I am currently building a VCO for an analog synthesizer. The circuit produces a triangle wave by taking a Schmitt trigger output and feeding it into an integrator. The constant output of the Schmitt trigger is integrated to a triangle wave. When the triangle output hits the upper threshold, the Schmitt trigger flips and the output goes negative. However, the lower output is not square. It moves linearly with the triangle input. The entire circuit schematic can be found here,

and the entire article that discusses the circuit in question is linked here.

I have a simulation of the circuit running on falstad.com/circuit/, linked here.

My question is: Why does the upper trigger work so well, i.e. why is the upper threshold behavior of the circuit constant while the lower threshold behavior is variable (and seems to act like an inverting amplifier with slight nonlinearity?

  • \$\begingroup\$ Adding a proper schematic inside the question helps. Hint: have a look at the base of the left transistor and try to figure out why it does what it does. \$\endgroup\$
    – PlasmaHH
    Jan 13, 2017 at 21:08
  • \$\begingroup\$ @HunterAkins fascinating example of a faulty design. The large input swing need to the positive feedback switch starves the input stage of current. far better ways to do this \$\endgroup\$ Jan 13, 2017 at 21:56

1 Answer 1


try this if you insist on using 2 BJT's

What output swing do you want +15, 0 like above of something else?

This works far better enter image description here

  • \$\begingroup\$ The output swing would ideally be a square wave, with an amplitude of 5 Volts peak to peak, centered at 0 Volts. I think an op amp Schmitt trigger is clearly the easier circuit, although I believe the author of the circuit we are trying to replicate chose BJTs for speed. I also feel like there is a way to use some feedback to compensate for the drooping on the negative swings. \$\endgroup\$ Jan 15, 2017 at 20:58
  • \$\begingroup\$ Yes I know Hunter but this design is incapable of doing this with 2 transistors and an inbalanced collector impedance output using common base positive feedback. It's a bad design. If the person who indicated -1 wants to contribute something, correct my error otherwise I assume they know nothing about what I am saying.. and are equivalent to trolls who by definition contribute nothing useful. \$\endgroup\$ Jan 15, 2017 at 23:29
  • \$\begingroup\$ I'll send $50 to anyone who gets a clean +/-5V square wave with 10% hysteresis from the Op's question copied from a poor design by tuning the values. \$\endgroup\$ Jan 15, 2017 at 23:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.