Thanks as usual for the input/help with my golden age saga.
Having now had a clear moment to investigate this behaviour with more tests it is evidently not a component fault in any of the main audio circuitry. The problem is in the PSU even though the spike signal was not detected on the supply rails in the audio board.
I picked up the same spike/rings at the line out socket, across the balanced output pins (xlr 2 & 3), with all the audio circuitry disconnected. That means with no O/P xformer connected to the O/P socket, input and output sockets just run to connectors to the audio board and these are all disconnected. The 24V and 48V power connectors to the audio board are also disconnected.
The PSU circuit is always on when the plug from the AC pwr xformer is in, regardless of whether or not the front panel on/off switch is pressed (crap + I noticed that the phantom switch applies 48V even if the power switch and indicator light are off!). The input and output sockets are all mounted on the same physical board assembly as the PSU. The spikes only occur when AC power is connected to the unit (not entirely surprising, but...)
When isolated like this the only (non-parasitic) connection from the PSU to the output socket balanced output pins 2&3 is via the circuit ground, which only connects to the hot and cold pins via two small nominally equal value decoupling caps (as well as directly to shield pin1). So some imbalance between the impedance to ground capacity as seen by the scope versus that felt by the circuit must be at work here for anything from the PSU to appear on the scope.
I found I could pick up the spike/ring signal just holding a scope probe against the power cable before it enters the unit. After expending some braincells suspecting spikey interference coming in along the cable I started doing some more capacitive probing around the PSU circuit components. The signal increases to a maximum around one particular capacitor located in the early stages of the PSU circuit. (I haven't removed the board to trace out the circuit yet.)
I now measure the pulses as occurring at bang on 10ms apart (not 5ms, sorry, I must have misread the scope previously) which corresponds with the zero crossing frequency of the 50Hz power input.
So, diodes opening/clanging shut into a failing reservoir capacitor? There is an inductor in the circuit somewhere too. Interesting that the phenomenon is seen back down the cable between the rectifier diodes and the power transformer. Is it actually some ringing on the power transformer side?
I also noticed the magnitude of the spike/ring gets a fair bit larger when the PSU is under load supplying the audio board. Larger amounts of charge rushing into the reservoir.
The cap is currently a prime suspect, but I'm also wondering if the design of the board is maybe not so good regarding the paths of ground currents, which will take a bit more pencil-chewing. Maybe this behaviour was there all along - this is the first time I've had a scope on it.
Anyway I'll get the psu board out, trace out the circuit and replace the suspect cap for a start.