Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

O oh! Kernel panic. Is it RAM? GPU? What next?

The storm clouds were gathering.


Thunder and lighting outside. I kept using my Mac Pro, but during the storm it froze. Ouch!


It's an early 2009 / Nehalem machine, series MacPro4,1, with 6 gigs of RAM and a NVIDIA GeForce GT 120 video card.


The machine starts booting. Sometimes it fails half way and freezes. Sometimes it gets to the desktop for 5 or 10 seconds before freezing. I've tried everything. At most, it is booted for just a few seconds before failing.


-Zapped the P-RAM (NVRAM).

-Safe boot makes no difference

-Swapped out RAM cards in different combinations, makes no difference.

-It can't boot from the install DVD. Halfway through it powers off.


Not looking good, is it?


When it freezes half way through booting, it gives a 'panic' error message, which I've copied at the bottom of this post. Can anyone decipher it?


Can you tell if the problem is RAM, or the graphics card, or something else? If it is the graphics card, how can I find a replacement for this old Mac? They don't sell the original graphics cards any more. It's currently running Snow Leopard OS.


Here's the error message on startup:


panic(CPU 0 caller 0x2aaf41): Machine Check at 0x002b1eb9, thread:0x57607a8, trapno:0x12, err:0x0,registers:
CR0: 0x8001003b, CR2: 0x000a0928, CR3: 0x00100000, CR4: 0x00000668
EAX: 0x00000000, EBX: 0x00000022, ECX: 0x7139688c, EDX: 0x00000023
ESP: 0x3092bb78, EBP: 0x3092bb78, ESI: 0x00000023, EDI: 0x00000000
EFL: 0x00000002, EIP: 0x002b1eb9

Debugger called: <panic>
Backtrace (CPU 0), Frame : Return Address (4 potential arcs on stack)
0x819e1

Posted on Mar 30, 2014 11:52 PM

Reply
9 replies

Mar 31, 2014 10:54 AM in response to veg

I have ridden out a lot of storms with CyperPower's UPS 1500VAs, one for each computer (and on different circuits too). Everything electronic basically in house has UPS (TVs, some lamps) it is essential to do it properly, and to protect both any modem and routers or switches as well.


http://www.amazon.com/CyberPower-CP1500PFCLCD-Sinewave-Compatible-Mini-Tower/dp/ B00429N19W/


You may need a 2nd unit smaller for peripherals, and depends on how much run time to ride out a storm or gracefully shutdown. With APC RS1500 a 2nd batter pack can be connected to add an additional hour of runtime (it is just a basic 24v car battery).


And done properly would have warranty protection.


And duplicates of all your files and system OFF LINE copies (clones, backups) so you can be back up and running if necessary.

Apr 1, 2014 5:34 AM in response to Grant Bennet-Alder

When the Mac Pro boots to the desktop, it's only there a few seconds before it freezes. So I don't have time to get into the System Profiler.


Some interesting stuff:


I put the broken machine into Firewire Target mode, and used another Mac to read its hard drive. The broken machine crashed after a few seconds.


So I tried it again in Target mode. But this time I removed the broken Mac's nVidea graphics card. It then stayed afloat for about 8 minutes before stopping. That was enough time to rescue a few of my files.


That makes me think the problem may be with the graphics card.


There is a known problem where Mac Pros can freeze when in target mode for a while, so that might have caused target mode to end after 8 minutes.


I don't think I can afford an Apple Center repair, so I'm thinking of ordering a graphics card to see if it makes any difference. Otherwise, I assume the next suspect might be the logic board, but I don't really know. I dread to know how difficult a logic board might be to install.

Apr 3, 2014 2:45 AM in response to Grant Bennet-Alder

Thanks Grant, but I could not find any LED indicators near that RAM cards.


Some new info:


I finally managed to get the machine to do a short Apple Hardware Test without crashing. It can't do the long test because it always crashes. The short test gave this error message:


Alert! Apple Hardware Test has detected an error.


4SNS/1/40000000: TNOD- 127.875


Pass Number: 1,

Total Time Testing 47 secs


According to a CNET article that explains error code, this message means that the heat sensor that monitors the Northbridge chip on the main logic board is showing it is too hot.


Someone else posted a picture of the offending heatsink and sensor in an older forum. Apparently these sensors come loose and can be fixed with glue, but if you take it to an Apple Centre they only replace the entire logic board.


I've replaced RAM, graphics cards, optical and hard drives, but tinkering with the logic board is something above what I've done before. I'm a novice, but then again, I guess I've got nothing to lose. I really need my Mac Pro to work.


I think I will try to unscrew and inspect the Northbridge heatsink, and failing that I might order a new logic board on the internet. The thought of installing a new logic board scares me, but I don't think I can afford to pay an Apple service centre $1000+ to do it.

Apr 9, 2014 6:32 PM in response to veg

I fixed it!


Below I will show what I did, for anyone searching similar problems in the future.


The problem was the heatsink that sits on top of the Northbridge processor. It is located on the main processor board. See my photo below.

User uploaded file


This heatsink became slightly loose because two plastic rivet screws that hold it down had become worn.


I bought some 3mm wide / 12mm long nylon screws from my local electronic store to replace the originals. The new screws only just fitted. 10mm long would have been better.


I had to remove the main processor board to do this. Something I'd never done before. To remove the processor board I had to remove both the main processors and heatsinks. But this was also a good opportunity to clean dust away from hard to reach places, using a new paintbrush and a can of compressed air.


To reinstall the processors, it required cleaning with isopropyl wipes and reapplying the thermal paste. I used Arctic Silver 5 brand paste.


It was a similar process for the offending Northbridge heatsink. I thoroughly cleaned it. I made sure the little heat sensor underneath was properly positioned, and used some glue to hold it there. I also applied new thermal paste to the top of the Northbridge processor, before attaching its heatsink with new nylon screws.


I had nothing to lose by doing all this myself. The other alternative would have been sending it to the Apple store to get a new processor board.


I was thrilled when the machine booted up and was running properly again. I also feel a bit more confident with computers now!

Apr 10, 2014 2:15 AM in response to a brody

Thanks A Brody,


I don't have that much knowledge.


I just read other forums about how to remove processors.


The other incentive was that the logic board (processor board) would have been written off anyway if I had a professional repair done. They never fix things at a component level.


Having done it once, I feel confident if I have to remove processors and heatsinks in the future.


If you have an out-of-warranty machine, it's a good place to learn these things.

O oh! Kernel panic. Is it RAM? GPU? What next?

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple ID.