Mac IIfx Refurbishing Phase II - PSU? Log of work, first repair on it in 7 years

[IIfx]

Been a while since I did any Mac stuff. Wanted to get my favorite, the IIfx, running again. It's been shelved for about a year now due to seemingly power-related stability issues. When fully loaded with it's typical configuration (128mb RAM, multiple Nubus cards, 1 Radius Rocket 40mhz, IBM 2GB SCSI HDD from 96) it would boot-loop just like it did before I did the full logic board recap 7 years ago. New axial electrolytic (Nichicon) and surface mount polymer (Panasonic) caps were installed. Solved the boot looping back then. I immediately assumed that it was PSU caps. This may be a partial truth.

I was proceeding with the PSU recap, but stopped when noting that the majority of the original Rubycon caps Sony used were still perfect when tested with a cap tester. Only a 1uf one was notably out of spec. Reassembled the PSU with new caps replacing the 11 I pulled. Boom! Seemed like the IIfx was stable again, but it really wasn't. Lots of strange crashes.

So I have started diving into what could be the problem. Voltages out of the PSU are perfect and there is no heat building up at the motherboard power connector. It seems Bank B of memory is the problem - it gets hot and causes the system to be very unstable. Pulling Bank B gets the system to pass MacTest Pro's memory check.

This RAM was perfect in the past - but it was in Bank A. Bank A is currently occupied by new SIMMs made using the Doug Brown design I bought a few years ago.

System Temps:

 

[mdeverhart]

How does the fan on the power supply seem? Are you running your tests with the lid on or off? I’m not sure which direction the air flows on the IIfx, but you may not be getting airflow over the RAM with the cover off, or with a bunch of other cards installed. You could try balancing a small fan on top of the ram blowing down to see if that helps - it’s not a permanent solution, but it would be useful to know if it is a heat problem.

 

[IIfx]

How does the fan on the power supply seem? Are you running your tests with the lid on or off? I’m not sure which direction the air flows on the IIfx, but you may not be getting airflow over the RAM with the cover off, or with a bunch of other cards installed. You could try balancing a small fan on top of the ram blowing down to see if that helps - it’s not a permanent solution, but it would be useful to know if it is a heat problem.

Lid off, office is cold though. Fan is fine.
Airflow is right to left in a IIfx, IIx, and II.

I figured out the likely problem. Buffer chips in the path between the memory controller and memory. The ones connected to Bank B are likely marginal. The higher electrical load of the tall RAM simms is putting enough strain on them to cause instability. You put the tall RAM in Bank A, and the system is 100% stable. The more modern, single row SIMMs work well in Bank B. The system now passes many runs of MacTestPro.

My solution? Do nothing. I put heatsinks on all of the warm or hot running chips in the IIfx to improve longevity.

The power supply recap resolved the problems I was having with the system unable to boot with more than just a video card installed.

Eventually I will look into replacing the buffer chips - if the problem ever comes back.