Thanks for all the advice, folks. My computer seems to have stabilized for now -- nothing funny's happened in the last few on/off sessions. But it's durned strange, in the way that apparent crashes in Windows led to repeated failure in the CMOS startup.


> Ok, first things first. Be sure not to jump the gun with a BIOS/CMOS FLASH. If you happen to get the wrong one you're basically dead in the water and not much is gonna rescue you. And even the right FLASH can screw things up royally. FLASHING is left as a last resort.

Heh. I'd figured if the Bios settings had suddenly toasted, flashing wouldn't help that much since the problem could equally likely be physical... like a spike short circuit inside the chip. It's not as if I needed an "updated" flash, since everything's been working all together, just fine, for over a year -- my Bios didn't suddenly become "obsolete" with my particular setup. All I'd done immediately before this problem started was: uninstall Dimension4, a time-setting program; and install Webshots, a desktop background changer. And yes, I always do a restart after each install.


> > Okay, so I've been having a strange intermittent hardware failure, which a few minutes ago became a MAJOR failure. Previously, I'd be in Win98, and have the file manager constantly windowed (which I'm aware opens more than 300 files to do its job). Then my HDD would seem to freeze with the access light lit continuously, and while it was doing so, it would also jump back and forth to accessing my CD-ROM -- even though I don't keep my CD-ROM loaded with any disk. Weird.
>
> On a healthy machine, this is called paging, and can be fixed with the addition of RAM. It just simply doesn't have a large enough paging (swap) file to accomplish all it wants to do. A bit more RAM will usually clear this right up. As for your machine, if this is part of the same problem you described further down it could be something else, but probably not.

I have 64 MB of RAM, and let Windows control the size of my swapfile (usually 2x the size of my RAM bank). But why would an OS want to page swap to the CD-ROM? It can't write anything to that device!?


> >
> > Frankly, it sounded like my IDE cable(s) had gotten loose from the motherboard, but I've checked and everything's secure. While I was at it, I disconnected the 24x CD-ROM -- on the off-hand chance that that's the part that's defective (or so Dave's been claiming for months).
>
> Bad CD-ROM was going to be my guess. If you have disconnected it and are still having errors, but not the same one(s) did you change the jumper of the hard drive it was connected to? If the HDD was Master w/Slave present, did you change it to Master/Single? If the HDD was set to Slave, did you change it to Master or Master/Single? The drive should have a diagram.

At first, after I completely disconnected the CD-ROM, I couldn't even get the puter to boot. It kept reporting "Secondary HDD failure". I broke apart the machine again, re-examined the jumper settings, put everything back in its place again. This time I was able to start into Windows, but as soon as I accessed the Secondary Master it crashed again. Dave started suggesting that my motherboard had gone bad. I still had no idea what was going on, since I could use a remote connection to retrieve files from *both* HDDs with no problem. Finally, yesterday I was able to get Windows to start without crashing and we defragged all my partitions. I haven't had any problems since. If it keeps behaving well, I may even put back the CD-ROM later to see if it caused the original failures.


> (1b) Could a messed-up CD-Rom connection or driver have really led to all this? I thought not, but...
>
> (Ans. 1b)Connection? Maybe. Driver? Again maybe, but probably not. Bad drive? Yes, but in any case disconnecting it should have cleared it up if it was the sole problem.

The CD-ROM was Slaved to the Primary Master HDD, so removing it should indeed have immediately solved the problems. Instead my Secondary Master/Single reported as failed :-(


> (2) WHAT EXACTLY is the CMOS *doing* when it auto-detects the HDDs? If it tries to get a response from the ID-FAT from the HDDs and not getting anything, then I'm toasted because the main HDD could be fried. But surely not both drives.
>
> (Ans. 2) From my understanding of things, what you described is exactly what it does. But some drives are wierd. For example, most WD's need to have the jumpers set across pins 3 and 5 for Master/Single even though that is not on the diagram on the drive. If you're running WD's you might try that.


Dracimas: I put the jumper across pins 4 and 6 for Master/Single on the WD. Could that possibly make any difference? Are you saying that what the diagram says for WDs is *not* optimal?



> (3) Why the heck does my CMOS think both my Masters are Slaves? True -- I don't have any Slaves. But it's never had problems doing Auto-Detect before... Though it *did* always read my WD as "Matsumi 4something" and the Maxtor as "None".
>
> (Ans. 3) Again, I'd make sure the jumper settings are in order. Connect each HDD to it's own IDE controller as Master/Single. Save the BIOS FLASH as long as is possible, and hard set the drives in BIOS instead of Auto Detect. Hopefully that'll get you going.

Thank you, I set the BIOS to skip the auto-detect sequence for Slave drives :-)


> If it does, you might take your CD-ROM drive in somewhere to be tested to see if maybe it needs replacing. If not then you might contact someone locally who can run some diags on this machine to try to pin-point the problem.

I have hardware testing programs. What I don't really know is What to troubleshoot in a possibly defective CD-ROM drive?


> Drac "Any other ideas?" imaS

Pin oxidation? Random overheating? Radiofrequency emissions conflict between daughtercards? My hydro got spiked by a lightning hit 15 km away? A new bizarro hardware virus? Defective IDE ribbon cable? Well, fortunately, I do know where pin 1 connects on the cable :-)

Thanks!

Wolf "Cat hair sucked into the CD-ROM bay?" spirit