I thought I had solved this erratic misbehaviour back eleven days ago, but I hadn't.
Now I have definitively solved it after a few forays with the jewellers loup and my soldering iron. Just to refresh the memory - the issue was an erratic buzzing and strange horrible distortion of the sidetone. It was extremely unpredictable and I thought I had beaten it at least twice. The tone and amplitude of the buzzing seemed to be related to multiple presses of the edit keys, and it would change as I entered different stages of setting the time. The second hour digit buzz would get louder or sometimes softer when I moved onto the minutes one I was clearly hearing some artifact of the logic chips interactions, but could not understand why.
Yesterday I discovered that I could make the buzz and strange tones come on and off by tapping parts of the main board. I had previously realised that knocking on the case firmly with my knuckle could make it come and go, so it was clearly a bad connection somewhere. I had thought the pins which mount the board to the front panel were the issue because once it went away when I wiggled it. Likewise a couple of tiny solder flakes I found and removed.
Tapping the board and ICs more and more lightly with a plastic probe, pointed the way to the centre and rear part of the board so I set too re-flowing these connections. All of the through the hole joints have a tiny bit of the component leg - and then I found one which didn't. I floated the iron over the? negative pin of C24 and there was no wire there at all, just a grey soldered pad. I inspected C24 - I mean, I wobbled it and one end was moving. It seems that the builder may have cut the wires before they put the electrolytic capacitor into the holes, cut one leg a little too short, and soldered the pad but not the wire which did not have the length to get right through the plated through hole. I must have looked over this board with a 5 times mag loup, at least a half a dozen times and missed this. The radio would have passed inspection because C24 negative lead was mostly touching the plating on the plated through hole - touching until it wasn't ,because the board had warmed up, and and then the unsatisfactory behaviour would start and the just go away again randomly, as contact was restored for whatever reason. Oddly, I could always make it go away by switching off and on again. I don't know why that worked but it did systematically solve it for a while.
The solution was to push the capacitor's loose side down to the board and flow more solder into the plated through hole and get enough heat in there to firmly connect the leg. Now the radio is perfect and I feel pleased that I was finally able to find and solve the problem.
I'm really enjoying the radio in spite of this for a while rather puzzling issue.
