I have a simple non-inverting op amp buffer, to buffer the input to my circuit. It works fine, but I wanted to add a bypass capacitor to the power rails to make sure that it is as low-noise as possible. I referred to this answer and decided to use one rail-to-rail capacitor because it was one less part. But, when I add the capacitor, it gets way more noisy.


simulate this circuit – Schematic created using CircuitLab

When I zoom in on the noise, it actually looks like an oscillation at 35 kHz (see image below). The yellow trace is the output of my whole circuit. The blue trace is where I have re-created just the part shown in the diagram, with a single op-amp and capacitor on a breadboard with a constant voltage input. Both circuits are using the same power supply. When I take out the bypass capacitors, both lines return to being almost totally flat. What is going on here?

enter image description here

  • 3
    \$\begingroup\$ What does the physical layout of this circuit look like? Is the TL072 the opamp? Also, are you using ±12V supplies or one 24V supply? Where are you referencing the oscilloscope ground? \$\endgroup\$
    – Voltage Spike
    Commented Jun 18, 2021 at 16:32
  • 1
    \$\begingroup\$ You've got a 35kHz glitch in something. If this is in the power rail, then you need to correct that. What does it look like when you measure +24V with your scope? \$\endgroup\$
    – TimWescott
    Commented Jun 18, 2021 at 17:08
  • \$\begingroup\$ What's going on with the "clean" trace? What's that ringing? If you have a DC input it should just be a flat line. \$\endgroup\$
    – Shredder
    Commented Jun 18, 2021 at 17:27
  • \$\begingroup\$ Is this your Vac (SMPS) noise coming thru. poor ground ref. \$\endgroup\$ Commented Jun 18, 2021 at 17:33
  • 1
    \$\begingroup\$ Your 'scope probe has a flying GND wire ending in an alligator clip? Try clipping the alligator to the probe tip, forming a loop. Wave the loop above, around your DC supply and see if you see some 35 kHz glitches. \$\endgroup\$
    – glen_geek
    Commented Jun 18, 2021 at 18:28

1 Answer 1


As the answer referenced by Bruce Abbott says, you need to use bypassing to ground, not between the rails. In your circuit, the load current flows through the impedance of your power supplies, providing a feedback path. ( complicated by the details of your input and load ground)


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.