PDA

View Full Version : Bad CPU or Bad RAM?



DarkAvenger
15th September 2007, 11:46 PM
I have a Powermac G4 Quicksilver Dual 1.0 GHz with 1.5 GB RAM running OS 10.4.10 (all applications are the most current updates). Ever since I bought it I've had continual problems with it always crashing (i.e. total system freeze so the Mac has to be turned off then back on). Macs do occasionally crash of course, but mine's crashing 9-10 times daily! (over typical 8 hours usage a day).

Eventually I became so frustrated I took it to my nearest Apple Centre and explained the frequent crashes. I explained the following symptons:

Short double beeps followed by pause then short double beeps on booting. This also happens when I shut down. This happens 2-3 out of every ten times and will generally last over several minutes. Sometimes it can take several minutes just for the Mac to boot even without any double beeps.

Applications continually crash. Mail, Quicktime and Safari are the biggest culprits I've observed, with Quicktime guarenteed to crash the computer 'without fail' every 9 times out of 10 (will happen a few seconds into playing any video), but it will 'only' happen if I have another application open.

Continually get the 'spinning beachball of death', sometimes with the double beeps. No specific applications observed, but computer always seems slow like its only running with 256 MB RAM instead of 1.5 GB RAM.

Everyone here will agree it 100% sounds like a symptom of 'bad RAM' and thats what the tech guy said too. So he replaced both RAM modules (both bought there) with brand new ones. Guess what though? Its made absolutely no difference!

It seems improbable that two lots of RAM, both bought from the same Apple Centre, could both be bad! So I'm wondering was it really bad RAM or something else? Could it be the original factory installed 512 MB RAM the machine originally shipped with that's actually bad? I suggested that and was told that was extremely unlikely.

Now I'm thinking could I have a faulty CPU? The crash reports always start with:

Crash Report: CPU=1

Is this significant or does it just mean in the event of a crash CPU 1 will always get haulted before CPU 2?

I'm about to soon take my computer in for the second time. Only its not good enough that they'll simply replace the RAM again because there's going to be a 99% chance it'll still make no difference and that's just not acceptable!

Does anyone here have any suggestions on what could be causing the frequent crashes? I've run 'TechTool Deluxe' numerous times and all tests come back 100% no hardware/memory faults detected.

Is it possible the supplier has 'pulled a fast one' on the Apple centre and supplied them with 'generic' PC133 SDRAM (3.3v) that's not recomended for use in Macs?

For your info here's what the System Profiler says about the installed memory:

Memory Slot Size Type Speed Status

DIMM0/J21 512 MB SDRAM PC100-322S OK
DIMM0/J22 512 MB SDRAM PC100-322S OK
DIMM0/J23 512 MB SDRAM PC133-333 OK

And yes I'm aware the System Profiler simply reads the ID (which is often incorrect) and so that doesn't mean the wrong type of memory is installed. In fact the tech guy verified all three are PC133 (3.3v).

g5agogo
16th September 2007, 05:34 PM
Forget system profiler, try memtest or rember (http://www.kelleycomputing.net:16080/rember/) to test your ram.

Also, if you really did have some bad memory and installed any major software upgrades, this could cause you problems after the faulty ram was replaced.

I wouldn't really be keen on the fact that some ram is reported as pc100 and some is reported as pc133 with cas2 or cas3 latencies - although most ram is pc100/133 these days, as I recall it was critical for those machines that the ram was matched for speed and latency. Actually, my memory (poor as it is) makes me think that it was critical to have CAS2, maybe I'm wrong (you could try googling it). Have you tried removing the "new" ram and seeing if you still have the same problems?

Currawong
16th September 2007, 05:42 PM
Could be a dead motherboard as well.

kim jong il
16th September 2007, 05:58 PM
As you have, seemingly, eliminated RAM as the culprit I was reminded of a similar issue discussed here recently. Brains was talking about this a few weeks ago in response to another member with a DP MDD that experienced constant kernel panics. He suggested that it is the CPU daughterboard (flaky level 3 cache) that might be the issue and causes all sorts of weird behaviour however Currawongs suggestion that your motherboard may be fried is also a definite candidate. My next stop in troubleshooting would be to try your CPU daughterboard in another QS or MDD even. It may be possible that a local MacTalker could help you with this but you would have to appeal. :)

g5agogo
16th September 2007, 06:09 PM
I honestly don't think that a ram conflict has been eliminated yet, at all. I think your first move is to run rember, if that gives you the thumbs up, then your next move is to remove the 2 new sticks and see if you still have the same crashing problems.

Only then would I consider that ram problems have been eliminated.

Arkhum_Eramak
16th September 2007, 06:26 PM
I had a similar issue with a dead Hard Drive

g5agogo
16th September 2007, 10:35 PM
Just to revisit this, I'd say the supplier has definitely screwed up. The ram should be PC133 not PC100. The Quicksilver has a 133 MHz bus.

Mctastic
16th September 2007, 10:50 PM
yep, . i'd say from your post above, ... the ram will still be an issue...

you have pc 100 & 133 chips mixed. i also know for instance that if all the ram was pc133, .. but 2 chips 322 s & 1 chip 333, .. this would also cause a problem, because that relates to the speed/timing of the chip. take it back, get them to match the ram properly.

kim jong il
16th September 2007, 11:57 PM
Just to revisit this, I'd say the supplier has definitely screwed up. The ram should be PC133 not PC100. The Quicksilver has a 133 MHz bus.


How embarrassing*, I didn't pick up on that. :o

*(PC 100) It's never going to work.

DarkAvenger
17th September 2007, 02:32 AM
As I'm typing this reply I get my eighth kernal panic (shut down). Here goes again...

Firstly all installed RAM is PC133. I was told by the guy at the Apple Centre that System Profiler is incorrectly reporting the PC133 RAM as PC100 because its simply reading the manufacturer ID from each RAM module. In other words its the ID that's wrong, not the wrong type of memory installed.

Anyway I ran Rember (with applications and finder turned off), just once as it takes a few hours to complete. Here's the end result:

*** Memory Test Failed ***

Here's the transcript: (just the highlights)

Running 1 test sequence...

Test sequence 1 of 1:
Stuck Address: ok
Random Value: ok
Compare XOR: ok
Compare SUB: ok
Compare MUL: ok
Compare DIV: ok
Compare OR: FAILURE: 0xbff22f94 != 0xbff22f9c at offset 0x055d75d6.
Compare AND: FAILURE: 0xbf720584 != 0xbf72058c at offset 0x055d75d6.
Sequential Increment: ok
Solid Bits: ok
Block Sequential: FAILURE (on testing 93 of 256): 0x5c5c5c54 != 0x5c5c5c5c at offset 0x055d75d6.
Checkerboard: ok
Bit Spread: ok
Bit Flip: ok
Walking Ones: FAILURE (on testing 38 of 64): 0xfbfffff7 != 0xfbffffff at offset 0x055d75d6.
Walking Zeros: ok

So does this positively prove 'at least' one of the RAM modules is faulty? Do I need to verify by running the test again or even several times? Would the results be different each time e.g. different areas of the tests failing?

At this point I get another kernal panic (shut down) so I have to type the reply all in again!

kim jong il
17th September 2007, 02:47 AM
If it's failing it's failing. I personally don't believe that the system profiler is reporting incorrectly as my G4's (100&133MHz bus) seem to manage okay in telling me the truth. Furthermore if the RAM test failed once it's unlikely to tell you something you want to hear the second time around (IMO). You really need to test it with modules that are known to work at the correct speed.

EDIT: a shame you don't live in Melbourne as half a dozen Coopers sparkling would probably get you a home troubleshooting visit. :)

DarkAvenger
17th September 2007, 03:06 AM
I saved the memory test transcript so I can show it to the guy at the Apple Centre. I'll give them a call tomorrow. Thanks everyone for your comments.

g5agogo
17th September 2007, 07:47 AM
1. You should be able to work out which ram module(s) are reporting faults from rember
2. I'm a bit suspicious about the reported pc100 modules, and in any case as best as I can recall these systems were a bit fussy about having the same CAS latency. You can check which modules you have by visual inspection - check what's printed on them (if the speed and latency is not obviously printed on them, which it usually is, then you can google the manufacturers chip code).
3. I'd want to check the system without the two new ram modules. Take them out and run rember again. Then, assuming rember reports no errors in the original module, use the system for several hours and see how it behaves - I'd be willing to bet you a beer that the kernal panics and crashes disappear.
4. It still sounds to me like a screw-up by the store. I'll apologise if I'm wrong.

good luck!

stewiesno1
17th September 2007, 09:46 AM
Like all the others that have replied , I too think it is a bad Ram chip or two , whether the incorrect speed or just a failed chip. I would strip it back to just one chip and test it to see if it works OK ,then another etc. You should be able to track down the culprit fairly quickly by this method. The other thing is to disconnect any peripherals and PCI cards and plugins and start with just a known good keyboard. Once again introduce them back one at a time. I presume you have done a hardware test with TTP or similar. What about your PRAM battery ? Is that OK? Maybe a new one of those is in order.
Failing all this look at maybe a bigger hardware failure like the Logicboard as KJI suggests. Try all the little things first and work your way up.
Stewie