cheesestraws
Well-known member
Following my normal habit of posting non-obvious fixes for other people to find:
Symptom video:
(thanks, @Paulie)
Macintosh Plus. Definitely an analogue board problem; swapping AB made it go away. AB already recapped. Capacitors of decent quality and known to work in other ABs.
Diagnostics:
Solution
This turned out to be a corroded solder joint or joints in the bottom centimetre of the AB. Several were in a worse condition than I strictly liked the look of, so I cannot pinpoint precisely which one it was, but they were all in the switching core of the PSU (this leads me to wonder whether the 10 Hz was an artefact of the feedback controller and the switching core fighting with each other; it's an odd frequency to have induced otherwise). One or two had had corrosion creep down the component leads into the joint, and these are my best guess as to the immediate cause of the wibble. Removing the corruption and old solder from these and reflowing fresh solder onto a cleaned pad and lead has led to a wibble-free afternoon, and will hopefully be a permanent fix.
Symptom video:
(thanks, @Paulie)
Macintosh Plus. Definitely an analogue board problem; swapping AB made it go away. AB already recapped. Capacitors of decent quality and known to work in other ABs.
Diagnostics:
- Examination of video showed that screen wibble had a frequency of 10 Hz; that is to say, each line of the screen returned to its original position after 100ms.
- Amplitude of wibble reduced, but frequency did not change, as computer warmed up. Eventually disappeared entirely.
- Examining HSync on an oscilloscope showed a solid and unchanging HSync frequency.
- Examining 5 V and 12 V lines out of the floppy port showed a pronounced 10Hz ripple which faded as the wibble subsided.
- Upon deciding it was the PSU, the computer obligingly rebooted randomly a couple of times to prove the point.
Solution
This turned out to be a corroded solder joint or joints in the bottom centimetre of the AB. Several were in a worse condition than I strictly liked the look of, so I cannot pinpoint precisely which one it was, but they were all in the switching core of the PSU (this leads me to wonder whether the 10 Hz was an artefact of the feedback controller and the switching core fighting with each other; it's an odd frequency to have induced otherwise). One or two had had corrosion creep down the component leads into the joint, and these are my best guess as to the immediate cause of the wibble. Removing the corruption and old solder from these and reflowing fresh solder onto a cleaned pad and lead has led to a wibble-free afternoon, and will hopefully be a permanent fix.
