Well -while
@loz_the_guru can report his own thoughts - I'm glad to say we met earlier today, and spent more time chatting than really worrying too much about the amp!
The punchline: yes, the ticking reported is thermal in origin, and the three 240K resistors
@a.palfreyman first suggested as running warm. They are.
Basically I sent some minutes going over it first with a bright light, and the boards themselves are clean- put together quite neatly enough
and free of flux and obvious things that
might cause a problem (dots of solder, debris, dust, not a hint of 'tracking' etc.) All the HV caps etc are reputable quality and appear suitably rated based on build notes in threads like those linked above by others.
Power -up, quick nominal voltage check: all spot-on, rails bang-on 480v where noted, so initial thoughts are positive; power down. The supply HV rails bleeds-down close to 0v very fast. That's nice behaviour to know...
Re-apply mains power and let it idle. I went over all the heatsinks, discrete parts and all resistors with my Fluke IR gun. Used close-up this has a fine enough field of view to clearly register the centre of a non-SMT transistor or a 2w resistor on a packed board. No immediate surprises, so left it on. The amp is totally silent mechanically, as is the transformer (which settled 4degC above ambient). Idling with the top off, the heatsinks on the power amp boards all settled quickly to modest and sufficiently- equal values - 38-42degC typ. All insulators on transistors in place, neatly assembled, clearances ok - no issues. Ditto on the supply board, including the 15ohm resistors which run significant current.
But those three 240K resistors come immediately up to 65-68degc and after a bit of a soak, settle near 90degC - yes, with the lid off. That's rather too much for long-term reliability imo. In chatting about this, Loz commented the reported ticking had first been noticed when he'd had his dac sat on top, but didn't think it had happened since. Aha.
Looking at the 240K parts more closely then closely - they are mounted right down tight to the pcb, and the leads are cut short and soldered tight -no pre-bend in the lead-outs , no possibility to expand at a different rate to the pcb, and indeed, while monitoring voltage after turn-off - I heard a very slight tick from the central one.
tl;dr: Proposed remedy:
The 240K resistors are dissipating up to 1w each, installed on a pcb footprint for 22.5mm lead spacing.
I'm inclined to replace the three 240K parts with 5w metal-oxide resistors, spaced-off the pcb by 15-18mm. Plenty of space for this...and metal oxide is very, very robust against overload. A quick look at suppliers suggests the 5w size is available readily with a 700v+rating. Standing-off with a careful pre- bend in the legs will make space for the longer 5w resistor body, and allow
much better convection around (and flex with any thermal expansion) without getting close to adjacent components.
Only two other things I've suggested we address:
1. yes, that indicator LED appears to be connected to the transformer AC
input! That has to go; replace with standard 5mm part fed from tapping a 15v rail from the opamp servos would be ideal, and easy to do. We've left is disconnected for now.
2. Also - there's no secondary insulation over wired connections to the IEC input socket and voltage selector switch. Again - easy to improve-upon. The case
is properly bonded to mains earth, btw.
I suspect, hope, this could lead to another meet, with more music, one evening TBA. Very nice to meet you, Loz! : )
