There are a pair of phonos on the chassis admitting L and R audio from the preamp.
The ground of these phonos is firmly grounded to the chassis (metal bodies phonos, metal chassis).
A ground tag is on each phono, the tags linked by a wire.
Each phono has a signal wire plus a ground wire running to the PCB some 200mm/8" away. These form a tightly twisted pair, with the signal wire running to the signal in point. The ground wire is however unterminated, aside some insulating heatshrink. FWIW, the signal in points are close together for the two channels.
A fifth wire runs from the middle of the wire joining the ground tags to the ground connection on the PCB between the two signal inputs.
The ground plane of the amplifier has a 5Ω 100W resistor to chassis.
I can understand (a) lifting audio from chassis grounds, (b) twisting pairs, (c) avoiding hum loops with single shield etc, but what is completely confusing me is what's the point of all this complexity given the phono socket grounds are bolted to the chassis at the very start of this?
Chassis often don't make very reliable conductors and can have voltages/currents induced by magnetic fields i.e. the transformer. A wire is a better way to connect grounds.
Let me be more specific:
1. If the designer chose to lift the audio ground from chassis ground via the 100Ω resistor, why are the phonos bolted to the chassis?
2. Because they're bolted to the chassis the single wire that goes from the phono tags to the audio ground on the PCB now shorts out the 'lift' completely - why?
3. Why run a twisted pair for each channel to the PCB, but not connect the ground? Yes to avoid loops if BOTH grounds are connected hence the single ground wire (forgetting the issue of 2 for a moment) - but why run the two unterminated ones?
I really can't speak for another designer's choices, I can't even say that they are good but often it is a case of if it works, it works, unless you have some underlying noise or hum issue.