...
The motherboard is fixed, it got a damaged cia chip. Also some upgrades were made, PIO2 upgrade etc. So a new fixed mobo in town.
Is it certified that the cia chip was properly replaced and that the CIAs were not just swapped around?
Everything was put together and back in action. My Cyberstorm PPC, with 128MB of RAM, my Picasso IV -Paloma TV card module included-, my Deneb USB card, a DVDRW and a hard disk, everything powered from a Liteon PSU. The on board RAM is 4MB chip and 16MB Fast Ram on five sticks, 4MB each. Those sticks were not the ones that my mobo used to have, but they were reported working on another 4000.
If you're having problems it's never wise to test with a loaded system. Strip it down to the bare minimum first, install a basic OS and begin from there. It sounds like an OS independent hardware issue anyway so it wouldn't matter much what is installed.
This configuration was working well in the past.
But this is not the same configuration that was working well in the past. RAM sticks are different, upgrades have been performed... it's not an accurate comparison.
...
HDD light stays on.
No drive activity.
No booting.
There is caps lock on the keyboard and I am able to reset.
Being able to reset is little comfort, it just means that the circuitry which detects the 500 ms KCLK low pulse is working ok and once this specific pulse is detected it asserts _KBRESET (hard reset).
HDD light has nothing to do with the motherboard, it's driven by the hard disk itself.
And sometimes a green screen on boot and then black screen, or as I said freezing on Deneb's intro screen. If I turn off the machine and turn on again for example on three or five minutes, it is still not booting.. But if I leave it off for a half an hour, or a quarter. It would boot normally.
The green screen with a Deneb present is common even with working systems, I wouldn't put much weight on it. May have to do with the state of Deneb's flash ROM. Anyway, you should have the Deneb removed (along with everything else) while testing this out, why would you insist on testing with a loaded system is beyond me.
So,
"Green screen means bad chip RAM", I thought. So I changed the RAM. Pulled out the five sticks of 16MB I was ussing and tried to boot with no RAM at all, thinking the the RAM on Cyberstorm will do the job.
The first stick on revB motherboards is Chip RAM, no real Amiga can boot without chip RAM. So at least that stick must be present. RevD motherboards have the chip RAM onboard, so it's a given there.
No booting. I tried with a single 8mb stick. Machine booted. I think I was using this 8MB stick initially on this Amiga.
No booting as expected due to the lack of chip RAM, yes booting as expected with chip RAM back.
Now the MorphOS is booting fine. So that kind explains that this should be a RAM problem. But the same prob remains. After second or third rebbot the nachine locks up on boot.
I have put the 8MB stick on the upper chip RAM socket.
Which is the chip RAM socket.
The machine reboots perfectly on every soft reset from keyboard.
It's not a soft reset, it's a hard reset as described above. Soft reset on the Amiga is not present on all machines (it's actually called "reset warning") and should not be relied upon.
This PIO2 upgrade mainly improves hard disk speed, trying to test the hard disk behavior copying sn entire partition through AmigaOS 4.1, everything worked well, no crushes. Done that thinking that this upgrade could be the problem. It is actually a chip replacement on mobo.
At this point I would try to forego using IDE and just install a SCSI disk on the CS PPC. I never understood the obsession with the PIO2 upgrade anyway, but that's a personal choice. Useless to me. So if the GALs that were replaced have some sort of initialisation trouble after they've heated up, it could be the culprit. You could try keeping them cold with heatsinks etc and see if anything changes.
Do you think is a memory issue?
No.
Is there a way to bypass the on board RAM and use only the Cyberstorm RAM? Yes it is brand new.
You definitely cannot bypass the chip RAM as already explained.
You can leave the motherboard fast RAM off and only use the RAM on the CS PPC, after all access to the motherboard RAM is much slower as it's done via 68030 bus emulation.
Any further thoughts, ideas, advice?.
Sounds like a bad IC problem. Or a bad solder joint problem, but that usually works the other way round.
If the Buster chip is socketed, you can pull it out and carefully clean its pins by brushing them softly with a white rubber. Not that much should be expected from this.
