Submit Hint Search The Forums LinksStatsPollsHeadlinesRSS
14,000 hints and counting!

A possible fix for a highly unstable G3 or G4 Tower System
Do you have a G3 or G4 tower that acts "flakey"? Does it crash a lot? Problems formatting your hard drive? Does it just boot into open firmware with some strange error, like "Default Catch!"? Well, then you may have a bad IDE bus. Don't throw out that tower or buy a new logic board just yet...

The IDE bus that the hard drive is plugged into by default can get "worn out" on a computer that is in use a lot of the time (work horse). I don't know if it's something to do with solder or what. The fix is to put your hard drive on the other IDE bus (the one the optical drive plugs into). You will have to do a master/slave setup to use both drives. I've also found that it can be very convenient, but a little tricky, to put the hard drive into the zip drive bay (if it's empty of course).

I've seen this hardware problem on three computers in the last three years -- a Blue & White G3, Graphite G4, & just recently on a QuickSilver G4. So I can't even imagine how many towers are out there with this issue. I have one of these "fixed" G3 Blue & White towers in my home today, and it has worked just fine for two years after this fix.

[robg adds: I obviously can't test or confirm this one through personal experience, but I thought it was intriguing enough to post as a possible solution for what might at first seem to be OS X issues.]
  • Currently 1.00 / 5
  • 1
  • 2
  • 3
  • 4
  • 5
  (1 vote cast)

A possible fix for a highly unstable G3 or G4 Tower | 27 comments | Create New Account
Click here to return to the 'A possible fix for a highly unstable G3 or G4 Tower' hint
The following comments are owned by whoever posted them. This site is not responsible for what they say.
A possible fix for a highly unstable G3 or G4 Tower
Authored by: genericuser on Sep 07, '05 09:40:19AM


could this also have something to do with the 'Default Catch!' error i keep gettin gwhen using XPostFacto?

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: bedouin on Sep 07, '05 10:33:19AM

Probably not. That's a pretty general kernel panic error just telling you something bad has happened.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: Gerk on Sep 07, '05 11:13:15AM

actually it's not a kernel panic (there is no kernel loaded yet). It's a firmware error trying to load the elf executable (bootloader) during quiesce.

It's kind of hit and miss with this sort of thing, if XPostFacto is triggering it all I can say is to try again. If it happens every time there is a problem. If it happens occasionally you'll probably just have to retry and live with it. Debugging stuff in OF is not fun at all :(

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: xSmurf on Sep 07, '05 12:51:30PM

I have experience similar errors while installing NetBSD on a 6500 (OpenFirmware 1.0.5). As explained earlier it happens when the firmware can't load the bootloader. This could be because it the drive hasn't started yet, the floppy is messed up, etc. One thing you have to remember is that OldWorld machine don't have much support for the Open Firmware (it was never inteded to do anything else than boot MacOS, if it did that, there was no need to debug any further). It can also occure if the real-base and load-base variable aren't set properly. This said, the default cache error can also occure "randomly". Indeed, on the 6500 if I tried the command that gave the error a few times it would eventually boot. The NetBSD documentation has some good information about this common problem. I would strongly suggest that you give it a look @

Free iPods, now in Canada to! Get yours :

PM G4 DP 800 / 1.25gb / 120Gb+80Gb / CD/DVD±RW/RAM/DL
- The only APP Smurf

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: xSmurf on Sep 07, '05 12:56:54PM

May I also add that from what I've heard xPostFacto will not work on every machine. Or at least OS X won't. I have a friend who was trying to get it to run on a 7500 with a CPU upgrade to no avail (OF & Kernel errors). He even tried a different board and tried swaping most components. No go. He ended up buying a 9500 motherboard. So beware, even if xPostFacto allows you to install OS X, it doesn't mean it's gonna run properly.

Free iPods, now in Canada to! Get yours :

PM G4 DP 800 / 1.25gb / 120Gb+80Gb / CD/DVD±RW/RAM/DL
- The only APP Smurf

[ Reply to This | # ]
Authored by: bedouin on Sep 07, '05 09:49:12AM

How does an IDE bus get 'worn' out through normal use?

I've always had mysterious corruption issues when copying large files to and from my slave drive, but was never able to pinpoint the issue. Most recently I had a kernel panic when using Carbon Copy Cloner to clone an OS X install to a SATA drive and adapter I had installed; the panic would happen each time while copying the Library folder, and could be repeated by using the ditto command from Terminal.

I've never had startup issues though. This is a 2002 Quicksilver and has been a bit strange in this regard for the three years I've owned it. I don't want to overstate the problem though; there's been a few isolated strange incidents, but nothing to put my finger on with certainty.

[ Reply to This | # ]
Authored by: bdog on Sep 07, '05 10:26:22PM

I really wish I knew how the bus gets worn out. Or at least what triggers the 'break' (start of flakiness).

Your issues sound like this problem.
If your Quicksilver has had issues, try swapping busses! See if those strange problems go away.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: Shiver758 on Sep 07, '05 10:22:42AM

This makes PERFECT SENSE to me.

I was having odd startup errors, impossible installs, mysterious data corruption..

then I swapped busses (this is on my dual 533 digital audio G4 tower).. I don't even know what made me do it.. Just "old geek" instinct I guess, it was the only thing I HADN'T tried.

Magically it all worked (ok, a week later the power supply tanked, but that's hardly a related issue).

I concur.

[ Reply to This | # ]
'worn out' might not be best word choice
Authored by: Gerk on Sep 07, '05 11:18:02AM

but the solution seems valid enough. One thing with doing this though, is that if you're using things like XPostFacto (as someone mentioned above) or other bootloaders you may run into a couple of weird things.

In open firmware there are what they call devaliases ... (issue the command devalias at OF prompt to see them). By default it gives you some short form names for devices, like hd, ide0, ide1, ultra0, ultra1, cd, zip, etc ... so changing to this suggested configuration can do things like make the devalias to no longer be hd ... but likely it would be zip instead :)

Not a big deal, but if you're trying to do things in OF or with bootloaders that are looking for this sort of naming it's good to know ahead of time. If you're really adventurous you can 'fix' the devaliases in OF yourself. If you're lazy like me, you can just figure out which one really points to where your hard drive is and use that alias as required :)


[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: Forced Perfect on Sep 07, '05 11:41:49AM

One think to consider is that the second bus on most older Macs is MUCH slower than the bus for your hard drives. For example, in my G4 AGP Graphics (Sawtooth) my 80 GB Maxtor gets 30-40 MB/second or so on the normal HDD bus and only 8-11 on the "optical drive" bus since it isn't an ATA/66 interface like the HDD connector.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: koehn on Sep 07, '05 12:27:48PM

I ran into this problem when I added another drive to the other IDE bus, and worked around it by pulling my PCI video board and using the AGP one that came with my machine. Not a big deal since it's a server, but it was a bizarre problem.

[ Reply to This | # ]
Using the "zip" bay can cause a heat-related problem.
Authored by: Panjandrum on Sep 07, '05 04:33:37PM
I've been running a 120gb drive in my 3.5" "zip" bay for years now (this is in an older G4 Sawtooth). Since it is an ATA/33 bus instead of ATA/66, I use it for things like my MP3 collection, where speed isn't an issue. However, I strongly advise that you ADD COOLING to your tower when you do this. That drive bay is extremely enclosed, and running a hard-drive in there is obviously not something Apple really intended. A Zip drive doesn't create much heat, but a hard drive does. After installing the drive I noticed that disk in my optical drive were heating up much more than they previously did. I opened the system and I could barely touch the metal frame around the 3.5" drive it was so hot. I went to a local computer store and bought three small fans and attached them, blowing INWARD across the drive. I bought fans that each had a power "y" adaptor on them and ran their power cables up beside the drive bay (not easy). I just removed the plastic plate from the front of the tower and used a thick gel-type superglue to attach the fans in place. They don't interfere with the optical drive bay door, but it is a close-call so test before you permanently attach them. The end result is a drive that runs cool and doesn't cause any overheating of the optical drive above it. Here is a link to a quick snapshot of the fans:

David Butler

[ Reply to This | # ]

A possible fix for a highly unstable G3 or G4 Tower
Authored by: sgi_oh_too on Sep 07, '05 10:20:28PM

I believe this to be complete and utter nonsense. A bus controller could fail to operate properly due to defects or improper environmental conditions (e.g. heat, humidity, etc), but I seriously doubt that it could have "worn out". Integrated circuits are generally very robust. I have experience with many many servers and raid units that have withstood ten or more years of near constant use without failure of a bus controller. Hard disks and PSUs are much more likely to go belly up. Also, working in an office with easily 50+ Powermacs over the years (from the G3 on), I have never observed this problem.

A likely cause is perhaps that the software driver for the higher-speed ide bus controller was changed at some point during an update and resulted in these problems. Switching to the slower bus would fix any problems since this uses a different driver. (unlikely due to few, if any, reports)

A more likely cause, since this problem is exhibited on different machines, is that your facility is inappropriate for operating such equipment (humidity, heat, dirty unregulated power, dust that could impede airflow etc.) Perhaps invest in an uninterruptible power supply or an air-conditioner. The dirty power thing could be it since some integrated circuits are more picky about steady power than others, and Apple's PSUs are generally less reliable than those of some other vendors.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: bdog on Sep 07, '05 10:57:59PM

Please don't call this "nonsense"!!!! I hope this hint can help people have a working computer again...

I don't know if it gets 'worn out' exactly or if it suddenly goes flakey. I would think both are possibilities.

The 3 computers I've seen have been in schools. A lot of dust is the only major unwanted environment variable. What they did have in common was a lot of use (I/O on the bus) since 2 of these towers were servers (MacManager/File Server & NetBoot) and one was a movie editing station. So I concluded that the busses got 'worn out', or so it appears, at least.

A software problem is out of the question, since I tried different HDs (so different CLEAN installed systems) and I even tried different IDE cables, removing the SCSI card, testing and swapping RAM, ect.

Good point about dirty power! That very well be the cause. In the last 3 years, 2 schools have been struck by lightning. ;-) (but none of these 3 computers were in those schools)

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: Big Red on Sep 08, '05 10:42:37AM

Well, I was convinced for 4 months that my Smurf was toast. I could boot up from CD, but could not boot up on a drive attached to the primary IDE bus or a drive connected to an IDE66 PCI card, despite what felt like a dozen reinstalls. I thought it was the PCI card, memory, tried to pull everything. I finally detatched everything from the primary bus and booted with the only hard drives connected to the PCI card. It's been working like a dream ever since.

It was a nice excuse to buy a new machine, and now my girlfriend has a nice functional G3 tower for simple computing tasks.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: RoyalJimP on Sep 08, '05 04:11:52PM

I bought (on ebay) a sawtooth logic board to put in my Smurf.
I also found the only way to get stable operation was to put the IDE hard drives on the secondary slower optical bus. It would not boot for 9.2 to osX on the primary bus. And did silly things. Since switching busses it is very stable, no problems two years now. I've considered a PCI ATA 100/133 card just for the faster transfer speeds, but just haven't sprung yet cause it's been adequat for the use it's getting

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: obiwan2u on Sep 08, '05 12:55:14AM
Like everybody else, I'm a little unclear on exactly what "worn out" means.

I've noticed weird interactions with firewire and USB devices and my Mac G4 boot process. Remember that USB and firewire are involved with the boot sequence of the machine.

Specifically, I had a situation where my machine woudn't boot until I removed a firewire compact flash-mem card reader.

Could there be some other weird interaction like this causing your problems? It's easy to imagine changed behaviors if you're not actually conducting a controlled experiment.

Ben in DC
PublicMailbox at benslade
(append 030516 to the subj to bypass spam filters)
"It's the mark of an educated mind to be moved by statistics"

[ Reply to This | # ]

A possible fix for a highly unstable G3 or G4 Tower
Authored by: technomage on Sep 10, '05 06:20:48PM

I'm right with you. IDE controllers can fail, just like any other component, but that is an extremely unlikely event. These may be old machines, but such a failure is much, much less likely than a failing hard disk (servers and video production machines, you say?), mis-configured or buggy disk firmware/jumpers, or bad cables. If this is a school setting and happening on multiple computers of varying models, add to this the possibility that your master image is corrupt. It is well-documented that some older Apple IDE controllers are buggy, and replacing the OEM drives with new-generation ones can expose those issues.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: vsopal on Sep 08, '05 07:05:48AM

When I saw this "hint" I thought it will solve my 6-months old problem. But the problem is a bit different. Whenever I try to boot system freezes while grey circle is spinning at white screen with apple. Booting from CD also makes same effect.

So I went to see if when I disconnect hard drive will it boot from CD.
Unfortunately it didn't.
Before booting everything works fine i.e. Target Disk Mode, Open Firmware, holding down alt key lets select startup disk.

Can Anyone suggest me what else to do.
Shiver758 said that it was the only thing he hadn't try.
Now I've tried everything.


[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: Arturia on Sep 08, '05 08:52:45AM
Well... no ;-]

If you've got problems with one of your drives, first unplug your comp,
and remove the little lithium battery from the motherboard [replace it if needed]

Then, reset the cuda button that's onto the motherboard. [similar to a little pill]
Resetting means that you push it ONE [1] time.
Put the battery in place and replug your comp.

Some static electricity (or thunderbolts) can affects these connections.
Unpluging and repluging your drive may help.

I've got huge thunderbolts lately and my drives weren't showing up at startup. [the dreaded "No System" message]
I got a surgebar, so I don't know if the problem is really due to thunderbolts, or even static electricity... or else

Anyway... It was like my drives were totally vanished [TechTool, Disk Warrior and tutti quanti... were of no help]

Resetting the cuda and the unplug/repluging fixed it for me :-)

[ Reply to This | # ]

A possible fix for a highly unstable G3 or G4 Tower
Authored by: Arturia on Sep 08, '05 09:13:34AM
Here's the correct procedure that I've found somewhere on internet

Many system problems can be resolved by resetting the PMU chip.
Whenever you have a unit that fails to power up, follow this procedure before replacing any modules:

1. Disconnect the power cord and check the battery in the battery holder
The battery should read 3.3 to 3.7 volts.
If the battery is bad, replace the battery, wait ten seconds, and then reset the PMU (refer to the next step).
If the battery is good, go to the next step.

2. Press the PMU reset switch (S1) once on the logic board and then proceed to step 3.
Do NOT press the PMU reset switch a second time because it could crash the PMU chip.

3. WAIT ten seconds before connecting the power cord and powering the computer on.

4. Run Apple Hardware Test.

[ Reply to This | # ]

A possible fix for a highly unstable G3 or G4 Tower
Authored by: vsopal on Sep 08, '05 10:37:15AM

I'd checked it at the beginning of "slow death" of my G4 and I've check it now once again.
It didn't helped then and it didn't helped now.
Thanks for trying. Maybe you can think something else? I've run out ideas.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: theodour on Sep 08, '05 03:21:24PM

I had this problem with an old Performa 6200. It was highly unstable, and I narrowed the problem to the IDE bus.
It would boot fine from an external scsi drive -- or an internal, but you had to do some crazy things to keep both a CDROM and HD in there at the same time.
I think I even gave it to a friend to use for a while.
Poor friend.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: johnsawyercjs on Sep 08, '05 04:15:05PM

It's just as likely, if not more so, in most cases, that it's the hard drive that's beginning to have problems. The submitter of this hint didn't say if he tried a known-good hard drive on the logic board's hard drive ATA bus. Whether the problem is the hard drive bus or the drive that's plugged into it, the symptoms can be the same. However, I could see where the components in the hard drive's bus could age too--like everything else, they age due to heat, wear due to electrons flowing through them, etc.--not much you can do about it. You might be able to reduce the wear by keeping your Mac cool (install an extra fan or two; don't let sunlight directly hit the Mac; etc.), but I don't think this would help much unless your Mac is otherwise running unusually hot.

Placing a hard drive on the logic board's CDROM drive's ATA bus simply makes the hard drive run slower, which can eliminate a lot of problems if the hard drive's components have begun to age, and are no longer able to run at ATA33 or ATA66 speed. In the B&W G3, my timing tests show it to be about 75% of the speed of the G3's ATA33 hard drive bus; the speed difference will be more pronounced in the G4, which has an ATA66 hard drive bus. The CDROM drive's bus in the B&W G3 is ATA-3, and in the G4 too, I believe--not ATA33. ATA-3 runs slower than ATA33, etc. Some say ATA-3 runs at half the speed of AT33, but that's not the case--just about 75% of the speed.

The B&W G3 is the model that's particularly sensitive to these issues. The size of the drive on a B&W G3 can also be an issue--I've found that the B&W G3's hard drive bus, even the rev 2 B&W logic boards that have an improved ATA bus controller chip that allows the hard drive bus to run a master/slave drive configuration without data corruption, can't take drives larger than about 20 or 30 gigabytes. Larger drives on a B&W G3 can often cause intermittent bad behavior--intermittent failure to boot, crashes, directory and file damage, failure to copy files to and from the drive, etc. I don't remember if dividing the drive into 20 gig maximum volumes helps, but I don't think it does.

Though some hard drives will have problems in the B&W G3, many others will work fine--you just have to try them to find out. One utility I've found that will pretty reliably tell you if there's a problem with read/write accesses, is LaCie's old "TimeDrive" utility, version 3.0 or later--it came with the Silverlining formatter. Unfortunately, they don't seem to have it on their web site any more, and it's not on VersionTracker. Other utils like Techtool might properly tell you if there's a problem though. Another test that can often indicate if your drive is working OK, is to use the Finder to duplicate a multi-gigabyte (2 gig or larger) folder full of files, and see if you get any read/write errors.

On G4s, these bus problems are much less common, so I'd try a known-good drive on the logic board's hard drive bus before you move the potentially problem drive to the CDROM bus.

Also check out Applecare document 31347: "Power Mac G4 (Digital Audio): 40 GB and 60 GB Hard Drive Issue", for info on how some of the drives Apple installed, have these problems.

One way to get a G3 or G4 to handle larger, newer hard drives properly, at full speed, is to install a PCI slot ATA card (Sonnet Tempo, etc.), and connect your hard drive to it. The rev 2 B&W G3 logic board will allow two or more drives to be connected to these cards, but if the G3's logic board is a rev 1, often you can't connect more than one drive to the card, and you might not even be able to plug another drive into the logic board's hard drive bus--doing so can introduce instability again. If you need to run two drives in a B&W G3 rev 1, especially larger newer drives, the most reliable arrangement is one drive connected to a PCI slot ATA card, and the other connected to the CDROM drive bus.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: bdog on Sep 08, '05 10:13:38PM

There are so many troubleshooting steps I didn't want to list them in the hint. But yes, I did try replacing the HD, every time. This hint is one of the last things you should try, since a bad HD, bad RAM, corrupt PMU and/or PRAM are all much more likely.

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: imageworx on Sep 09, '05 09:15:30AM

Why not get a PCI-ATA card that can boot an IDE drive? Might even try a ATA-133 RAID-compatible bus card (make sure you use the appropriate cable).

Like all things, its likely that we get a lemon now and then (a delicate trace, a resistor issue, a logic problem, a nicked ribbon cable, environmental...) and this could be interpreted as such.

However, I've seen some oddities where they are unexplainable at the moderate level. Some tricks like setting drives to Master and Slave from CS have resolved this. Or replacing the CDROM/DVDROM drive with a Burner has improved transfers...on the other bus!

Its like SCSI Voodoo all over again! But without the IDs...

To BeOS or Not to BeOS

[ Reply to This | # ]
A possible fix for a highly unstable G3 or G4 Tower
Authored by: bdog on Sep 09, '05 10:03:48PM

Well, this fix dosn't require the user having to buy anything.

Yea, CS versus Master or Slave can also make a difference in certain cases.

[ Reply to This | # ]