I want to make a 100~200 kHz square wave signal. I tried with a Schmitt trigger oscillator . The circuit below worked well with another op amp but the slew rate was too low so a tried with the LM318 which is faster. But I got this on oscilloscope :
Do you know what is happening and how to cure it?
Here is the schematic:
Your op-amp inputs are connected incorrectly for a relaxation oscillator: -
Positive feedback via the 10 kΩ resistors should be to the non-inverting pin like this: -
Image above from here.
You also need op-amp power rail decoupling (a capacitor) and, the capacitor you mention on the output is a no-no.
You need a capacitor from pin 4 to pin 7 near the chip and no capacitor on the output (also inverting and non-inverting are switched, but that was a transcription error you said).
In fact, put a 100Ω resistor in series with the output (between that and your probe).
The decoupling capacitor prevents inductance in your power supply leads from resulting in imperfect power to the chip when there are fast changes in demand from the chip. One way you can cause such a rapid demand is to add a capacitor to the output. If it's set up as an amplifier, that will likely cause oscillation unless the capacitor is tiny (100pF is usually safe, but not necessarily for micropower op-amps). As a comparator it will cause big surges from the supplies.
A small series resistor between the output and whatever it might be driving (but not in series with the feedback) that is capacitive (such as a x1 oscilloscope probe or a cable with significant capacitance) can help prevent that kind of problem.
