Previous 1 2 Next 21 Replies Latest reply: Jan 12, 2009 9:14 PM by Peter Nelson1
Peter Nelson1 Level 1 Level 1 (10 points)
Hello,
I just recently acquired a B&W G3 (450 MHz) to replace my old Beige G3. I transplanted the 40GB & 80 GB IDE hard drives, along with a USB 2.0 PCI card, from the Beige to the B&W. Although it generally boots up OK, it seems to have a problem freezing up--several times a day lately. It would appear that this is probably hardware related, because (1) it froze while running the Panther installer (booted from the install CD), and (2) it froze last night while idle, waiting for its next scheduled Retrospect backup run.

Can anyone on this board offer advice for conclusively diagnosing a freeze problem like this? It would really be a shame to have to junk the whole machine over a problem that may be fixable by replacing a minor component.

I did run TechTool Pro (some pre-OS X version; don't remember the exact version number) under OS 9.2.2 on the original 9 GB SCSI HD, before transplanting the 40 & 80 GB drives. No problems were detected at that time. I have not yet re-run TechTool since the transplant; I am ultimately planning on upgrading to an OS X compatible version (Micromat's current version has a CD image for a Panther-compatible back-level version); in the meantime, I will see if anything shakes out under the 9.2.2 version.

Thanks in advance for any help you can provide here.

Pete

B&W G3 450, Mac OS X (10.3.x), 640 MB RAM; 40GB & 80GB HD's
  • Dave Hamilton Level 5 Level 5 (6,805 points)
    Have you reset the PRAM?
    http://support.apple.com/kb/HT1379

    Also, if you are using the OS install from the Beige in the B&W, it may not have the correct files for the B&W.
  • Grant Bennet-Alder Level 9 Level 9 (51,750 points)
    That sounds like you have the "HeartBreaker" Rev 1 Blue & White G3. It is sort of flaky... never the same thing twice... It almost works right... sometimes it crashes... sometimes the Drives get corrupted... seems to happen at times of highest disk activity, like installing software.

    Have you made sure that you do not have the deadly combination of Rev 1 Blue & White AND a Hard Drive Faster than the original or Larger than about 40 GB (which are almost all Faster than the original)? This article contains the "litmus test" to look for under "3. Revised IDE Controller Chip":

    http://www.xlr8yourmac.com/G3-ZONE/yosemite/newfeatures.html
  • John Helock Level 2 Level 2 (195 points)
    The B&W G3 is very picky about ram. It is best to use matched ram that is PC100, 8ns, CL2, non-parity. There is a lot of ram that is PC100 but is 10ns. It has been my experience that this will lock up the system with annoying frequency.

    If you have OS 9 installed, try downloading and running Dimm First Aid to check your ram.

    http://www.mactcp.org.nz/dimmfirstaid.html
  • Peter Nelson1 Level 1 Level 1 (10 points)
    Grant,
    It is definitely not the Rev 1 B&W--I have verified that the IDE chip has the "402" marking as illustrated in your reference link. This is the 450 MHz edition that shipped with a 9GB SCSI drive and 128MB RAM.
  • Peter Nelson1 Level 1 Level 1 (10 points)
    John,
    I am inclined to suspect bad RAM myself after reading some other posts in this forum. I have downloaded the freeware Rember utility, and will be running it to see if anything shakes out. I have already rotated the 128MB DIMM from the J19 to the J22 slot (J20 & J21 each having a 256MB DIMM) to see if it will either minimize freezes or put potentially defective memory in an address range that can be tested.

    Some additional background: I acquired this machine from another family member who bought it new back in late '99. The original owner had added the two 256MB DIMMs within the last year; the 128MB DIMM shipped with the B&W originally. When he upgraded to an Intel-based iMac this year, he did experience some freezes on the B&W while copying files onto a flash drive for transfer to the iMac--had to resort to copying a few files at a time. Since he was still running OS 8.6 on the B&W, I figured that things might be more stable on it after an OS upgrade.

    When I first got it, I installed OS 9.2.2 on the 9GB SCSI drive, ran TechTool Pro 3.0.5 for a hardware check, and then installed Panther on the drive. Bringing Panther up to 10.3.9 took a couple of extra iterations with Software Update, due to freezes. Once I had Panther installed on the SCSI drive (along with Retrospect), I disconnected the power cable from it, to save it as an emergency boot drive. So at this point, it is running with the IDE drives only.

    I may consider purchasing additional 256MB DIMMs (to take it up to 1GB) in the near future; at that point I could try a DIMM rotation/holdback strategy to flush out the bad module. For now, I'll just rotate what I have, and see if I can find anything that way.
  • Jeff Level 6 Level 6 (11,290 points)
    "Can anyone on this board offer advice for conclusively diagnosing a freeze problem like this? "

    To rule out a problem caused by an overheated processor, have you checked out the chassis fan and the power supply fan to be sure that both are clean, running smoothly, and moving air? Given the age of the computer, you may wish to remove the processor's heat sink and remove the old thermal pad from it, apply a good quality thermal compound (like Arctic Silver) to the processor, and then reinstall the heat sink. If you do this, be sure that the heat sink retainer clip is installed correctly, so that the offset/bend in it is positioned directly over the center of the processor.
  • Peter Nelson1 Level 1 Level 1 (10 points)
    Jeff,
    I did check the fans, and both appear to be in working order. I may consider recoating the heat sink at some point.

    After running Rember all day (about 10 complete passes, with 450MB under test) with no errors reported, and then getting still more freezes this evening, it occurred to me that maybe I should try removing the SCSI card for a while. I'm not using any SCSI devices anyway, and with the freeze history going back to the original owner, a bad SCSI card is as reasonable a suspect as bad RAM. So, the SCSI card is now pulled, and we'll see what happens...
  • Dave Hamilton Level 5 Level 5 (6,805 points)
    If the SCSI card is the OEM Adaptec 2940U2B, there is a firmware update for it.
    http://docs.info.apple.com/article.html?artnum=60409
  • Peter Nelson1 Level 1 Level 1 (10 points)
    Dave,
    I did come across that particular firmware update, but it looks like you need to be running on OS 8.6 in order to apply it. I already had upgraded to OS 9 by the time I discovered the FW update, and do not have a copy of OS 8.6 to revert to (did not get any type of restore CD with the B&W).

    In any case, I am still getting freezes with the SCSI card pulled, so that hasn't been the culprit. I have even had it freeze up in mid-reboot, right after the startup progress bar window appears (the wave animation on the filled-in part of the progress bar freezes). No amount of RAM rotation has caused any bad RAM to be detected by Rember or TechTool. I was at first hopeful after pulling the SCSI card, as it looked like I was getting random app crashes instead of freezes. Alas, the freezes have returned--sometimes in as little as 20 minutes after a cold startup.
  • Grant Bennet-Alder Level 9 Level 9 (51,750 points)
    8.6 was the current version when that firmware update was issued. You should read the article as if it said OS 8.6 through 9.2.2.

    It will install under OS 9 just fine.
  • Grant Bennet-Alder Level 9 Level 9 (51,750 points)
    You may want to download the free Utility DIMM First Aid to take a look at your RAM settings:

    http://www.mactcp.org.nz/dimmfirstaid.html
  • Peter Nelson1 Level 1 Level 1 (10 points)
    Got a couple glimmers of hope, but alas, the freezes keep on happening.

    I downloaded and ran DIMM First Aid, and discovered that my 256MB DIMMs are 8ns, while the 128MB DIMM (the original) is a 10ns. Recalling a previous post here that indicated 8ns DIMMs are preferable, I have pulled the 128MB DIMM and have been running on 512MB. Hasn't worked to cut down on freezes.

    The last freeze that happened during a Retrospect scheduled backup actually took the form of the multilingual black-box "need to restart" message. I was hoping that its appearance would mean that some diagnostics got logged somewhere, but nothing even remotely useful has turned up anywhere.

    I am still hoping that, one of these days, I will actually find some conclusive evidence to identify the failing component, so that I can replace it with confidence and resolve this freeze problem once and for all.
  • John Helock Level 2 Level 2 (195 points)
    Peter,

    In a way, the kernel panic is good. By this I mean that a panic log has been recorded. You can read the panic log via the Apple System Profiler or at /Library/Logs/panic.log. If you can locate and post this log file, we may be able to get to the root of the issue.
  • Peter Nelson1 Level 1 Level 1 (10 points)
    Here's the KP log I found:

    Wed Oct 22 23:02:38 2008




    Unresolved kernel trap(cpu 0): 0x400 - Inst access DAR=0x00000000E3341000 PC=0x0000000000000000
    Latest crash info for cpu 0:
    Exception state (sv=0x1F688000)
    PC=0x00000000; MSR=0x40009030; DAR=0xE3341000; DSISR=0x40000000; LR=0x00000000; R1=0x0CC8BA40; XCP=0x00000010 (0x400 - Inst access)
    Backtrace:
    0x00000000
    Proceeding back via exception chain:
    Exception state (sv=0x1F688000)
    previously dumped as "Latest" state. skipping...
    Exception state (sv=0x1F681000)
    PC=0x90010C8C; MSR=0x0000F030; DAR=0xE3341000; DSISR=0x42000000; LR=0x902B569C; R1=0xBFFF9F30; XCP=0x00000030 (0xC00 - System call)

    Kernel version:
    Darwin Kernel Version 7.9.0:
    Wed Mar 30 20:11:17 PST 2005; root:xnu/xnu-517.12.7.obj~1/RELEASE_PPC


    panic(cpu 0): 0x400 - Inst access
    Latest stack backtrace for cpu 0:
    Backtrace:
    0x00083498 0x0008397C 0x0001EDA4 0x00090C38 0x0009402C
    Proceeding back via exception chain:
    Exception state (sv=0x1F688000)
    PC=0x00000000; MSR=0x40009030; DAR=0xE3341000; DSISR=0x40000000; LR=0x00000000; R1=0x0CC8BA40; XCP=0x00000010 (0x400 - Inst access)
    Backtrace:
    0x000DF41C 0x000DF700 0x000DF978 0x000BEC44 0x000BE68C 0x000CB2C8 0x00246D84 0x000941C0
    0x00090009
    Exception state (sv=0x1F681000)
    PC=0x90010C8C; MSR=0x0000F030; DAR=0xE3341000; DSISR=0x42000000; LR=0x902B569C; R1=0xBFFF9F30; XCP=0x00000030 (0xC00 - System call)

    Kernel version:
    Darwin Kernel Version 7.9.0:
    Wed Mar 30 20:11:17 PST 2005; root:xnu/xnu-517.12.7.obj~1/RELEASE_PPC



    *******
Previous 1 2 Next