DP MDD System Failures...

[wolfpweb]

Ok, here I go again; trouble in paradise...

1.25MHz DP MDD crashes, freezes & generally acts up about 1/2 hour into use, sometimes sooner. Doesn't seem to be HD or sw related, tried & switched several from 9.2 to Panther... Also went back to 2 x 256MHz OEM Ram, for testing. Same difference. And, yes, did several Firmware resets, to no avail.

Worked a-ok for some 4 months, got it as a 'refurb' from reputable dealer. Still has a gd 6 months of A.C. on it. Down time (for repair) will be a bummer. Did several system checks via A. Serv. Diagnostics vers. 2.1.4... One hour loop on everything. All "passed"... 3 hour loop on mobo, processors & 2GB of Ram. All checked GREEN. Photographed 'System failure' from screen, film not yet processed. Could be a heat issue since troubles start only after about half an hour into use... Fans seem to be working fine. Heatsink always got very hot. Once, when CS crashed, got "error code -600"... Once, after installing Raptor w. Seritek SATA PCI card for testing, reboot (after system failure) got "waiting for application services" upon reboot-freeze.

I've become rather savvy, especially after feedback from various threads on this here board, but current trouble seems way over my head.

Any educated advice would be most welcome. Thanks!

[rguising]

And I agree with you its most likely heat (but hard to tell too).

What kind of system temps are you seeing? Download Hardware Monitor from Bresinki (sp?) and it will give MDD CPU card temps (single temp).

We have an MDD here and I always enable Nap mode after a restart, CPU temps usually fluctuate at about 30C (vs 50C when purchased new). Nap is turned on by installed the CHUD package (latest CHUD works with 10.4.4). Believe its 4.x.x

Keep the MDD on a desk (ie off the carpet), pulled back from the wall. Only load 1 or two hard drives internally (if possible). Upgrade to the Copper Heatsink (usually available at SmallDog.com).

Finally, you could even become more daring and add more cooling fans (ie slot PCI fan), or more drastic measures.

MDDs are notorious for Heat issues.... keep it cool.

[wolfpweb]

Ok, makes sense...

Except, why now, not warmer in here than previously..? Only 1 HD at a time, in the bay away from the heatsink. Occasional ext. FW HD for file transfer. Positioned away from wall, not on carpet but on small 'printer desk' type base.

I don't have Tiger, only up to Panther. Anyhow, as I said, same trouble in 9.2 as well as in 10.2.8 (on the Raptor)...

Tecchie at vendor's said (on the phone), "try re-installing the processors"... Hm. Can the stuff between heatsink & CPU go bad..?

Of course, in my book, with my previous experiences, it could be the mobo.

To be contd.

[kaye]

As long as the case feet are on the printer desk, and not overhanging/dangling with the case itself on the printer desk.

Nicolas and I have a thread in Custom Configurations on what we did to our MDDs. He did far more than I and it is long. http://macgurus.com/forums/showthread.php?t=20094 k

[TZ]

crashes, freezes & generally acts up about 1/2 hour into use, sometimes sooner.

As an experiment, I removed all internal drives, including the Combo drive (and leaving a Pioneer) which helped air flow and bring temps down noticeably (more important in summer than now). Added a small fan up in the front top bay to blow air out of there and where the PSU and hot air seems to gather. Never tried pushing air in the front and/or pulling it out the back - but those do work.

Can't say enough good about Hardware Monitor
http://www.bresink.com/osx/HardwareM....html#Download

While an Open Firmware reset and set defaults is good for kernel panic, no reason to think that is what you need or seeing.

You are running off Raptor, so that means you have SeriTek or something. I would try removing any/all PCI cards, even if it means booting off FW and see how that is. Would also be nice and help to have a video card to try and help remove that as a source.

The thing with crashes, is that you really need to run fsck on restart. And Applejack. You don't want to just restart (though OS 9 will automatically check the boot volume). Disk Warrior would help. While software may not be the cause, you could end up with corrupt directory, file system, or just cache files and preferences.

It happens with any mix of programs running, not just one (Mail seems to give trouble to a few/some people for unknown reason).

Norton is notorious for some systems to cause trouble. But you say it isn't software.

I am always leery of refurbished except Apple (had bad experience with SmallDog on a number of Apple systems, mine and two friends, so I don't send or recommend to others). Would like to know what/why the system ended up there, whether they really replaced PSU or logicboard or something else.

I think NOT having a system on UPS can weaken or shorten the life of a PSU, I know my UPS died after two years from having to constantly adjust for house current, doing a lot of boost and trim work.

I've had cable to drive in ATA/100 cage vibrate or shake loose enough to cause I/O errors (picked up by SoftRAID).

Some people have found the Apple cooper heatsink helps (and for that, I would grab one if Small Dog has any, or someone else).

It doesn't seem or sound like memory (Memtest 4.1.2+) would be my choice, and change the order of the RAM chips, as well as it can be run from Single User Mode.

Kaye ran into trouble with MDD 1.25DP and ended up having someone check out to see if it was cpu or logicboard. But it sounds like you've ruled that out if you have something more thorough than the Apple Hardware Test CD.

I think Jaguar was totally beta quality. Not a good test. Panther is pretty hard to install and upgrade, it is "messy" to do all the updates and patches. Not pretty either. Which is why I prefer and recommend Tiger, esp. now as it ships on DVD and comes with 10.4.3 - not totally current, but almost.

[wolfpweb]

Thanks, Kaye, but feet are posted solid, same place -give or take an inch- from the very beginning. Correction on that one; been in my use since end of Aug. 2005...

Dr. TZ, I always appreciate Ur expertise, but if U read carefully from the start, U will notice that I've been through some of Ur suggestions, especially the HD's... Only installed the Raptor once for testing the possibility that trouble might come from IDE connect. Apparently doesn't. Also booted from an ext. FW HD with 9.2. Crashed or froze equally.

One of Ur tips could be the hot one... Thanks. Will exchange the nVidia TI 128 for one of my ATI 9000's 128 tonite after an afternoon outing. (Hope is in the air...)

To be contd.

[wolfpweb]

By the way: absolutely no PCI cards in this unit. Used strictly for heavy duty files in CS. All other stuff being handled by 2 smooth running QS'...

[TZ]

Then maybe there is something about using SATA with the bridge (PATA to SATA). And only way to tell would be to add SeriTek; and, put a nice PATA drive in there. Something like 250GB 16MB cache Diamond Max 10.

Out of curiosity, what kind of benchmark do you see with the Raptor? I don't think I've seen QuickBench from anyone with the same setup.

Yes, I know you said that you tested the drives or felt you had ruled out that being a problem. But the only way I could rule out hard drive would be to zero the drive and do a clean install of the system. Or, is it based on the fact that it happens when running off other drives on FW.

Heavy duty files in CS usually means 2GB RAM for MDD, along with a fast 4-8 drive RAID to compensate for lack of having 4-8GB memory. And, using either ATA controller or Serial. No harm in having 3-4 PCI cards, but always necessary when troubleshooting to remove 3rd party and bring the system to a 'stock' condition. (kaye used ATTO UL3D, SeriTek, and ACard ATA controllers, plus USB2 I believe. I have two UL3Ds and USB2/FW in mine.)

I was even thinking that if you didn't have a spare AGP video, might have PCI Radeon just for testing. Radeon 9800s (and 9600 OEM modified units) frequently do seem to 'go bad.'

[wolfpweb]

Too bad...
After almost an hour of flawless performance with a 320 MB file in CS2, I was ready to pour a huge bucket of praise over Ur head, Dr, TZ... But, no, all of a sudden the screen froze and I was prompted to restart the computer. This with one of my ATI Radeon Pro cards instead of the OEM nVidia GeForce TI...

Prior to that another curious thing happen'd. Replaced the 2 OEM Ram sticks with the four 512 GB dimms. Here 'System' only recognized 1.5 GB's... The stick in slot # 0 didn't read. Removed it, carefully put it back. No change. Took it out again to place one of the 256 MB OEM's... 'System' now read
1.75 GB's, aha...

Too bad I didn't pay attention to the Ram readings before. Though, if I remember correctly, during the Hardware tests all 4 dimms got a 'green' rating.
Don't know whether this is significant or just a freak.

Anyhow, Dottore, let me repeat why I believe that HD's or sw are not the cause of the troubles... Everyday single HD in the unit is a WD 120 GB, with only about 40 GB's of files on it. That one runs Panther with nothing but CS2. When things started to go South, I removed said HD, installed the Seritek card + the Raptor, with 10.2.8, CS & various other sw's... Got 'system failure' just the same. Replaced that setup with the 160 GB OEM drive - OS 9.2, PhShp 7 & various other sw's... Again, trouble in paradise. Repeated the experience with a FW ext. HD, again running OS 9... No improvement.

Before all these issues arose, I once experienced a 'Kernel panic' which led me to a clean install on the 120 GB WD... I did post the experience here, back in Dec. 05, if I am not mistaken.

One more idea... Could it be that Apple somehow 'torpedoes' the use of a Dell LCD..? In my case a 20.1 wide model. My other 2 working QS units; each have the sgi lcd connected to Formac PCI cards.

Last, not least, I'd be glad to perform a 'benchmark test' with the Raptor in the QS with the 1.33MHz GigBump, OS 10.2.8. (Can't go any higher because of the sgi lcd...) Is this something that can be done internally..? If so, how..?
Last year I performed a crude test with a timer. Can't find the posting, but the Raptor opened a large file in CS at least 3 to 4 times faster than a WD 120 MB SATA HD.

Let me know how I should proceed, please.

Over & out.

[wolfpweb]

Forgot this... My slight suspicion as to the Dell lcd results from the fact that 'System' does not recognize the (albeit) working monitor connected to the AGP.

Hm...

[wolfpweb]

And the saga continues...
This morning the MDD ran for some 3 hours without a freeze or system failure. 4 crashes, though, one in CS2 the others in SAFARI... Worked for more than an hour in the Shop, large file. Kept 'saving' every once in awhile. However, I then noticed a slowing down, as if some part (or parts) were getting tired.

Now, that's an improvement, whatever may have caused it; ATI card instead of nVidia, different Ram stick configuration. or both..? But not a situation I'd wanna live with. Got to report 3 crashes to Apple. On the 4th I got stalled half way into my 'problem description'... A color wheel came up that kept spinning & spinning. Force quit didn't get me any further. Disgusted, I shut off the juice.

Am back on one of the QS'. Will Apple reply, by the way..? So far not even an automated acknowledgement of the crash reports. So much time wasted.

[wolfpweb]

Sorry for boring the community. But once in awhile one gets real good advice here, therefore I dare to continue what has turned into a monologue...

Actually got a bit of work done yesterday, saving as often as might be advised under the circumstances. Today, however, after only 'bout 1/2 of work in CS2, another system failure occured. 00000001 (corrupt stack), whatever that may mean. Found a thread on Apple's discussions, involving Nathan Terry & his DP 1.8 G5, with an identical problem. The discussion, unfortunately, ended inconclusive.

Called a few service shops within a 50 mile radius, some of which shouldn't even be on Apple's list, by their absence of any Mac technicians, and had to learn that, no matter what, they'd have to keep the unit for about a week.
Before I bring it to one of them, I guess I'm gonna have to deal with A.C.
Called once, got frustrated after more than 10 minutes waiting...

To be contd.

[unclemac]

.....Just nothing very constuctive to add.

I would agree that heat may be the cause, but realistcally most Macs don't have heat problems, at least not that cause stability issues. It happens, but not that common in my experience. Failing fan, or bad sensor keeping it from kicking into high mode when it should?

No, Apple does not reply to any crash reports. If we are lucky they go in a data base somewhere and someone looks at them sometime.

Have used a multitude of monitors, and never a problem, at least not in modern times (post OS 9), so I would say the Dell is not an issue.

I would hash it out with Apple Care. Let 'em replace the logic board, power supply, whatever they want. Do take it back down to the stock config first or they may blame it on after market RAM, PCI cards, etc. And be prepared for them to reinstall the OS. When I take a machine in for service, I always tell 'em that they are free to wipe the drive and reinstall the OS - cause I already did three times and it didn't help (which is usually true before I give up and pay for help).

[wolfpweb]

Righto, "Uncle"...

The Dell was only a hypothetical far shot, cause it is not recognized by 'SYSTEM', though it works... The Ram stick which doesn't register in one slot, does just fine in another. Which really seems to indicate a troubled mobo.

Shoot happens.

[kaye]

The details are getting sketchy now but when my MDD died and before going to AppleCare the Gurus here helped me a ton in trying to test everything that we could but no luck. So finally I decided to get all of my data off of my main boot drive and purchased a Diskology, an older model than what is available now. Got mine from Granite. I did this because I did not want to risk my QS in any way, at the time my backup machine.

After backing up the drive and reinstalling it, but in the ATA66 drive cage, I stripped out all of the PCI cards and drives except the OEM drive which was a backup boot drive moving it into the ATA100 drive cage. I left the OEM RAM stick and the remaining RAM installed. Reinstalled the GF4ti which came with the MDD as an option.

Took the MDD into the local Apple Store and was directed to the Genius Bar. I found out it helps to call ahead and get an appointment and the store was jammed with near Christmas shoppers. So I checked in and waited. When a Genius called my name, the MDD went on the counter and he opened it up, removed the added RAM, did a visual inspection, and closed it up. Connected his monitor and power source and tried to start up. He then went through a number of tests but I could not see what was going on so I strayed away from the Bar. After about 30 minutes, he told me my mobo was bad.

So we did all of the paperwork and he said they would call when ready. This was in the afternoon. Next morning I get a call that it is ready. They did replace the mobo and zero cost. Ran it for awhile and decided to GigaBump it. k

[TZ]

Monologues get read and can be your own "note taking" note pad to help organize what you try.

Some thoughts I've had over the last couple days: OS X will look and load fonts from Classic/OS 9 system, and can be trouble. Printer drivers.

A similar report to yours regarding memory was posted on MacIntouch's "Bad RAM" report:

Experiences - MIchael Krayewsky

I reported a few days ago about numerous Kernel panics and sleep problems I was having with my Quicksilver 2001 G4 upgraded with the PowerLogiz Dual G4 1.6 GHz card. Glad to report that the problem has been solved.

Its been interesting the last couple days. I re-loaded Tiger and upgraded to 10.4.3 on Saturday. Sunday night, I re-started to install an old Canon printer driver, and my Quicksilver G4 Mac with the PowerLogix Dual 1.6 GHz froze on restart!

My frustration continued as I've been having sleep, kernel panics, and loss of the 2nd cpu over the last week or so, that I let it go and went to sleep.

On Monday, I did some surfing on sleep and kernel panics and Tiger 10.4.3 and 10.4.4 to help diagnose my problem. When I got home I tried to start, via option key, from my external firewire drive and my Mac froze. Then the bell went off that perhaps its a RAM problem (based on my surfing earlier in the day).

So, when I forced a power shutdown, I removed one of my three 512MB RAM modules. Upon restart, I got further, and again froze. So I removed a second RAM module and my system started AOK, from either the Firewire drive and my hard-disk inside the Mac. I replaced each of the suspect 512MB RAM modules and was unable to start when either was installed.

I then reloaded Tiger again and upgraded to 10.4.3 as I suspect the bad memory corrupted some of my disk files.

I have been running with no problems. I suspected my problems all along were memory RAM related, but didn't want to believe it as the symptoms were somewhat consistent.

Neither the PowerLogix Dual G4 1.6 GHz card nor Apple's Mac OS X 10.4.x was the problem. For those having similar problems, please focus your troubleshooting efforts on your RAM. It certainly was mine.

All of the memory in my MDD, OEM was Micron and all the additional 1.5GB was identical only Crucial.

[wolfpweb]

Thanks again, Kaye & Dr. TZ...

Ram was one of my original suspicions, as U could've deducted from the way I went about my trials & errors. After testing several HD's with only the 2 x 256MHz OEM's installed, I still got crashes & freeze-ups. The best, almost (!) trouble free run came after the graphics card switcheroo. Indeed, the ATI Pro doesn't get as hot as the nVidia with its hi-pitched fan. Haven't had the time to pursue with Apple. Nearest store is 30 miles away, in a mall, at least 300 yards from the nearest parking. Called there, no store (phone) pickup, got switched to Apple main (where I hung up after 10 min's or so)... Am I glad to have 2 workhorse QS'..! I'm even more glad to have a Seritek in 1 of the 2, GigaBumped to 1.33MHz, with a fast spinning Raptor.

I still mean to follow up on Ur benchmark test, Dr. TZ... Just give me something to go by.

To be contd.

[TZ]

Just because RAM is OEM doesn't mean it is immune or perfect. And if someone bought a used system....

The more RAM is handled, while rare, the more chances for mischief.

Over the last (6 yrs?) I have learned to focus on getting RAM as an issue as #1 first, then move on to other potential 'issues' even if it means investing in one good 512MB module to do so.

At one point it sure sounded (and I read messages 2-3x if necessary and even copy to textedit to look through) where I got the impression you had a Raptor but rather than running off SeriTek, used on native IDE bus using a SATA <-> PATA bridge. Could be you don't or didn't, but looked like just another thing to 'test.'

When I get to the point with a clean, updated, stable system, one thing I do is create a 8GB disk image of the system. And insure that it can be restored to a partition and work.

In the past, some printer drivers (HP, Canon, Lexmark inkjet) have been known to cause problems. Same with fonts, and because you seemed to have trouble in 9 and X...

[wolfpweb]

I regret, Dr. TZ, if I haven't been clear enough in regard to HD's...
In fact, the Raptor (with 10.2.8) runs flawlessly in the QS with the 1.33MHz and a Seritek card. One of my tests of the faulty MDD was to install & run card + Raptor, to eliminate chances of any IDE connect problems. Troubles persisted with only the Raptor running. Once got a freeze even during startup.

Anyhow, today I got my slides back. As I am a pro, U could've bet that my screen shots of 2 different 'system failures' would show sharp & clear. They do. One shows said 00000001 (corrupt stack), the other: 00000008 (unknown failure code)... I'll print these out, something to show and to go by...

I once had to replace a mobo in a QS, remember... Early troubles were similar to the ones I am experiencing now. I'd say, chances are about 70% bad mobo, 30% bad Ram... I could be wrong. Only fools are never.

To be contd.

[wolfpweb]

Update... As I was cleaning & sharpening the respective files, I noticed that failure 00000001 is mentioned in connection with CPU 1, while 00000008 gets the honor with CPU 0 (..?)...
Are we talking processor problems here..?

I'll soon find out.