Long time user stumped

It might not be a bad idea to go back to 350. While it isn't the best, it may help. Your cpu may have an intermittent issue due to the speed. I've seen some weirdness on the blue and whites, but nothing compared to what I've seen from the MDD G4's. If need be, I have a spare 350 chip around the house. I have a pair of B&W g3 350's at home, one that I use occasionally, and one that's halfway parted out.

- iantm

I ran my b&w 350 at 400 MHz for months, without a single crash.
Then all of a sudden I was getting crashes every time I tried to do 3D work, then it spread to crashes in other programs.

Once I tried clocking down to 366, everything was fine. I'd tried cleaning the insides out with compressed air, too, thinking it might have been a thermal problem, but no. It just didn't want to run at 400 any more.

YMMV, of course, but iantm's idea is a good one.

Can I have it?

hi there!

i had a problem with a g3 b/w too, a few days ago and got great help from Hawaii Cruser, see http://www.applefritter.com/node/8872

in my case it was a faulty firewire module... maybe worth to check?

First off, your kernel panicking is almost certainly due to bad RAM*.

(* Or, to be more accurate, RAM your machine doesn't like. I've seen *waaay* too many instances of Macs hating memory that works perfectly fine in other machines to say "Bad RAM" without an asterisk. Ram tests for Macs almost never find the problem in my experience. It's spooky, voodoo magic, and it's one of the things that makes me guffaw every time I hear someone talk about Apple having "superior engineering" compared to the rest of the PC industry.

/rant off. Anyway.)

Secondly, what you're describing happened to me, just about exactly. My B&W has 544MB of RAM: One 256, two 128, one 32MB DIMMs. Some months ago I tried swapping two (known good in every other computer they've been in) 256MB DIMMs for one of the 128s and the 32, giving me 896mb. (Interesting coincedence, actually.) Within an hour or so the machine kernel panicked. However, since I'd *also* just put a second video card and a DVD burner in the machine earlier that day, I was stupid enough to *not* rip the new RAM out straight away. (The justification being that I thought the kernel panic might have something to do with the other hardware I'd added.) Shortly thereafter I was *monumentally* stupid, and hit *install* when after the kernel-panic reboot the "Software Update" window came up and told me there was an update available. Machine kernel panicked in the middle of the install, I rebooted, and... poof.

That crash in the middle of software update did *something* to the file system/partition table/karma of the hard disk which rendered OS X unbootable on the machine, even from an install CD. (The machine would boot, then kernel panic, presumably when examining *something* related to the hard disk.) I ended up installing another hard disk (there wasn't anything I particularly cared about on the other one), leaving the screwed-up one in the box disconnected, and the machine worked perfectly again. (With it's original 544MB of RAM, of course.) Connecting the cursed hard disk as a slave, however, caused the machine to kernel panic at boot again. It would boot into native OS 9 with the disk connected, so I used that to copy over what little I cared about from the old disk.

Just recently I took the cursed hard disk out of the machine, connected it to a Linux PC, and ran:

dd if=/dev/zero of=/dev/hdc bs=512 count=1000000

Which overwrites the first 1/2 gig of the hard disk with zeros, thus totally obliterating the partition table and any system filesystem information. After which said drive successfully took a "Tiger" installation and is working fine today. So what happened? Hell if I know. Apparently kernel panicking during "Software Update" is a *really* bad thing. I'd love to know exactly *what* it does to the file system which can render the machine unbootable even when booting from another disk, but... anyway.

From my experience all I can suggest is you throw your RAM away, and then use a PC to *blank* your hard disk before putting it back in the Mac. And I mean *blank*. (Just running an OSes repartitioning tool doesn't *really* wipe it.) Use "dd" under Linux or BSD, or see if the drive manufacturer puts a "blank drive" option on their diagnostics tools disk.

Of course, as to why your machine is croaking under OS 9, well...

A: Is your machine a Rev. A or Rev. B, and do you have both hard disks connected now? If it's a Rev. A perhaps the old famous B&W corruption-with-a-slave problem bit you during the install.

B: It could be that firewire module problem. (Firewire is working grand on my system, and it doesn't have an OS boot problem, so I suspect your OS 9 issues are unrelated to what happened to your X installation.)

C: Whatever problem triggered the OS X meltdown (be it bad RAM or flakyness from your overclocked CPU) corrupted your OS 9 install. Bad RAM often manifests itself in the form of random filesystem corruption.

--Peace

These posts, while very helpful and educational, make me love Classic and hate X more and more. It is almost as if Apple is attempting to accelerate planned obsolescence at an unnatural rate by putting 'old-system killers' in their updates (and no, I am not wearing my tin-foil hat here), forcing the unwitting and loyal Mac user to have to upgrade their system to a whole new box or live without the upgrade. I had a similar problem with X.2.8 on a Rev A iMac and had to pull X entirely - a kernel panic caused me to loose 75% of what I had on that machine when I tried to use Software Update to apply Apple updates to a stable X.2.8 install.

I don't think Apple *intends* to "kill" older systems. However, what I will say is OS X is doing a wonderful job of exposing how weak the engineering behind all the "transitional-era" machines really was. (Meaning basically all the Beige G3-derived systems, which includes the tray-loading iMacs, the B&W, original iBook, and Wallstreet/Lombard.)

Modern OSes expect *reliable* hardware. (That's why traditionally UNIX workstations often included fault-detection features like parity or ECC RAM.) It's pretty obvious that either due to faulty engineering or poor quality control those machines left over from the Beige era just arn't fundimentally reliable. It didn't show under the Classic OS because it's so fundimentally glitchy anyway. (I've been amazed how tolerant old-school Mac users are of having their computers crash. At least they generally didn't eat themselves when they did it.)

Feh.

--Peace

Well, I solved my problem I guess... I got a Mac Mini. As of now, I'm not really sure if its an upgrade or downgrade. My Powermac had 160gb HD and 896mb or RAM and the mini has a 40gb HD and 256mb of RAM. I really don't like the mini at all, but I don't have enough for anything else. I have never had any less than a Powermac, so this is quite differnt from what I am used to. I'll still try and fix my G3 though, I can't picture myself getting rid of it. Oh well, I'll live. Thanks for all your help guys. I know I can always count on the people of Applefritter when I need help. I posted this same question on Spymac right after I did here, and didn't get a single reply.