I'm having difficulties keeping this opamp from oscillating and heating up.The schematic is a slightly modified Mini3(mini cube )amp schematic with active ground.

A 1nF cap across the 330 ohm feedback resistor was added but can't measure any improvement also the datasheet mentions cautions about capacitance in the feedback path. It also says using a 400ohm resistor across inputs .But how?. Would simply sticking it between the inverting and non-inverting really solve something? It appears to be noise mixed in also. The Mhz bandwidth may be too high, but wanted to try with the original schematic.

Also does anybody have a schematic for offset voltage compensation(beside the resistor network that injects current)? Schematic Gerber Files PCB

  • 5
    \$\begingroup\$ Please post a copy of the schematic, or use the tool to draw while editing the question \$\endgroup\$
    – Voltage Spike
    Mar 7, 2019 at 18:02
  • \$\begingroup\$ In addition to the component schematic, please include a drawing of the circuit you are currently troubleshooting. \$\endgroup\$ Mar 7, 2019 at 18:09
  • 1
    \$\begingroup\$ This is a 500-MHz amplifier. You should also post your PCB layout if you want an advice how to fix your problem. \$\endgroup\$ Mar 7, 2019 at 18:10
  • \$\begingroup\$ That 1nf cap is suspicious. A few pf---ok. Is all this on a 2+ layer pcb? or some whiteboard? \$\endgroup\$ Mar 8, 2019 at 3:06

1 Answer 1


The original schematics calls for LMH6643 OPA, which has under 10 MHz bandwidth with gain +5. Don't use OPA690 without proper power de-coupling and power grounds, like shown here,

In your layout the ground path for bypass cap doesn't practically exist,

enter image description here

I think this layout won't work for any OPA. You need to change it.

  • \$\begingroup\$ I found another interesting thing,when i increased the gain to 2,by adding a 330 ohm resistor from the inverting terminal to the virtual ground all oscillations stopped and the OPA690 didn't get hot.The measured output (referenced to the virtual ground from the TLE) was only 5mV above which is ok due to the offset.It may be that the op-amp is not unity gain stable? \$\endgroup\$
    – Andrei112
    Mar 9, 2019 at 9:11
  • \$\begingroup\$ Also as i decreased the gain closer to 1 ,by using a 1.8k resistor instead of a 330 ohm ,the IC dissipated more heat ,but not as much as without any feedback.If i keep the Gain=2 ,but bypass the the output -to-inverting input resistor with a 1nF cap it enters oscillations again like nothing's changed.I belive it also doesn't like unity gain too much? \$\endgroup\$
    – Andrei112
    Mar 9, 2019 at 9:14
  • \$\begingroup\$ @Andrei112 Please take time to examine Figure 1 in OPA690 datasheet. As usual, the unity gain has widest bandwidth and amplification peak at high frequencies, which is very bad for stability. \$\endgroup\$ Mar 10, 2019 at 1:18
  • \$\begingroup\$ you are right,i initially implemented something like that ,but on wrong layout as you pointed out so i guess by increasing gain i narrowed the bandwidth and got out the problem somehow.Thank you for the help.I'm redesigning the pcb with a Gain =2 or more and better grounding.Also using the TLE2426 is better than the resistor divider(in terms of noise) method if i properly add bypass caps for the resistors? \$\endgroup\$
    – Andrei112
    Mar 10, 2019 at 18:45

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.