I'm having trouble building a circuit to generate square, triangular, and sine waves. In theory (calculations), everything is correct, and in the simulation (done in LTSpice) everything is also correct. However, when I finish the assembly, the circuit doesn't oscillate (the wave generation doesn't work for square and triangular waves). I'm using the values according to the image and two UA741 Op-Amps. Interestingly, by mere coincidence, I discovered that using the UA741CN K9E218 in the square wave generator circuit, and UA741CN K9B01W, the circuit works as expected. Only if I use these two specific Op-Amps, in that order. But I'm not sure what that code means. Do I need to add something? A buffer, perhaps? Could you help me make it work for any UA741? I imagine it could be something related to their configuration or even imperfections, etc.

enter image description here

  • \$\begingroup\$ You've probably wired it incorrectly. You will need to troubleshoot it. Measure important voltages around the circuit and report back. \$\endgroup\$ Jun 22, 2023 at 1:30
  • \$\begingroup\$ @KevinWhite So, I've assembled it around 10 times, and every time I've used those two UA741CN K9E218 and UA741CN K9B01W Op-Amps, it works perfectly. However, if I use two UA741CN K9E218 (in the same circuit/components), it doesn't work. I've measured all the voltages, and the output is always saturated, at approximately +Vcc or -Vcc. \$\endgroup\$ Jun 22, 2023 at 1:39
  • 1
    \$\begingroup\$ With the values of R1 and R2 you have selected the saturation voltage of U1 needs to be larger than U2 for it to oscillate. You could try reducing R1 somewhat, say 4.7k or 6.8k. The circuit should be guaranteed to oscillate then. \$\endgroup\$ Jun 22, 2023 at 1:51
  • 1
    \$\begingroup\$ @benjamin_ee Looking over the LTspice group, that square wave and ramp generator is credited to an idea from G. Salmonson. And as Kevin suggests, that earlier design uses \$R_1\lt R_2\$ by a smidge. (It also runs on a single 9 V battery and uses a divider to make the midpoint reference.) \$\endgroup\$ Jun 22, 2023 at 2:03
  • 1
    \$\begingroup\$ Please stop using 741s, they were obsolete in the 80s. Why does everyone insist on using them constantly? \$\endgroup\$
    – Hearth
    Jun 22, 2023 at 4:29

1 Answer 1


Technically, your circuit is linear. If U2 were a proper comparator, with slight input hysteresis, then this design would work well, but it's a regular op-amp.

This whole system has sufficient phase margin to turn what needs to be positive feedback into negative, at the frequency you want this to oscillate. Consequently, it can be quite stable, which is the last thing you want in an oscillator.

Try reducing U2's gain to 100, with R4 and R5 here:


simulate this circuit – Schematic created using CircuitLab

This has the effect of keeping loop phase near 0° (for positive feedback) over a much broader range of frequencies, including the expected frequency of oscillation.

To be honest, even though my frequency analysis revealed this phase issue, it's certainly questionable for huge signals, and when the op-amp is supposed to rail, as in this application. Still, I tried it, and it worked, so I suggest you try it too.

Alternatively, use a proper comparator for U2, like the LM393, or even better a push-pull output type, like the TLV1812.


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.