How would a scope help here? Remember I'm nearly tabula rasa when it comes to troubleshooting like this.
With a square wave source, an oscilloscope will show the transient response which should be clean. The transient response will also reveal the low and high pass frequency response. A sine or triangle wave source will reveal high levels of zero crossing distortion if present; this could be the result of an improper idle current setting. A triangle wave source will reveal high levels of non-linearity if present. If there is any spurious oscillation, then it will be visible on an oscilloscope.
Tests should be done with different realistic loads.
If your oscilloscope can perform an FFT, then the noise spectrum could be measured which might indicate something. In the old days one might sweep a bandpass filter to make spot noise measurements.
However these STK based amplifiers aren't famous for their low noise...
The hybrid's bipolar input stage should pretty quiet operating at a collector current of 2 milliamps but the input resistance at the non-inverting input and probably the inverting input as well is too high for good noise performance. The application example for the STK3102 from Sanyo shows a quieter design which assumes low impedance drive so I get the feeling that the Sony designers did not quite know what they were doing.
Counterintuitively, lowest noise will be with the attenuator set to either minimum or maximum so you might try that. With the input shorted, see how the noise changes with attenuator setting.
I do have a 2nd one of these that is much quieter, no audible hiss unless you're right near the speakers, and as I said the noise is not quite normal - it's choppier. It also has random spikes every once in a while. I've tried to record a sample of it below, the effect is far more pronounced on my floor standing speakers but it would take a while to move my recording equipment. Please forgive the ambient street and fan noise:
It sounds like it could be aluminum electrolytic capacitor noise to me.
I've managed to test a few of the electrolytic capacitors. They generally measure within spec on my Fluke. However the ESR meter shows what I believe to be too high a value on every one I've checked - at least going by the chart in the instructions I have. This is my first practical attempt at using the ESR meter so I could be completely wrong, and yes I did zero out the probes.
I am not a fan of ESR measurements because they are so often equivocal. Dissipation measurements are much more useful because they are consistent across a range of capacitance values.
I'll order a new set of capacitors and see how it goes.
Aluminum electrolytic capacitors are cheap, have a limited life, and are easy to change so it makes sense to replace them early when the cause of a problem is not readily apparent.