Try a different CPU card if you possibly can though, preferably a simple one like the basic '030 one.
Can you please confirm that your 68040 is 40MHz and not the 25MHz? Because if it's 25MHz you want the clock jumpers to be set to INT not EXT...
Agreed, AmigaKit will have certainly tested the board. Sounds as though something has been upset with the Cyberstorm board when it's been removed and refitted. Could be anything from dirty connectors to cracked/intermittent solder joints. It's not a
Cyberstorm mk3 by any chance? I'd agree that a known good CPU board is the best way to test.
I'm intrigued as to why the clock source jumpers should be set to internal when the CPU board is operating at 25MHz, please fill me in here.
By my understanding, a board such as the Cyberstorm is clocked using its own on-board oscillator, so obviously the A4000 main board needs to be exactly synchronised for all bus communications with the CPU board, which is the entire point of shifting the jumpers to 'EXT' so that the timing of the main board is identical to that of the CPU board. If the CPU board and main board are using different clock sources (jumpers set to INT) then both clock sources must be identical in phase and frequency. From memory the stability of those oscillator modules are around 20ppm or worse, so to have two oscillators running at exactly the same phase and frequency for any length of time is practically impossible. Which would mean the computer will either fail to boot, or not run for very long at all.
