Hmm, I was going to write about that 6 GB Seagate I have in the G3, but I already did that a few pages back, so.
I do want to talk about wipe some more, though. Given the events of the past few months, I would dare say they're not all that paranoid, but that they're concerned about the wrong things, Most of the threats they rattle off are ones hackers and libertarians were scared of in the late 1990s; modern compsec types would be more afraid of rogue states like the Russians (or possibly even Trump's brain trust if it knew its ass from a hole in the ground), or malware networks.
Also, unlike when a lot of these wiping thingies were made, hard drives are essentially disposable now. If you're that worried about the data on it, a hammer or a hot furnace are all you need. (There's also acid, but I understand the magnetic coating is pretty corrosion-resistant and would need aqua regia.)
The iMac G4 is no more. A component on the logic board let out the magic smoke at some point, and I was neither able to revive it nor find the failed component, so it's gone. RIP.
I FINALLY HAVE THE IBM IN A GOOD PLACE. Windows 2000 seems to be where it's happiest, and I have it running with a few modern accoutrements (Radeon 9250 PCI video, the Promise SATA card that's made a few appearances here before) as well as the AWE64 and an ISA network card (crappy, but the best I can do on a machine with only 2 PCI slots and no USB). It plays CANYON.MID pretty well! I haven't done much else with it yet, though.
The Power Mac G4 and G5 are still up and running, though the G5 is sloooooow for reasons I haven't figured out yet, and I've upgraded the RAM and the HDD and neither have helped. It feels like it's running with the L2 cache turned off. I haven't had time to investigate this further; maybe this weekend.
It turns out my 486 *did* have something wrong with it:
The BIOS is too old to support the Pentium OverDrive (which was what was causing the "random" lockups). No one seems to have archived all the old PB BIOS revisions, so I can't just update it. Dropping in a 5-volt 486DX2 has solved this problem neatly.
The board itself isn't stable at 33 MHz bus speed (no idea...probably the memory on the motherboard not being fast enough). It's rock-solid at 25 MHz.
Adding cache to the board actually makes it slower unless you go all the way and add 512k. Adding 512k requires chips that are impossible to get now, so...no L2 cache. The thing is plenty fast on the 486's L1.
And once again, it turns out the 486 had yet another thing wrong with it. This time, it seems to have been the IDE cable! I was using a cheap, rounded IDE cable before because the case is so tight, but those rounded cables have a unique property to them, and that unique property is that they all suck. It's even worse on a 486 without Ultra DMA (and thus no interface CRC). Replacing the rounded cable with a normal 40-pin IDE cable magically fixed a lot of issues (for one thing, Windows 9x doesn't lock up on the POD with the Cirrus card in now).
Also, I figured out why my 486DX2 isn't stable at 66 MHz...it's not rated for 66 MHz. Duh. If I run it at 50 MHz like it says to, it's rock stable.
I was able to find the special 128k x 8 static RAM this thing needs (Micron MT5C1008-20 if anyone's looking for it) and now the system has the full 512k of cache it supports. It doesn't help all that much, since the chipset still sucks, but it's nice to have.
Right now, I'm working on getting Windows 95 and DOS stuff going on it. It does very well in 95, not so good in 98SE (yes, even with the POD in).
This machine is actually a really good Doom machine, even with an ISA video card and the 486DX2-50. I took my mini-speakers downstairs, so sounds isn't always working, but I've listened to it on headphones and it's quite good.
Also, revisiting the Medalist Pro thing from a few pages back: I'm pretty sure these things can trace their lineage to the CDC Swift somehow, but I just haven't found the missing link yet. That missing link, however, may be the ST1400/ST1480, which were 4412 RPM versions of the Swift released around the same time as those weird 4500 RPM things. I have one on order from eBay, and yes, PCB pictures will be happening.
So! I got the ST1400 (since I couldn't find a 1480) from eBay. The one I have is in not-great shape, but it does seem to work. The command processor is an 8051, the servo processor is an AMD 80C188 (!!!), and I saw no trace of the Zilog or ST chips that were in the ST3283 series, though I *did* see a TI glue chip not unlike the ones on the Barracuda and Hawk of that era.
At this point, I'm thinking the ST3283 was a clean-sheet design that borrowed bits and pieces from the Swift, but wasn't an exact copy. It also seems to have cross-pollinated quite a bit with the ST3144 and ST3491 series later on, the drives with "the chip" (a servo processor thingy specific to the ST3000 and ST1000 drives) on them. The Medalist Pro 7200 came out of this mess eventually.
Comments