinking a
LM339 with a pullup resistor connected to 5 volts would work.

Do they make comparators that will give a positive output regardless of the
polarity of the input voltage?? The table in Horowitz & Hill (pp. 584-585)
leads me to believe that an LM311 can, which would eliminate the need for
an absolute value circuit. Any ideas??