New adventures in SCSI. I was digging around this week and I found an old SCSI card. The PCB identified it as an Initio INI-9100A. I seem to recall I got it with my CD burner a few years back, and that I ditched it when it wouldn’t work with Windows 2000 RC2. Curious to see if drivers were ever released for it, I checked Initio’s Web site, and lo and behold, the release version of Windows 2000 was supposed to support the card. Since I’ve got a couple of decent SCSI CD-ROM drives laying around, I figured, why not try it?
I was unhappy to see Windows 2000 failed to bring up the Add New Hardware wizard after installation, but when I looked in Device Manager, the card was there. So I powered down, connected an old NEC 12X SCSI CD-ROM to it, powered back up, and bingo! I even got activity during boot. A quick verification by reading the disc, and I’m happy. So I powered down again. What else can I throw at it…? I spied an ancient, ancient Quantum 52 MB SCSI drive. (Don’t ask me why I keep this stuff.) I knew the card wouldn’t boot off it, since it lacks an onboard BIOS, but would the drive still work, I wondered? So I powered down, plugged the ancient thing in there, power back up, and I thought I saw the drive’s LED flicker during boot. Yes, back in this drive’s day, hard drives had LEDs on the front of them. They even had faceplates! I watched Windows finish booting, opened My Computer, and sure enough, I had an extra hard drive up there. But what on Earth could be on it? I opened up the drive, and hit gold. I must have used the drive sometime within the past five years, because it contained a copy of Caldera OpenDOS. That, believe it or not, is extremely useful. OpenDOS’ FDISK will delete any partition, unlike Microsoft’s. So I’m very glad I tried the drive.
So now I’ve got two SCSI-equipped systems, one of which is bootable. I’ve got an excuse to go buy a $220 IBM or Quantum 10,000 RPM SCSI drive… Uh oh. Good thing I got some overtime at work this week and will get some more this weekend.
And it looks like it’s unplanned upgrade time. Tom Gatermann called me up yesterday. He was replacing our friend Tim Coleman’s hard drive and the system just wouldn’t come back up. He futzed around with it for an hour, trying everything he could think of, then called me. I had him try putting the hard drive on auto detect, reset all the PnP/ESCD data, and of course, check all the cables. Nothing. The board would POST, then die. Well, without having a POST card (I know how to make one except I don’t have an EPROM burner) I can’t diagnose it any further.
Come to think of it, I should have had him disable L2 cache and see if that brought it back to life. It’d be slow as can be, but that’s a good way to troubleshoot a Socket 7 system, or a 386 or 486 for that matter. Strip the system down to just motherboard, CPU, video, a boot device, and a single bank of memory (a single DIMM or one pair of SIMMs). Disable L2 cache. If it works with L2 cache disabled, it’s a motherboard problem of some sort. Check all jumpers, re-seat anything that’s re-seatable, and try again. If it still doesn’t work, it has to be a CPU, video, or memory problem. Then you’ve got a few more steps to try, including disabling L1 cache and switching out the video and memory.
Tom took the system home with him, so I’ll be giving it a look today.
At any rate, it looks like we’re dealing with a blown board, and every time Tom or I do anything with that system, something goes. Tim’s on power supply #2, motherboard #2, sound card #2… Tim’s got an army of cats, and the system’s on the floor, which gives me concern. The system can pull in cat hair, and with it, static electricity. And I don’t know how good Tim’s wiring is. It’s a very maddening problem. Had anyone else built the system, I’d be cursing them, but Tom and I built it ourselves, and we used the same calibre parts we use in our own systems. So we think it’s an environmental problem.
I’m thinking I’ll go ahead and pick up a Gigabyte motherboard with a Duron chip on it, then give Tim my two-year-old AOpen AX59Pro board. I normally run systems much longer than that, but I want to help Tim, and I really ought to try to stay somewhat current.
Linux 2.4 again. I was right. Within 4 hours of 2.4’s release, Alan Cox released Linux2.4-ac1. A few hours after that, Linux2.4-ac2 followed. When does he sleep, I wonder?