Joseph_Joestar wrote on 2024-02-24, 19:36:

Give it time.

Data corruption isn't always immediately apparent.

True. I admittedly don't use that other system very often so I probably just have been lucky so far.

Still, I'm curious as to why the 1TB setup is having more immediate problems including with the Windows setup itself, but my 250GB setup did not.

At any rate, I think I'm going to go with a 128GB drive for the Win 95/98 partitions and a 250GB drive for WinXP. Hopefully that will absolve any potential drive size issues.

Did another attempt at setting this up with a 128GB Lexar SSD. Spoilers: didn't work, same (ish) problems.

This time I set up a 40GB partition for Win98 (plus the 16GB partition for Win95).

When performing the surface scan (again, canceled before finishing) BootIt reported zero bad clusters. This is different from the 1 TB drive where it reported every cluster as bad.

The Windows 98 setup error was different as well, this time reporting a corrupted CAB file. But it still didn't work.

When I booted into Win 98 via my CF card and scanned the drive, I got the same errors as before: duplicate file allocation tables that didn't match, and parts of files stored in invalid locations.

I tried formatting the partition in Win 98 and then copying over files from the CF card. Same issues reported.

I further swapped the IDE cable and IDE-to-SATA adapter and did everything all over again. Still the same issues, but at least I know it's probably not a faulty IDE cable or adapter.

I also connected the drive to my Windows 10 machine and ran a drive scan there. It reported no errors.

Round 3.

Installed the latest (non-beta) BIOS for this motherboard. Don't think it would do anything, but I suppose it couldn't hurt.

Used AOEMI Partition Assistant to create a single 40GB FAT32 partition on the 128GB Lexar SSD.

Windows 98 setup actually ran this time. But upon rebooting it simply halted the system with a Windows Protection Error.

Tried installing the RLoew TBPLUS patch to see if that would do anything. Rebooted again and received a BSOD.

Booted with the CompactFlash Win98 install and ran scandisk and the SSD drive once again reported a whole bunch of file/folder errors.

*sigh*

Final test of the day.

Tested installing Windows 98 on the Lexar SSD using my Athlon XP 2000+ setup (Asus A7V600 motherboard).

Windows 98 installed without a hitch. Running scandisk results in no errors reported.

Re-installed the drive in the Pentium 4 system, booted from the CF card, and running scandisk there results in the same bunch of errors (invalid locations, etc.).

This at least narrows down the incompatibility to this specific combination of OS, motherboard, and drive.

Final test of the day (for real this time).

I did some searching on this forum regarding the SiS chipset and SSD drives. I stumbled across a post which mentioned changing the PIO mode in the BIOS.

Decided to try that and lo-and-behold, changing the PIO mode to 3 seems to have worked!

Ran scandisk on the drive and no errors reported. Finally!

Now I get to go back to the beginning, re-install BootIt, re-partition everything and try this whole thing over. But that will be for another day.

Shponglefan wrote on 2024-02-25, 04:11:

I did some searching on this forum regarding the SiS chipset and SSD drives.

Are you using those generic SATA to IDE adapters with the JMicron chip? They have compatibility issues with certain VIA chipsets (KT133A) but I wasn't aware of anything similar regarding SIS chipsets.

In my case, attempting to install Win9x using one of those adapters on my Abit KT7A motherboard resulted in terrible performance, but no data corruption. I read somewhere on this forum that more expensive SATA to IDE adapters (e.g. from StarTech) which use the Marvell chip work better on such cases, but I never tried that personally.

Shponglefan wrote on 2024-02-24, 19:47:

Out of curiosity, I did the same on my 250GB Windows 98 system and it reported no such similar errors. Not sure if the 1TB drive is triggering more issues?

here is MANY problems appears when original 9x/me used with newest machines.
please use RLoew patches http://lonecrusader.x10host.com/rloew/rloew.html

MERCURY127 wrote on 2024-02-25, 12:01:

Shponglefan wrote on 2024-02-24, 19:47:

Out of curiosity, I did the same on my 250GB Windows 98 system and it reported no such similar errors. Not sure if the 1TB drive is triggering more issues?

here is MANY problems appears when original 9x/me used with newest machines.
please use RLoew patches http://lonecrusader.x10host.com/rloew/rloew.html

In my troubleshooting of this, I had tried the RLoew TBPLUS patch. But that didn't fix the problems since the problems in turns out weren't related to drive size.

Joseph_Joestar wrote on 2024-02-25, 05:18:

Are you using those generic SATA to IDE adapters with the JMicron chip? They have compatibility issues with certain VIA chipsets (KT133A) but I wasn't aware of anything similar regarding SIS chipsets.

In my case, attempting to install Win9x using one of those adapters on my Abit KT7A motherboard resulted in terrible performance, but no data corruption. I read somewhere on this forum that more expensive SATA to IDE adapters (e.g. from StarTech) which use the Marvell chip work better on such cases, but I never tried that personally.

I'm using the Startech adapters with the Marvell chips. I've even tried multiple Startech adapters just to make sure the issues weren't specific to one of them.

I do have one of the other cheaper SATA-IDE adapters, but didn't try it. It might be interesting to see if it resulted in the same issues.

Try the cheap IDE to SATA converters and see what happens. There isn't any IDE to SATA bridge chip I know of that is 100% compatible with everything. The Marvell chips had compatibility problems with some Intel chipsets, so maybe something similar is happening with your SiS chipset.

grub is my bootloader of choice. it doesn't break at the drop of a hat--you actually have to do something to make it stop working. i even use it for win10 with no linux partition.

fat16 is a royal pain. are you sure you want dos 6.22 and win95 instead of just enabling a dos boot menu in win98? if you do go this route consider booting dos 6.22 off a floppy and storing the non-system files in a directory on the win95 fat16 partition during boot.

lti wrote on 2024-02-26, 01:09:

Try the cheap IDE to SATA converters and see what happens. There isn't any IDE to SATA bridge chip I know of that is 100% compatible with everything. The Marvell chips had compatibility problems with some Intel chipsets, so maybe something similar is happening with your SiS chipset.

At some point soon I'll test that. I'm going to set up a separate SSD for that, since I've already now installed Windows 98 SE using the Startech adapter. With PIO Mode set to 3 it seems to be working perfectly (so far).

I'm also curious to see how an NTFS / WinXP set up works with the Startech adapter and if it will also need to be set to PIO Mode 3.

stanwebber wrote on 2024-02-26, 03:31:

fat16 is a royal pain. are you sure you want dos 6.22 and win95 instead of just enabling a dos boot menu in win98? if you do go this route consider booting dos 6.22 off a floppy and storing the non-system files in a directory on the win95 fat16 partition during boot.

Oh, I'm definitely including DOS 6.22 and Win 95. That's not optional. 😁

In fact, I've already got DOS 6.22 setup. I ditched my original plan of trying to install everything on one drive. Instead I set up DOS 6.22 on a 4GB Compact Flash (dual 2GB FAT16 partitions). I did this by booting from the DOS 6.22 set up disks and using FDISK to create the partitions.

Windows 95 and 98 I'm putting on the same 128GB SSD and Windows XP will go on a 250GB SSD. I've got Windows 98 SE installed. Windows 95 and XP are next on the list.

For the bootloader I'm trying out BootIt Bare Metal. It functions as both a bootloader and partition manager. It's been trivial to set up booting from any partition, as well as hiding all of the other partitions from their respective operating systems.

Progress update: I've installed DOS 6.22, Windows 95 and Windows 98 SE.

DOS 6.22

I installed DOS on a 4GB Compact Flash card. I booted directly from a DOS 6.22 setup disk, used FDISK to create a pair of 2GB partitions, then installed DOS from the installation disks.

Since I already had another install I'd been using to test this system out for DOS compatibility, I simply copied over the drivers and AUTOEXEC & CONFIG files to this new install, along with about 50 games.

Altogether this took about 20 minutes to set up. I love DOS. 😁

Windows 95

I installed Windows 95 using setup files copied to the 16GB partition on the SSD. As part of the install, I used a boot disk with a CPU speed fix to make it compatible with the 2.8GHz processor.

I've never installed Windows 95 on such a fast system before. Before I do anything further, I'm compiling a collection of updates and drivers, and planning things according to this thread: Windows 95 setup guide for the 2020's

Windows 98 SE

I had to install Windows 98 twice. I ran into some issues with drivers the first time around.

I initially used the ESS Solo-1 drivers from the VOGONS driver library. However, installing the drivers caused the system to freeze up requiring a hard reset. Upon rebooting, the card seemed to install fine. However, I could not get any MPU-401 playback. No errors, just no audio.

The bigger issue occurred after installing the Diamond Monster 3D II drivers for the Voodoo2 card. This caused the Display Properties to no longer activate properly. Removing the physical card would allow the Display Properties window to show up, but would report a DESK.CPL error. I tried removing and installing different drivers, but the problem remained.

Finally, installing the LAN driver resulted in the system boot times increasing significantly.

As a result, I nuked the entire installation and started over. For the ESS Solo-1, I tried the drivers from Phil's Computer Lab instead. They appear to be a newer version and worked perfectly. MPU-401 playback now works. The only issue I'm now having is that DOS SB emulation doesn't work with digital audio on DMA channel 1. It results in distorted static. Changing to DMA channel 3 produces correct digital audio playback.

For the Voodoo2 card, I used the 3Dfx Voodoo2 reference drivers. These also worked with no issues thus far.

And last, I omitted installing the LAN driver. I don't know if I'm going to need to use this PC on a LAN, so for now I'm foregoing that driver.

Another small update:

To accommodate audio across all the operating systems, I'd been planning to go with a triple sound card setup.

• ESS Solo-1 (DOS / Win 95 / Win 98)
• Diamond Monster Sound MX300 (Win 95 / Win 98)
• Creative Labs Audigy 2 ZS (Win XP)

Since I have both CD audio and a wavetable daughterboard connected to the Solo-1, I wanted to use it with both DOS and Win 9x. But I also want the Diamond MX300 in Win9x for A3D supported games.

I instaled the MX300 under Windows 98 alongside the previously installed ESS Solo-1. Installation went well. I did have to disable the MX300's SB emulation since its default settings conflicted with the Solo-1's SB emulation. But that's fine, since I'd rather be using the Solo-1 for SB support.

Testing it out, it appears that everything works the way I want it to. I can use the Solo-1 for CD audio and MPU-401 playback, while the MX300 seems to be working for Windows sounds effects. Haven't tried any Windows games yet, so I'll need to test that next.

I've also been considering what to do with the Audigy 2 ZS. While I intended it primarily for Win XP, it could be useful in Win 98 as well. I had previously tried a MX300 / Audigy 2 ZS setup, but ran into stability issues with both cards installed.

Alternatively, it would be trivial to add another Windows 98 partition and just have a completely separate install for the Audigy 2 ZS. I might end up going that route.

Ran into some more shenanigans tonight.

Turns out the SSD drives do not like being set as 'slave mode' on their respective IDE cables. More specifically, it causes whatever drive is set as master to not work properly. In order to get everything functioning properly, both SSDs need to be set as master for each of their IDE cables. Then the CF adapter and DVD-ROM set to slave mode respectively.

I have IDE cable extensions so I should be able to make everything physically reach inside the case. But it's going to take some cable origami to sort it all out.

Fortunately this case has no window, so I'll be able to hide my cable management shame when I'm done. 😅

Another update: installed Windows XP.

For the most part this went fine. I did run into a weird error in BootIt when creating and formatting the partition for Windows XP. For some reason after the format it popped up an "insufficient memory to complete operation" error. Googling this message didn't return anything.

Formatting the partition during the Windows XP setup went fine and XP installed without problems.

For this drive, I left the BIOS settings at default which included PIO mode 4. This didn't cause any problems installing or running XP. Drive scans produce no errors. So I'm assuming the issue with the other drive is specific to Windows 95/98 FAT32 partitions perhaps?

I have been juggling a lot of IDE cable management trying to figure out an optimal solution, so also not completely sure what role different IDE cables may have played.

I also discovered that the Compact Flash does not work with the 12" cable extension. It works fine with the 6" extension though.

Conversely, the DVD-ROM works fine with the 12" extension cable. So I'm assuming the cable itself is fine, but may be hitting limits of IDE based on physical length of the cables.

After a bunch more testing and troubleshooting, I may have to pull the plug on this build.

Two main issues right now are:

1) Windows 98 SE is unstable when it comes to any hardware changes. The symptom is always the same. The system boots and works for about 10 seconds, then completely freezes. It does this on any change from the installed hardware. Even plugging in a PS/2 mouse instead of a serial mouse causes this issue.

I could probably live with this, except for the next issue:

2) The Audigy 2 ZS with EAX enabled results in a loud digital interference noise under Windows XP and Windows 2000. I've been troubleshooting this for the past couple days (thread here: Audigy 2 ZS noise issue under Win2k / WinXP (seems related to EAX reverb?)), but despite changing all the hardware (except the motherboard), trying different drivers, and even two different operating systems (XP/2000), the issue persists.

I'm at a point where it feels like there is an issue between the Audigy 2 ZS and the motherboard. While I could mitigate this issue by setting the Audigy 2 ZS to standard h/w acceleration, this disables the EAX capabilities of the card which defeats the whole purpose of using it.

On the upside, the multi-boot capabilities using BootIt Bare Metal has been a success. I've been able to successfully run a total of 6 different OS installs which includes four different versions of Windows (95/98/2k/XP) from multiple drives. I just need a new platform on which to try to build a multi-boot system.