Freezing G5 Quad: Logic board swap does not fix the problem

Posted by Pierre Igot in: Macintosh
October 2nd, 2006 • 1:54 pm

Regular Betalogue readers now that, for the past few months, I have been struggling with my G5 Quad computer, which has been experiencing freezes or kernel panics on a regular basis. I have made all kinds of attempts to identify the cause of the freezes, and have been unsuccessful so far.

Last time I wrote about this problem, I mentioned that I had been able to log in remotely from another machine even after the G5 Quad had become completely frozen. But it’s quite possible that this was just a one-off. Certainly when the computer does not freeze but has an actual kernel panic, it’s hard to image that there is any possibility that one could still log in remotely.

In any case, last week I also contacted the “local” authorized repair shop about the problem, after AppleCare told me to do so. Since getting there is a three-hour drive (to the provincial capital Halifax) for me and they are aware of this, they offered to order replacement parts for all the parts that might be involved (logic board, power supply, CPU, etc.) and call me once they had all the parts. (They can return unused parts to Apple afterwards.) They had all the parts on Thursday, and I went there on Friday.

After describing all the symptoms and all the troubleshooting steps I had taken over the past few months, I left the machine with them for a few hours. When I came back, the technician told me that, of course, they had not found anything wrong with the machine in their testing (I didn’t expect them to), but they did replace the logic board with the replacement one that they got.

They tried to also replace the CPU unit, but for some reason the replacement one that Apple had sent them didn’t work at all. In fact, the technician told me that it looked different, and he simply couldn’t get it to work. So they were unable to replace the CPU and he had to put the old CPU unit back in. It took another hour and it was about closing time, so he told me that there was one last step that needed to be done, which was thermal calibration of the CPU. It would take a little while and he didn’t have time to do it, so he said I would have to do it myself when I got home.

He put a CD-R of the software needed in the machine, and told me I just needed to boot from the CD and run the thermal calibration tool, which would take up to 45 minutes. I got home on Saturday morning and followed his instructions. The thermal calibration tool itself said that it would take between 10 and 30 minutes, and in actual fact it only took something like 20 minutes. At least I think it did. The tool doesn’t exactly have a user-friendly interface and does not tell you when it’s finished. It just reported in its log that both “CPU1” and “CPU2” had been calibrated successfully, so I quit when that was done, after waiting for a little longer and seeing nothing else happen. (I am assuming here that there are only two “CPU” units listed because the Quad actually contains two dual-core G5 processors, and not four separate CPU units.)

After that, I booted the machine normally. A few basic settings had been lost, no doubt due to the logic board swap, but other than that everything seemed to be normal… except that the machine froze again within a couple of hours. And again later in the afternoon I got a kernel panic—the old “ugly” kind of kernel panic with lines of white text over a black background on top of the frozen Mac OS X screen.

So obviously the logic board swap did nothing to fix the problem. The machine froze (or panicked—I can’t tell the difference when it happens with the monitors asleep) again once during the Saturday/Sunday night, but has not frozen since—which of course doesn’t mean anything. All it means is that the problem is still there, and we still have no idea what causes it.

The only other “new” thing I have noticed is that, since doing a clean install of Mac OS X a couple of weeks ago, after the freeze/panic and the hard reset, I now get the “Mac OS X has unexpectedly quit” dialog during the restart process, which I never used to get before. The dialog provides a log of the kernel panic. I have kept those logs, even though the technicians in Halifax told me that it wouldn’t help them in any way.

Typically those logs start with some error message regarding “cpu 1” and then provide a configuration list with the kernel version, machine model, memory modules installed, etc.

The error messages in the beginning of the logs are not all the same. One says:


panic(cpu 1 caller 0x0003EE58): wait queue deadlock - wq=0x50f9aa4, cpu=1

followed by the “latest stack backtrace for cpu 1.”

Another one says:


panic(cpu 1 caller 0x000A5048): simple lock (0x0038BF50) deadlock detection, pc=0x0003C5EC

again followed by the “latest stack backtrace for cpu 1.”

Yet another one says:


Unresolved kernel trap(cpu 1): 0x400 - Inst access DAR=0x00000000E0292000 PC=0x0000000000000000

this time followed by the “latest crash info for cpu 1.”

I don’t know if this means anything. All the logs mention “cpu 1” and not “cpu 2,” but this could just be a coincidence.

It’s all very frustrating. I am going to have to see if I can get the repair shop to try and order a replacement CPU unit again and make sure that they get a working one this time, and then go back there to get it installed and see if that makes any difference.

In the mean time, I am going to try a couple more things, such as borrowing RAM modules from a colleague who has the same G5 Quad and has no problems with hers. (The only situation where I haven’t yet been able to get a freeze is when running the G5 with its original 512 MB of RAM only. But it could only be because I didn’t wait long enough. Running with only 512 MB of RAM is quite frustrating and I must admit that I haven’t been able to do it for more than 3-4 days. I do have work to do with this machine, after all.)

The only good piece of news that I got last week was that my employer is going to buy me a new Mac Pro. (I bought the G5 Quad with my own hard-earned money as a “personal” machine.) The last one that they bought me was the G4 MDD four years ago, so I was due for a new one.

Once I get the Mac Pro (the shipping estimate is the middle of October right now), I will be able to bring the G5 Quad to the repair shop in Halifax and leave it there with them and get them to fix it once and for all. Hopefully if they keep it for a while they will be able to reproduce the freezes/kernel panics themselves.

But the whole experience is still very frustrating, especially for a machine that was supposed to be top-of-the-line a year ago, and didn’t have any problems for the first eight months (except for a slightly defective video card). I have a three-year AppleCare warranty for the G5 Quad, so there is no concern of being stuck with a freezing machine out of warranty, but obviously I want to get it fixed and in working order, whether I decide to keep it for myself or resell it.


3 Responses to “Freezing G5 Quad: Logic board swap does not fix the problem”

  1. apple4ever says:

    Hi Pierre. I’m an Apple Certified Desktop and Portable Technician. I have a question about the logs. They can be helpful if the capture the right information(sometimes they don’t, especially on Intel Macs.) Do they have any lines similar to this log:

    Kernel loadable modules in backtrace (with dependencies):
    com.apple.driver.initioFWBridge(1.4.7)@0x3bb55000
    dependency: com.apple.iokit.IOFireWireSBP2(1.7.5)@0x3bb38000
    dependency: com.apple.iokit.IOSCSIArchitectureModelFamily(1.4.9)@0x388ac000
    dependency: com.apple.iokit.IOFireWireSerialBusProtocolTransport(1.4.4)@0x3bb50000
    dependency: com.apple.iokit.IOFireWireFamily(2.2.5)@0x385a6000

    This was from a kernel panic caused by a bad Firewire drive. Do your logs have anything like this at all? From what I’ve read of what you have tried and seen, and I would guess that its probably the processors. A possibility is that it might be a bad HD. I’ve seen bad HDs cause kernel panics. An option is to install SMARTmontools (Click on SMARTmontools for OSX installer.) Post what that prints out, and I can tell you if anything is wrong. You can either post a comment here, or shoot me an email at mbutch@macoutfitters.com.

  2. Pierre Igot says:

    a4e: Thanks for the suggestions. Unfortunately, I have nothing like this in my logs. I am sending them to you by e-mail, just in case you can see anything.

    But as indicated in earlier reports, I have been able to reproduce the freezes/kernel panics even with NO peripherals attached other than the mouse and keyboard.

    As for the internal hard drives, I suppose it cannot be entirely ruled out. But I have tested both drives, and the S.M.A.R.T. status in Disk Utility is normal.

    But I’ll send you the stuff by e-mail just the same. Thanks again for your help!

  3. henryn says:

    Pierre:

    AAARGGGGG. I’m sorry that you’re having such problems.

    I hope “apple4ever” will be able to help you figure out what is going on.

    At this point, my intuition says “memory”, so I’m hoping your total swap-out will prove successful. If that doesn’t help, then maybe you ought to look at the internal drives. HDDs are so cheap these days, you might want to start completely fresh with a new drive, format it, install OS X, and then re-configure so it is the boot (and only) drive. Run for a few weeks…

    No comfort to you, but these boxes are incredibly complex inside. Some — most — Macs work continuously for very long times without glitches, or with very few.

    Good luck!

    Henry

Leave a Reply

Comments are closed.