|
Post by Tom Maneiro on Dec 13, 2005 9:50:12 GMT -5
Done with disk checkings: from more than 50 floppies, i recovered only 2 sane disks (HD, one Dysan and other Memorex), and one faulty floppy. The entire BTOS disk set was damaged, and i decided not to check the Windows/286 sealed disksets (why i want to break the seal?)
Also i made my own cleaning disk, just took a floppy, and stick the internal fabric coating to the magnetic disk. Put some drops of alcohol, and voila! Is either that or a $3.99+S&H cleaner on eBay ;D
|
|
|
Post by GiGaBiTe on Jan 2, 2006 19:27:19 GMT -5
you gotta get more ram, since you're probably using an overwhelmingly amount of virtual memory. i think i have a buncha 4mb and 1mb 30 pin simms, but they're for macs (no parity), since i was gonna use them for some homebrew dram projects ram is ram, there is no pc or mac specific ram (except for VRAM sticks and the ram in the IICX) 30 pin simms will work in both pc and mac, the same said for all other types of ram.
|
|
oompa loompa
I AM THE GOVERNATOR
"Git 'Er Dun!"
Posts: 1,301
|
Post by oompa loompa on Jan 7, 2006 23:44:26 GMT -5
30 pin simms come in two types: simms with parity, and simms without parity. ibm machines use parity for memory error correction, and can't go without it. early mac machines that used 30 pin simms didn't require parity, so a simm with a parity will still work because these extra pins aren't used.
30 pin simms that have parity can be identified easily, if there are 9 or 3 chips on the simm. if there are 8 or 2 ic's on a simm, the simm doesn't have the extra parity bit. i have 30 pin simms without the parity because my projects can do without it =P
|
|
|
Post by Tom Maneiro on Jan 11, 2006 9:18:01 GMT -5
Got another dead for the collection: an AcerPAC with a P133 and 24M of RAM. Case was in a very bad shape, with lots of dust. Also that box has a NEC floppy, a fried Fujitsu HD, and a nice led panel with a reset switch, Acer-custom (i tested it on the 386 box, and it works, now i have a reset button!).
The machine boots.. and die after exiting the POST tests and drive detection stuff, where it's supposed to search and run from a bootdisk. What could be happening there? The machine model is a CT0063D3, and i can't found nothing about it on Google (it was easy for the 386, and the info was on Spanish...)
|
|
|
Post by jlf65 on Jan 11, 2006 13:53:17 GMT -5
What size/age drive did you put in it? A P133 has got to be an old system. I doubt the BIOS would handle modern drives. You mght have to scrounge up a 400M (yes MEG) or smaller drive for it.
|
|
|
Post by Tom Maneiro on Jan 11, 2006 16:50:51 GMT -5
I tried with a couple of floppy drives (3.5" and 5.25"), and with a OLD 85MB HD. The BIOS detects all drives with the correct CHS settings (and supports large-size HDs), but, after detecting all drives, it simply blank the screen, then hang.
At least the processor and the RAM are in good shape, so i could raise some bucks with these...
|
|
|
Post by jlf65 on Jan 11, 2006 20:51:25 GMT -5
Sounds like an issue with the Shrug and Pray. Oh well, you can get more money selling the parts than for the machine as a whole. ;D
|
|
|
Post by Tom Maneiro on Jan 13, 2006 18:01:00 GMT -5
Just discovered the problem: faulty RAM! I got parity errors, and after playing a litte with the SIMMs, i got MSDOS 6 to boot... in "safe mode". Loading HIMEM with these SIMMs (two 8MB Hyundai and two 4MB Powmem) cause more hangs. It should be easy to find 64MB SIMMs and install Win2K in that box...
|
|
|
Post by GiGaBiTe on Jan 15, 2006 5:32:26 GMT -5
IBM machines did not need parity to run, neither did early macs. it did help to have parity back then though, because memory quality sucked. i can remember getting more than a few bad sticks of ram.
why not just tell them on 30 pin simms, there are 9 bits instead of 8, instead of the whole 9 chips instead of 8, or 3 instead of 2, lol.
also, you really dont want to waste your time on trying to find some 64 MB simms, most PCs didnt support them (64, 128 and the rare 256's were mainly for stuff like sun workstations and servers) the max you probably want to try to hunt down are some 32 MB EDO's, as FPM sticks are actually slower.
|
|
|
Post by Tom Maneiro on Jan 16, 2006 19:26:22 GMT -5
Did more tests in the 386: -GenRomSuite: It take a while to checksum a ROM, but for a weirdo reason, it fails to show the ROM header in the window. bug?
-BasiEgaXorz: it runs! it compiles! (albeit slow, but it works). Point that in the FAQ. Now... can you run your programs there? NO!
-I found some old copies of ACDSee 2.41 and PSP 4. Now the slogan is "Your JPG in 30 seconds" ;D
|
|
oompa loompa
I AM THE GOVERNATOR
"Git 'Er Dun!"
Posts: 1,301
|
Post by oompa loompa on Jan 18, 2006 19:09:51 GMT -5
x.x, i have a small 386 33 mhz (or maybe it was 25 mhz =P) sitting in a box somewhere. i'll have to take it out, and check out those non-parity memory simms on it. i remember i couldn't even get into the bios before because post didn't even detect any base memory
i've meant to get a case for the 386 motherboard, and develop my own bios code, but haven't gotten around to even start
|
|
|
Post by Tom Maneiro on Jan 19, 2006 16:08:32 GMT -5
Do 72-pin SIMMs have parity?
Back to the P133, i discovered some nasty issues: i put fresh SIMMs there, thinking in bad modules, but... i got the same error...
"RAM Parity Error. Press ESC to turn off NMI or any other key to reboot".
I managed to hook a floppy and run memtest86+. I let the test run by about 7 hours, but i do not got any error... Tried to run Linux (using muLinux, a floppy distro that runs on RAM), and got something that in my years with Linux, i have never experienced: a "kernel panic".
> RAMDISK: Compressed image found at sector 545 > Kernel panic: free space corruption
Finally, i went into the BIOS and disabled memory parity, but i still was unable to boot into full DOS. Booted in minimal DOS and checked config.sys: the problem was in EMS386 driver. Disabled it, and system boots fine! Win31 runs fast, DOOM is now playable at full speed...
Now time to check RAM modules: With two Hyundai 8M modules that i have, system is unbootable past the POST if both modules were installed in the first bank. With some other 4M modules, system seems to work "OK".
I checked the motherboard, and got no misconfigured jumpers (there is about 20 jumpers to setup the system), no broken traces, no nada!. Could it be a corrupted BIOS? Since i can't found info about that machine (an AcerPAC, model posted earlier), i can't get an updated BIOS (if it still exists).
ACERPAC Update: Got more sane 8M and 4M SIMMs, and a P75 (i thinked also in a faulty CPU), and now i got weird results (most of these are random): -Harddrive not detected -Graphics corruption -System unbootable past the POST -System unbootable past the Video BIOS message -System unbootable at all -System unbootable past the POST if using 8M SIMMs in SIMM1 and SIMM2 slots -Crashed starting DOOM
Help! I don't know what i can do now... (PS: i can trade a bunch of RAM and a couple of Pentiums by 4 30-pin 4M SIMMs)
|
|
|
Post by GiGaBiTe on Jan 23, 2006 16:12:34 GMT -5
any type of ram can have parity on it. and memory is not referred to as "sectors" its referred to as "pages" so the kernel is bitching about a sector on the floppy disk most likely.
|
|
|
Post by Tom Maneiro on Jan 24, 2006 10:25:34 GMT -5
Sector? Who mentioned sector here? ;D Another fine piece of old hardware was just added yesterday to the 386 box: a nice AT&T/Archtek 1914BAV fax-modem (a 8-bit ISA card!), that also works as a basic sound card. The driver search took me about 30 minutes, because the modem has no model printed on it, just a PCB code. Using the FCC ID i was able to "extract" the modem name, and finally, i found the driver on the manufacturer's website. How nice of them... keeping drivers for dead devices. Next: a SoundBlaster and a NE2000! (and maybe some extra RAM, if i can pay the 70€+S&H that i looked in a french store)
|
|
|
Post by jlf65 on Jan 24, 2006 15:05:21 GMT -5
It seems to me that a 386 is too old to support EDO or FPM memory. It more than likely needs PLAIN memory. That was a problem on old Amiga accelerators as well - some needed plain memory, some needed FPM instead of EDO, and a scant few could use EDO. By the time EDO became usable, PCs were all using 486DX4s and Pentiums.
|
|