I felt bad because I almost hijacked HomeLate's thread, and that's just not cool. I'm starting a new one as a bit of a repair log.
I have a Macintosh IIfx that I picked up from eBay in a... non-functional state. The batteries were dead but not necessarily pooping everywhere, but the metal shroud and power supply had their share of rust. On first power on, C1 burned up as they seem to do often, and I was happy that I finally got my first repair fire out of the way. Every tantalum on that board was replaced with new, higher rated tantalums, and the electrolytics were replaced with tantalums as well. After a single bodge was made, the IIfx was functional again, minus an issue with C24's negative pad preventing power on from working reliably, solved with a minor bodge to nicer pads.
A month or so ago, I decided to move it from System 7.1 to Mac OS 7.6 to see if that was a worthwhile endeavour. Pro tip: it's not, so after enough messing around with it, I powered off for the evening, and went to bed. It was probably the longest it was powered on since I owned it, maybe 8-9 hours. The next morning, I powered it on to start my move back to 7.5.5, and instead, I got the sad chimes. Brought it over to the bench, and decided I finally had the opportunity to try those serial diagnostics as put together by Rob Braun and enhanced by Adrian Black. I was greeted with `000000000012`, a SCC IOP failure.
After a bunch of messing around with diagnostics and tests, I found that after 15 minutes or so, I could get a normal boot chime followed by a sad Mac chime, and a diagnostic status of `00007FE61092`. At the time, I felt they were related, where time and/or temperature would affect things, however a power off and then power on would bring me back to an immediate sad chimes, so I didn't think it was temperature. The 8530 was obviously fine since the serial port was working. Everything beeped out fine on that corner of the board, just in case the few electrolytics left had pooped all over the board. Oddly -- it turned out to be temperature. The SRAM connected to the SCC IOP was failing, If I pointed a hot air gun at the SRAM, I could get it to chime nicely. I eventually was able to start communicating with the 6502 through the serial diagnostics, and found that cold, it couldn't read back memory properly, but warm, it would at least get all the way. Adding solder to the pads didn't help, so I picked up a new SRAM and installed it, and now I get a chime every time.
... followed by sad chimes, or even worse, absolutely freaking nothing.
So, next steps here are to determine the now unrelated failure elsewhere in the system. I feel like the OSS is involved in some way, but the fact that it's intermittent, where most of the time I get a chime and nothing else, it's harder to determine. Things I know so far? Memory tests all pass. Data bus tests pass. Address bus tests pass. During the long pause, I can hit the interrupt button and usually still get into the serial console with no obvious error to trace.
Next steps?
* Tomorrow, I will play with some freeze spray to see what I can suss out
* Will start beeping out the Nubus side of the equation to see if there's a question by the ROM not being answered
* I have 16mb SIMM PCBs showing up to potentially start building replacements, but that'll be a while
That said, if anyone has seen similar, I'm very open to new ideas.
I have a Macintosh IIfx that I picked up from eBay in a... non-functional state. The batteries were dead but not necessarily pooping everywhere, but the metal shroud and power supply had their share of rust. On first power on, C1 burned up as they seem to do often, and I was happy that I finally got my first repair fire out of the way. Every tantalum on that board was replaced with new, higher rated tantalums, and the electrolytics were replaced with tantalums as well. After a single bodge was made, the IIfx was functional again, minus an issue with C24's negative pad preventing power on from working reliably, solved with a minor bodge to nicer pads.
A month or so ago, I decided to move it from System 7.1 to Mac OS 7.6 to see if that was a worthwhile endeavour. Pro tip: it's not, so after enough messing around with it, I powered off for the evening, and went to bed. It was probably the longest it was powered on since I owned it, maybe 8-9 hours. The next morning, I powered it on to start my move back to 7.5.5, and instead, I got the sad chimes. Brought it over to the bench, and decided I finally had the opportunity to try those serial diagnostics as put together by Rob Braun and enhanced by Adrian Black. I was greeted with `000000000012`, a SCC IOP failure.
After a bunch of messing around with diagnostics and tests, I found that after 15 minutes or so, I could get a normal boot chime followed by a sad Mac chime, and a diagnostic status of `00007FE61092`. At the time, I felt they were related, where time and/or temperature would affect things, however a power off and then power on would bring me back to an immediate sad chimes, so I didn't think it was temperature. The 8530 was obviously fine since the serial port was working. Everything beeped out fine on that corner of the board, just in case the few electrolytics left had pooped all over the board. Oddly -- it turned out to be temperature. The SRAM connected to the SCC IOP was failing, If I pointed a hot air gun at the SRAM, I could get it to chime nicely. I eventually was able to start communicating with the 6502 through the serial diagnostics, and found that cold, it couldn't read back memory properly, but warm, it would at least get all the way. Adding solder to the pads didn't help, so I picked up a new SRAM and installed it, and now I get a chime every time.
... followed by sad chimes, or even worse, absolutely freaking nothing.
So, next steps here are to determine the now unrelated failure elsewhere in the system. I feel like the OSS is involved in some way, but the fact that it's intermittent, where most of the time I get a chime and nothing else, it's harder to determine. Things I know so far? Memory tests all pass. Data bus tests pass. Address bus tests pass. During the long pause, I can hit the interrupt button and usually still get into the serial console with no obvious error to trace.
Next steps?
* Tomorrow, I will play with some freeze spray to see what I can suss out
* Will start beeping out the Nubus side of the equation to see if there's a question by the ROM not being answered
* I have 16mb SIMM PCBs showing up to potentially start building replacements, but that'll be a while
That said, if anyone has seen similar, I'm very open to new ideas.
