Put 100mV @ 400Hz to 1000Hz into the input. Amp controls set as you would when playing through it and the problem occurs.
You'll be probing along the signal chain starting from the input and moving toward the output.
Setting the scope is another matter and depends on where in the signal chain you're probing. I'd suggest a 10:1 probe and just set the division so you can see the wave clearly enough. The right division setting for each node will be different. Perhaps after the distortion is identified a more finite setting for accurate measurements will be recommended. But for now you just need to "see" the wave form.