MacBook Pro 2,2 won't pass memory test at boot

Question

Hi Apple and Mac folks,
I have a MacBook Pro 2,2 15", 2007 year model, equipped with 2,16 GHz Intel Core 2 Duo and 2 GB 667 Mhz factory-installed DDR2 SDRAM, running Mac OS X 10.4.11. SMC version is 1.12f5.
Three months ago, when I was installing the Safari beta and the installation requested a reboot, the machine wouldn't start at all. It didn't give me any startup chime; the only thing to be heard was the initial movement from the DVD scanning mechanism (?), but after that nothing happened. If I waited for a while I could assume that the CPU was working, since I could hear the fans spinning up. But the screen was black and the computer wouldn't load any CDs or DVDs or even listen to any keyboard commands. The only thing that was working was the force shutdown by the 10-second press on the start button.
After some trial and error I realized that the system would reboot without problems if I let it rest (cool down?) for about 15 minutes. Sometimes, connecting an external hard drive via FireWire would also help the system to start up, bit it seemed to work rather randomly.
I ran the Apple hardware test and it reported an error on one of the SDRAM modules the first time I checked, but running the more thorough test wouldn't report any error and the system didn't show any signs of errors whatsoever once it was up and running.
I have also tried to reset PRAM and SMC, but it does not alleviate the reboot problem.
Since I then suspected that I had a faulty memory module I decided to upgrade the system with 2 Kingston DDR2 2GB SODIMM modules (yes, I know it only supports 3GB of memory, but I wanted the modules to work in pair). Unfortunately, the system exhibits the same problems with the new modules, and now things are getting even worse - now the 15-minute cool down period does not seem to be sufficient to get the system to pass (as I suspect) the initial memory test.
Last night I had to fiddle around, switching memory modules around in the slots to get the system restarted. I could get it rebooted once with one of the older 1 GB modules in the upper slot, and once it rebooted with an older module in the lower slot. Things seemed to act quite randomly and the previous cooling period method didn't seem to make any difference. Finally, I sort of gave up and put back the two new 2 GB modules and tried a last restart while having an external hard drive attached via FireWire and then it gave me the chime and restarted!
My interpretation of the problem is that the system has some kind of glitch (maybe not the correct term for a small gap in the circuits?) in the circuits connecting one or both of the SODIMM modules and that this glitching is getting worse as the system grows older.
So, I have three questions to you:
1. Is there any way to turn off or circumvent the initial memory check, so that I can temporarily live with this problem and not risk having an unusable system at next reboot?
2. Have you heard about this kind of problem before and maybe have knowledge about what part of the system that would have to be replaced and what it will cost me if I bring the computer to a service station (I'm located in Sweden).
3. Are there any SMC or EFI updates I should try before going to service?
Thanks for any help on this,
Ulf

Answer 1

cornelius wrote:
The firewire bridge is the most likely component in a FW HDD to fail. I have seen it a few times, and all that was needed was to remove the drive from the case and place it in a new firewire enclosure. I insist on getting an enclosure with an Oxford chipset, as they are the most reliable.
Hi again Cornelius,
first of all, thanks for the support in the struggle troubleshooting this issue.
Secondly, I don't think the problems has to do with the cabinet of the externa FW drive, but rather the problem would be located in the FW port in the computer.
Doing some research today, I found indications on that the AHT actually is installed on the hard drive of MacBooks after 2007. So, I decided to see if I could launch the AHT from my internal hard drive. Reboot needed, so hold your thumbs...
Reboot. Dead computer. Attached firewire disc. Dead computer. Let computer rest and cool down for half an hour. Restart -> dead computer. Connected firewire drive -> chime and boot!
Pressed D to load ATH. System gave me a dimmed screen with a text box telling me that I had to reboot the computer. I gave OK to that and the computer restarted normally, but reported an abnormal shutdown and gave me the opportunity to send a report about a kernel panic to Apple.
When the system came up the attached FW disc was not reachable. Looking in the system log, I find the following error message repeated:
Jun 26 16:24:49 Ulf-Kronman kernel[0]: FireWire GUID ffffffffffffffff is invalid!
The research on the net is getting me more and more confused: Is the ATH really located on the hard drive of MacBooks? Is the Power On Self Test really run at reboot? According to this wikipedia note (http://en.wikipedia.org/wiki/Power-on_self-test) the test is not run at reebot:
" Apple's Macintosh computers also perform a POST after a cold boot. In the event of a fatal error, the Mac will not make its startup chime. "
But my problems occur both when I cold start and when I reboot.
Isn't a lack of chime a very un-Apple-like way of reporting an error? Where/how can I get some more information about what's going on before the chime? And how do I know if the Self Test is run or not? System profiler only reports that self-test at start hasn't found any errors, but only when the system starts (due to that it not has found any errors )
Somewhere I saw instructions for doing a
Terminal> sudo nvram boot-args="-x -v"
to get a report of whats going on during startup. Could that be a way to proceed or is there a leaner way to get info out of the self test?
I'll continue to some research on the FireWire GUID, whatever that is.
Regards,
Ulf