Documenting this very obscure issue - and the solution too - in case anyone else ever runs into a similar problem in the future:
Earlier this year I did a full "restoration" on my Twentieth Anniversary Macintosh - I say "restoration" because the thing worked fine beforehand, but there was just a lot of jobs that needed doing regardless. Part of the project was doing a full recap of all boards, including the PSU (thanks Amiga of Rochester - link to Thomas' post on this project), and also implementing the fix for the "Bose buzz" issue. I've been very pleased with the outcome of the project, but ever since then I had occasionally encountered a bad clicking/tapping sound in the base unit that I could tell was being caused by something rubbing against the fan blades in the PSU. I had also replaced this fan during the project since the original one was quite loud, and I subsequently realized that I'd left the fan wiring harness too long when I replaced/spliced in the new fan - i.e., left too much wiring loose inside the PSU and it must have been dangling into the fan sometimes. Usually some "percussive maintenance" would resolve the issue at least temporarily, but I really wanted to solve this problem once and for all since the blades would eventually damage the wires.
So, I recently got around to tearing down the entire base unit (again), being extremely careful not to worsen the condition of the incredibly brittle plastics (again), and slowly and tediously peeling back all of the adhesive-backed acoustic foam on the PSU (again). Got the PSU apart, chopped out some of the excess length on the fan cable that I had left before, re-soldered and re-heat-shrinked the wiring, and re-assembled the PSU back into the bottom case of the base unit.
Powered the unit on with fingers crossed (also in part hoping I hadn't broken anything this time around) ... and almost immediately, right after the startup chime, heard the clicking sound once again
Turns out the issue wasn't my wiring job at all, it was just this small green cylinder (I presume some kind of ferrite bead / EMI filter or something?) that resides on the wires going from the board to the transformer. When I had the PSU apart the first time, it must have allowed this part to slide down the wiring and get then too close to the fan blades after reassembly.
From outside of the PSU case (using a screwdriver to reach through the fan blades), I was able to just push it further down the wire and behind one of the structural braces/arms on the rear of the fan where it can't slide down the wire anymore. Didn't even have to take the PSU apart again.
So, that is problem solved. Quite a silly and minor issue to have, but thankful that I at least didn't break anything. And it still feels good to have the fan wiring cleaned up regardless. Let this be a lesson to future tinkerers to always check the obvious things first ... lol.
Photo of the green part (ferrite bead / EMI filter?) that can slide down the transformer wires (red/black/blue), since it is just loose and not fixed in place in any way, and come into contact with the fan blades. It should be pushed either to the left or right, so that it is *behind* the structural brace/arm of the fan (partially visible to its right, behind the fan blade). It can be seen at top-left in the second photo, which was taken earlier this year when I had the board out.
Some fun facts / interesting observations about the TAM PSU (which is a modified version of that for the PM 6500): it has one of those sockets for connecting a monitor power cable, but the socket is blocked and not visible behind plastic when installed in the base unit case. Not sure if it actually works.
There is at least one wire that comes from the umbilical cable but does not actually connect anywhere on the board (seen here in pink) - I have read that the umbilical cable is actually much thicker than it needed to be because they had made provisions for a more powerful CPU than what actually ended up shipping; not sure if this is related to that or not. The entire thing was very clearly cobbled together based on how messy the wiring connections are between the board and the umbilical cable, and also how much of the board is unpopulated.
Also, there is a two-pin header (visible at bottom left) on the Bose amplifier board that is labeled "TEST", which lines up with this square hole molded into the lower base unit case next to the subwoofer volume knob. But, this hole is blocked off by the label on the underside. I assume this was for some kind of internal diagnostics or development purposes as it was never used in production.
Earlier this year I did a full "restoration" on my Twentieth Anniversary Macintosh - I say "restoration" because the thing worked fine beforehand, but there was just a lot of jobs that needed doing regardless. Part of the project was doing a full recap of all boards, including the PSU (thanks Amiga of Rochester - link to Thomas' post on this project), and also implementing the fix for the "Bose buzz" issue. I've been very pleased with the outcome of the project, but ever since then I had occasionally encountered a bad clicking/tapping sound in the base unit that I could tell was being caused by something rubbing against the fan blades in the PSU. I had also replaced this fan during the project since the original one was quite loud, and I subsequently realized that I'd left the fan wiring harness too long when I replaced/spliced in the new fan - i.e., left too much wiring loose inside the PSU and it must have been dangling into the fan sometimes. Usually some "percussive maintenance" would resolve the issue at least temporarily, but I really wanted to solve this problem once and for all since the blades would eventually damage the wires.
So, I recently got around to tearing down the entire base unit (again), being extremely careful not to worsen the condition of the incredibly brittle plastics (again), and slowly and tediously peeling back all of the adhesive-backed acoustic foam on the PSU (again). Got the PSU apart, chopped out some of the excess length on the fan cable that I had left before, re-soldered and re-heat-shrinked the wiring, and re-assembled the PSU back into the bottom case of the base unit.
Powered the unit on with fingers crossed (also in part hoping I hadn't broken anything this time around) ... and almost immediately, right after the startup chime, heard the clicking sound once again
Turns out the issue wasn't my wiring job at all, it was just this small green cylinder (I presume some kind of ferrite bead / EMI filter or something?) that resides on the wires going from the board to the transformer. When I had the PSU apart the first time, it must have allowed this part to slide down the wiring and get then too close to the fan blades after reassembly.
From outside of the PSU case (using a screwdriver to reach through the fan blades), I was able to just push it further down the wire and behind one of the structural braces/arms on the rear of the fan where it can't slide down the wire anymore. Didn't even have to take the PSU apart again.
So, that is problem solved. Quite a silly and minor issue to have, but thankful that I at least didn't break anything. And it still feels good to have the fan wiring cleaned up regardless. Let this be a lesson to future tinkerers to always check the obvious things first ... lol.
Photo of the green part (ferrite bead / EMI filter?) that can slide down the transformer wires (red/black/blue), since it is just loose and not fixed in place in any way, and come into contact with the fan blades. It should be pushed either to the left or right, so that it is *behind* the structural brace/arm of the fan (partially visible to its right, behind the fan blade). It can be seen at top-left in the second photo, which was taken earlier this year when I had the board out.
Some fun facts / interesting observations about the TAM PSU (which is a modified version of that for the PM 6500): it has one of those sockets for connecting a monitor power cable, but the socket is blocked and not visible behind plastic when installed in the base unit case. Not sure if it actually works.
There is at least one wire that comes from the umbilical cable but does not actually connect anywhere on the board (seen here in pink) - I have read that the umbilical cable is actually much thicker than it needed to be because they had made provisions for a more powerful CPU than what actually ended up shipping; not sure if this is related to that or not. The entire thing was very clearly cobbled together based on how messy the wiring connections are between the board and the umbilical cable, and also how much of the board is unpopulated.
Also, there is a two-pin header (visible at bottom left) on the Bose amplifier board that is labeled "TEST", which lines up with this square hole molded into the lower base unit case next to the subwoofer volume knob. But, this hole is blocked off by the label on the underside. I assume this was for some kind of internal diagnostics or development purposes as it was never used in production.
