[sdiy] Very fast CCO core, pulse CCO!

[karl dalen]

Hi!

There was a debate earlier with Ian, Rene,Harry and others about
""What-is-a-good-ramp-reset-time? How-fast-have-it-to-be?"" etc.
and how the whole chain have to be fast unless the OP buffer
could down grade the comparators performance. The idea is
(although not new, at least 15 years old) about doing a
pulse/current out CCO?

It excludes the OP buffer from the chain and bases its
oscillation around the a comparator only. Actually it's
perhaps the lowest component count CCO core around!
Even with a OP in buffer configuration..!

Do this around the comparator:
(look at page 6, "squarewave oscillator" in LM393
data sheet for assistance.)

1:Take a comparator e.g. LM393, take two 100k resistors
  connect them serially between +V and AG, connect the
  middle point to the +input.

2:Take a 100k resistor from the output to the +input.

3:Take a 100k resistor from the output to +V.

4:Remowe feedback resistor and insert either a diode or
  a transistor connected as a diode (base+collector)and
  connect the cathode to the -input.

5:Take a 100pF-1nF capacitor and connect between -input
  and AG or -input and +V.

6:Connect your regular NPN expo converter to the -input.

You now have a saw current/pulse voltage CCO core!

You can use a OP amp instead of a comparator just
remove the resistor who goes from the output to +V.

An even better approach is to do a PNP/Fet/NPN switch solution,
of the diode switch. Connect a PNP transistor, emitter to +V,
collector to +input, and base to the comparator output trough
a low value current limiting resistor (1k) to the comparator's
output. ( remember to swap inputs if using PNP tranny).

The capacitor -input junction has sawtoth current flowing,
if this is buffered with a FET transistor or with a CMOS
or FET OP amp there will be a sawtoth voltage at the output.

With OP amp buffer this allows:
1:Polar sawtoth independent of frequency and thresholds.
2:Bipolar sawtoth independent from frequency and thresholds.

A normal Tierry Michaels VCO don't allow these things.
If you change the thresholds in a TM VCO you change 
frequency as well as amplitude.

Idea is:
To get very high oscillations use a fast FET/CMOS comp/OP,
The "imagined" idea is that its easier to get high frequencies
out of this CCO that's buffer less its also neat and simple
for driving logic e.g wave tables!

Quick tests:
I did a test setup and a LM393 with a 1nF cap and NPN
instead of PNP switch, 2,2M in the divider and a 22p
compensation from the divider junction to AG,and 2,2k
feedback from out to +input, and 2,2k from output to +V,
it did easily 0.5hz to 840khz at +5V supply!!....Neat!

Now, one might ask the question, what and where is
the drawbacks of this CCO? Well there must be some
but at least fewer then the typical ASM/TM buffer in
chain designs! One obvious, the NPN switch adds
certain aspects.

Comments wellcome! 

Reg
KD

Höstrusk och grå moln - köp en resa till solen på Yahoo! Resor på adressen http://se.docs.yahoo.com/travel/index.html