On an MDD G4, dual 1GHz (from 2002), I tested a (probably fried) Firewire drive a few days ago. The drive didn't show up there either, but to top it off, I got the first kernel panic I've seen since the MDD's motherboard was exchanged for a new one under AppleCare, due to incessant and intermittent problems at/after shut-down (don't beg me to bore you with the details).

Without a problem since then, until now, flying along under Mix Plus, Pro Tools, Logic etc. G4 presently at 10.3.9. 1.75GB RAM. A couple of ATA's. Etc. What happens is this: I shut down, go to sleep, remember I own a computer in the morning, switch it on -> kernel panic. Wiggle scratchily past that sooner or later into a successful OS X boot (via OF, or a few times safe-boot which doesn't work each time, repeated PRAM-zapping, lots of etc).

A few times it refused to even find OS X, and started from 9.2.2, upon which it was sooner or later possible (also after false starts) to choose the desired OS X drive and boot into it. Somewhere along there I reset the PMU according to Apple's instructions. Here's the weird part: once it's correctly booted, everything works without exception (well, except the normal healthy behavior of being able to be shut down completely and then restarted using the power button). So by everything I mean everything. PRAM-zapping kills the time, which I noticed (in OF for instance), ie it showed 1904 or something. I would think a faulty PRAM battery would not be able to retain the time, but if I leave the PRAM-zapping alone and concentrate on getting back up and running via Open Firmware, the time is intact after an overnight rest. Should exclude that battery, but who knows. Maybe the PMU got scrambled - there are anecdotal notions to that effect here and there, both on macosxhints and elsewhere.

Anyone have a clue? I forgot to mention I used all the software easily available (Disk Utility, the correct AHT, TechTool Pro, memtet/rember, DiskWarrior 3.0.3) at various points without uncovering anything of interest. I'd love to not have to reformat and install it all, because it really is lots. Donating $150/SEK equivalent to our local authorized guy to double-click MacTest Pro, or Apple Service Diagnostics seems a drag too. Buying those apps is out of the question. Anyone have any truly scintillating suggestions? I'd love to hear them!
exdc

This Apple KB article should enlighten you on what the PRAM reset does:
http://docs.info.apple.com/article.html?artnum=86194
Part of it is that it resets time.

post-Shut Down blues: RAM RAM HAIRY RAM?
 

Further to the introitus of this thread: after removing hardware (Digi Mix+, FW-drives), though (lazily) retaining non-original RAM, and after having reformatted drives and reinstalling Panther in various guises (3.0 3.4 3.9), and after having run AHT in loop-mode (ctrl-L) for 12 loops finding nothing in particular,
the bottom line is: after Shut Down, the G4 will start in the expected manner IF the power button is hit within "a little while", whatever that is.

So far it starts fine after a Shut Down and a pause of 15 minutes, and after a half hour - haven't yet tried other periods, except overnight, which decidely does NOT work - then I get the kind of kernel panic where text is printed from the top to the bottom of the screen (and the whole rigmarole to get it up and running again, which has so far been successful, was sketched in the opening message here).

So for this computer-tech non-literate, that "feels" like "bad-battery-like behavior" - what would I know. The on-board battery, when removed, shows 3.65V, but I don't have what might be needed to test its actual charge as well - perhaps it gets uncharged entirely overnight? Perhaps this is a wrong idea, since a verbose enough kernel panic shows correct time, and time would be screwed if PRAM goes south. I assume that there is yet physical RAM-testing to do.

Rember/memtest smiles upon this RAM (all 1.75GB of it), as does AHT, though I am told MacTest Pro and Apple Service Diagnostics can detect faulty RAM the others can't, as well as faulty RAM slots, though that seems unlikely. So if RAM has gone bad, does that mean checking every slab in every slot and maybe even every combination (yuck!)? Anyone have a better idea?
exdc

My guess is that the bad drive somehow interfered with your boot drive and now you need to run Disk Utility (boot from installer disk & select it from Menu) or Disk Warrior on it. You could also leave it on overnight so the system can run its maintenance utilities.

Ignore the voltage of the battery. Without a load, it's not an indicator of the batteries strength. Look at the date on the battery. If it's near 5 years or older; Replace it!

problem solved? las vegas jackpot?
 

Las Vegas,
Thank you for your incisive logic. For $15 I replaced the motherboard battery - couldn't find a SAFT LS14250C, but I used a 3.6V (1/2AA) equivalent that an ASP in Stockholm had in stock. Smacked that battery into place, powered up (after an overnight rest and no battery for a few hours) and it started correctly immediately, for the first time since the problem showed up. I will put in a final word about this in a couple days if it remains stable (or otherwise): chances are, this battery was the problem, and if so (remember: they're not rechargeables) most people with MDD's from 2002 and thereabouts will run into this same issue right about now. Pretty easy fix if it is.
Thanks again!
exdc

motherboard battery was not a fix
 

About sizes it up.
Now I'll be removing and repositioning memory slabs (3 x 512 plus 1 x 256 = 1.75GB) - could be one of them went south. Anyone know an easy way to be systematic about this?
exdc

You can check the RAM with 'memtest':
http://www.memtestosx.org/

how long should memtest go?
 

hello hayne-
thank you for the tip. i tried memtest, but couldn't see that it ever arrived at a conclusion. does it actually stop at some point and provide a report, or does it keep on cycling until stopped?
exdc

It keeps cycling until it finds a error. Let it run overnight. If it finds no problems, your memory is fine.

memtest running
 

am presently running it in single user mode, 5 iterations, looks to take about an hour or so each - more tomorrow. quite a few combinations with 4 slabs and 4 slots, if memtest is to be run with its iterations for each physical RAM configuration...

5 iterations of memtest ok, still k-panics
 

All night did it, 5 iterations of memtest from single-user mode, passed fine. :cool: Should probably move some RAM around and do it again, but I got optimistic (read: lazy:o). My mistake. Recall there's a new pram battery in there (wonder if it does anything else...). Started fine from a several hour shutdown this morning. But, after the second several-hour long shutdown, it didn't start fine. Not the 4 language kp, but the screen filling up with "stuff" (i don't know programming!), however I just let the drives keep on grinding, and after a while booting broke into the blue screen, programming salad still being generated at the top of the screen, but once the Finder showed up the salad stopped. Machine appears to work fine in every respect, once booted, again. Except for this. What can it be?! A complete Shut Down, and a pause for say an hour, and the machine starts perfectly, works perfectly once booted. Shut Down again, wait say 4 hours, and then this kernel panic salad again...:confused:
exdc

Can you boot off another drive? I'm sticking with my initial guess that it's something on your hard drive.

Two bad drives at once? Very different drives, different ages, different sizes, different sets of software (10.3.0 vs 10.3.9). Bad in the sense that this weirdness can happen as described in disgusting detail above? Doesn't make sense. Yes I have about 8 drives I can boot from, same problem with two of them (haven't tested them all). But the reformatting and reinstallation of system software etc is being done on the internal drives only.

Not a second bad drive, but one that got corrupted by the bad drive. It's worth a shot, especially if you can easily boot off another.

Booting from 4 different drives, connected in various permutations, has elicited the same problem.:D

Will you copy down as much of the scrolling kernel panic text as you can and let us see it? Or perhaps if you have a digital camera you can upload a photograph.


Trevor

scrolling text
 

Trevor:
Not easy. Two problems - the panicky behavior is erratic, to say the least. The scrolling kp text changes from kp to kp. And the entire process changes, so sometimes it keeps this up, and I force Power Down, until I break down and use Single-User mode, or an OF-boot, and finally can get booted. But tonight, as erratic as ever, 4 forced reboots is what it took to get completely up and running. Boot 1 resulted in the 4-language kernel panic blurb. Boot 2 resulted in a frozen gray screen, with that little thingie that whirls around below the darker gray Apple. Boot 3 resulted in a scrolling kp text. Boot 4 went all the way to complete boot, OS X 10.3.9. Then restarted using a 10.3 drive (also internal ATA), resulted in everything coming up as expected except for the frozen USB bus... Force Shut Down, then reboot from same drive, and now it all works. Etc. Have included a jpg of the text, a jpg under 100kb is not very informative... Thanks for having a look. On getting further kp's (and images of same): the weird part is the pause required. So far it looks like less than 3 hours is not enough, i.e. if I wait a little over four hours then it kp's out on me; if less, it doesn't... Go figure.
exdc
__________________
"He that drinketh strong beer and goes to bed right mellow, might well wake up on sheets and sheets that all are colored yellow." - Anonymous

update to kp-weirdness on mdd g4
 

:eek:
Our local Authorized Service Provider is presently screwing the G4's head off and on again. I was delighted (jaded sense of h?) to hear he had the same trouble as I, ie kernel panics on start after at least 4 hours of Shut Down, but not after shorter periods. He was delighted to find a problem that stumped MacTest Pro, which he ran for about 18 hours with only this-dude-is-healthy to show for it. On a hunch (separates the artists from the technicians?) he put in a new graphics card - same as the original, whatever that was (dual 1GHz MDD G4 from late 2002) and it started nicely, both at once and after a 5 hour Shut Down. A new test is planned for tomorrow morning - stay tuned... Problem is, if he gets it to work a number of times, but I mean, he only gets a few days on this, by me, and yet there's no real diagnosis, just a hunch that yields a fix-for-now - I dunno, would feel better to have a diagnosis... Any comments? Btw: drives have been repeatedly formatted, gotten their OS X installs, one of them (the night before I finally relented) even got as far as Pro Tools 6.0, then 6.4.1cs something or other, and the G4 only died when I decided to screw around with (configure) the Digidesign hardware. Oh - it passed all DigiTest-ing before (and after, on another G4) the aborted attempt. Never did get as far as Logic Pro... Curiouser and curiouser...
:confused:

Nice sig :-)

At this point, I'd go with the tech's hunch and replace the video card.

The odd cool-down period required to see this kernel panic makes me think the problem might be a hairline crack in an electrical trace. The metal expands when heated and makes a good-enough connection. When cool, the metal contracts and the connection becomes not-good-enough.

This problem (I'm guessing) could be in your motherboard, or could be in one of your expansion cards. If swapping the video card makes it go away, then that's a good indication that the problem has been found.

Trevor

hairline trace crack?
 

possibly a good-enough explanation - always the detective i'd love to pinpoint that trace though... the techie/artist called this morning; a new boot from overnight shutdown worked fine-ish. he did end up having one problem, typical for the glut of problems this g4 has generated these weeks: after the first start this morning, the screen slept, and he couldn't wake it. usb freeze? who knows - there have been usb freezes among the rest of it. sounds more central than the graphics card. what he had done is put in another AGP-graphics card (not identical with the original). he wants me to take it home and run it with the digidesign cards as a test. thanks for the suggestion!
exdc

update on mdd g4 cold start weirdness
 

The G4's still at our brave local ASP. Not having a ready supply of logic boards has erected a hurdle, since so far everything but switching a new logic board (or a new dual processor) has been tried with no success, and - maddeningly - MacTest Pro hurtles along, insistently returning a clean bill of health. Apple apparently hits them up for a re-stocking fee if they need to return a new logic board they ordered but turn out not to need, but the ASP is going to request a special dispensation today. The drives and the pci cards (Digidesign Mix Plus) have been throughly tested by them in another MDD with absolutely no problems. Still at square one. Oh btw: a different video card made no difference with regard to the weirdness described at the beginning of this thread.

g4 blues
 

just gets worser and worser.
picked it up, third time running, this afternoon.
had last week neglected to provide the technician at the asp my logic and protools keys, we figured what the hell, he had some diagnoses (see below) so while he ran digidesign's digitest on the mix cards with no problems (transmission problem of course, but expected, since no interfaces), he never actually could start and check either logic pro or protools to see what happens when they access the cards in those two pci-slots. anyhow, once i got it back home, it took about ten minutes to get the first kernel panic. to summarize: both logic pro and pro tools elicit kernel panics. nothing else appears to be broken. digidesign insists the cards have passed their diagnostic software (digitest) and are thus completely clean. they work fine in another machine. what remains? is this a pci slot problem? i'm pretty soon into this for more money that i care to divulge, and it's still broken. isn't there some kind of system whereby an authentic asp can get the kind of support required to figure even this kind of sh_t out? His MacTest Pro ($1400 USD) showed no hardware faults after many many many iterations. Still the machine kernel panicked. Only when he put in a new dual processor and a new graphics card was it possible to run more or less everything - notable exception of the two apps i got the machine for (logic pro and pro tools). does anybody have any new ideas?
exdc

final installment hehe
 

OK, here we go. What fixed it? Yes, it's fixed. How - "don't ask me"? OK, that's uncool, considering all the great ideas that showed up here. The G4 has been put so thoroughly through its paces now, and the results tonight were so spectacular, that the bottom line must be: the technician fixed it. And why did the Digi gear still not want to play along? I suspect the answer is that formatting with complete zeroing, reinstallation of system software, on this particular drive for the fourth time, reinstallation of Pro Tools, both on the OS 9 partition and the OS X partition (of that more thoroughly expunged drive), updating of all relevant items (system to 10.3.4 = highest Digi-approved; PT to 6.4.1 = highest Digi produced for the Mix cards; Logic to 7.1.1). Before a couple of the updates on the OS X side (including Pro App Support 3.1), the OS 9 (9.2.2 of the system) version of PT accessed the Mix cards perfectly, while the OS X version still caused kernel panics. Now after a few more updates these are all gone. Oddly, these updates had been performed on every re-installation detailed above. Go figure. The take-home-message is: enough is not enough, maybe one more time is...
exdc
PS - if it's broken again tomorrow, you'll read about it here...
:eek: :) :rolleyes: :cool: :p ;) :D

That's quite a story. Mine is eerily similar, right down to the hardware part-diagnosis but without the expert-techie involvement. I'm intrigued - if you're still around - whether you're still using the G4 now, the OS version you settled on, etc or whether the whole kit met an infortunate accident which mine is definitely begging for... I'm currently back to 10.3.6, the most stable Panther release I can use. 10.3.9 was the start of my problems and - probably - the best way to fix them (more recent updates, etc) - or did you find peace with 10.4? This is where I stand now: spot for 10.4 or cut my (considerable) losses.

Or maybe it's just the PRAM battery...

IME, I have had similiar problems with a loose processor card. Simply reseating it fixes the issue, maybe that is all that the technician did.