What kind of SCSI adapter or chip does it use?

Disruptor wrote on 2023-07-04, 21:07:

What kind of SCSI adapter or chip does it use?

Integrated SCSI (Adaptec ASW-B626 - chip Adaptec AIC-6260AL)

This is actually the best reference on the built-in chipsets.

https://mastodonpc.tripod.com/personal/4253.html

Seems like you need ASPI2DOS.SYS
mkarcher and me never have checked controllers like the AHA151x or AHA152x for large disks - like he did with 274x and 284x and ASPI7DOS.SYS

Disruptor wrote on 2023-07-04, 22:17:

Seems like you need ASPI2DOS.SYS
mkarcher and me never have checked controllers like the AHA151x or AHA152x for large disks - like he did with 274x and 284x and ASPI7DOS.SYS

I am booting off of the drive, so with ASPI2DOS.SYS and the appropriate DISK.SYS it says that the disk is under BIOS control. At least that is on DOS 6.22. I haven't dried putting a secondary SCSI drive on there to see, mostly because there isn't space inside the case for it. If I am feeling froggy I will get a SCSI2SD set-up to see if disks loaded after that have < 1GB workable. That may able to be in there, and can at least get it working.

One of the reasons for using SCSI is I was hoping to not have to deal with legacy IDE limits.

Warp seems to have drivers for it, but I haven't tried any advanced partitioning software. My hope was to get the kernel as up-to-date as possible, I think with the Warp 4.5 kernel was back-ported in a Fixpack. I also haven't looked much into the Warp config.sys.

Perhaps you can try whether the limitations just apply to the boot drive - try to connect a second drive then (for testing purpose: first try a 4,5 GB drive, then 9 GB or larger).
Probably your WARP driver doesn't have those limitations.

Disruptor wrote on 2023-07-04, 22:32:

Perhaps you can try whether the limitations just apply to the boot drive - try to connect a second drive then (for testing purpose: first try a 4,5 GB drive, then 9 GB or larger).
Probably your WARP driver doesn't have those limitations.

Warp 3.0 Connect will recognize the full disk, but is not stable to install or partition over 1GB. It will throw disk errors when trying to install. I've tried with and without the boot loader. At the moment I delete all partitions with a Windows 98 boot disk, then do partitioning of a clean disk with OS/2 at > 1024M and it works fine. FWIW, I would hope the Windows 98 disk would support the max, but it doesn't see it as being larger than 1024M.

Will try with a second disk once I get SCSI2SD or something similar working. I think I can fit that inside the case without an issue.

FYI, an alternative to scsi2sd is zuluscsi

weedeewee wrote on 2023-07-04, 22:44:

FYI, an alternative to scsi2sd is zuluscsi

That's what I may try next. My SCSI2SD was an older version, and with the firmware updated it was unstable. Drives configuration would vanish whenever the SD card was removed. Problem could have been between computer and chair. I think that considering the form-factor and space involved, I may use the Centronix connector and go with that form-factor for disk expansion.

It's also "enough" disk, but I would love something that could allow for solid large-file transfer that doesn't involve splitting files onto virtual floppies, or burning physical CDs.

un1xl0ser wrote on 2023-07-04, 22:40:

Warp 3.0 Connect will recognize the full disk, but is not stable to install or partition over 1GB. It will throw disk errors when trying to install. I've tried with and without the boot loader. At the moment I delete all partitions with a Windows 98 boot disk, then do partitioning of a clean disk with OS/2 at > 1024M and it works fine. FWIW, I would hope the Windows 98 disk would support the max, but it doesn't see it as being larger than 1024M.

Will try with a second disk once I get SCSI2SD or something similar working. I think I can fit that inside the case without an issue.

I guess the problem is the BIOS support.
If it has problems with partitions over 1 GB it will be more likely that systems that rely on BIOS support (like Win98' DOS portion) will have a problem. So I recommend to create all partitions for DOS (FAT16) to be completely within first 1 GB - data exchange 😉. For a Win98 FAT32 partition you may try to place it that the first sectors are within first 1 GB; if you do not boot from it it may pass the border. The HPFS partition may be completely after first GB.

Perhaps you can extract the controller BIOS or even update it. Perhaps you'll get access to disks > 1 GB or even INT13E support.

Ye, SCSI2SD is a nice thing but seems to be too slow when used in Versions below 6 where it hits almost 10 MB/s. And the mentioned zuluscsi also lacks of Ultra, LVD or Wide support.

Your controller has an AIC-6260AL, my controller has an AIC-6370Q
My 1520B BIOS neither supports disks > 8 GB with BIOS 1.15 but > 1 GB is no problem.
The required ASPI2DOS.SYS in its newest version 3.68 supports 6260, 6360, 6370. It also does not support disks > 8 GB.
The 32 bit driver from Windows 95 B makes no difference.

So you may have a chance at least to get disks < 8 GB but > 1 GB working.

You may try to boot DOS from a floppy and run FDISK and tell me what size it reports. Perhaps you can run SPEEDSYS 4.78 too and tell me which CHS translation it reports for a disk > 1 GB and < 8 GB.
Then try the same with ASPI2DOS.SYS loaded (ASPI2DOS should report that it routes INT13 through ASPI driver).
So we can see if perhaps the controller BIOS is the problem...
Perhaps you can dump it with debug.
First you have to find the location of the BIOS. when you dump with d you quickly will see Adaptec strings. Try to start from c800:0, and increment in 16k steps: cc00:0, d000:0, d400:0 and so on
When you have found it, continue with the further strings.
Run from DOS prompt:
debug
dc800:0
rcx
4000
n adaptec.rom
wc800:0
q

I have a few items with the AIC-6260AL (two motherboards and a 1520 plain) and none of them support larger than 1GB HDD iirc.
One of the motherboards is similar to OP but is a Unisys PW 300/486 25-50, it has built in IDE also.
Attached the bios just for reference, the scsi bios part is (hex read): Adaptec ASW-B626 BIOS...Version 1.0. AHA-1520 1990 Version 1.02.
I added the jumpers sheet and picture from 1994 here: http://www.vogonsdrivers.com/getfile.php?fileid=1582

Horun wrote on 2023-07-07, 00:45:

Attached the bios just for reference, the scsi bios part is (hex read): Adaptec ASW-B626 BIOS...Version 1.0. AHA-1520 1990 Version 1.02.

Looks like a 64 KB All-in-one Phoenix BIOS. No extra SCSI BIOS.

Yes it is an All in One with WD90c11 video, main 486 board and adaptec 1520 bios all wrapped together.....

Disruptor wrote on 2023-07-05, 20:32:

So we can see if perhaps the controller BIOS is the problem... Perhaps you can dump it with debug. First you have to find the lo […]
Show full quote

So we can see if perhaps the controller BIOS is the problem...
Perhaps you can dump it with debug.
First you have to find the location of the BIOS. when you dump with d you quickly will see Adaptec strings. Try to start from c800:0, and increment in 16k steps: cc00:0, d000:0, d400:0 and so on
When you have found it, continue with the further strings.

Adaptec ASW-B626 S2 BIOS
Version 1.0
Copyright 1991 Adaptec, Inc.
All Rights Reserved

The provenance of the drivers is Adaptec EZ SCSI 5.0c for DOS / Windows 3.1. ASPI2DOS.SYS is version 3.86 from the attached screen-shots.

I created 2 x 4GB drives with BlueSCSI, made them targets 1 and 2. They are recognized fine. They have been low-level formatted with SCSIFMT and passed validation to the full 4GB.

It says no logical drives though, and the DOS 6.22 FDISK seems lame, or I am missing switches. Will be spending more time to RTFM, but if anyone has any pointers, it would be appreciated.

So I still have something to learn about DOS and partitioning multiple disks. I'm still going to try to figure out the right way of doing this. I'm also going to dig out some utilities that I had that showed the CHS and other details of the disks especially the virtual ones.

In OS/2 WARP 3.0 with XR_W040 applied, I am able to select the different disks and do partitioning. The aforementioned controller works fine with a 4GB disk and a single partition with HPFS. I tried to create a 2G and 1G partition formatted with FAT, but ran into a cylinder limitation of 1023 for FAT (at least as far as OS2 disk). I was able to make a 512M and partition it, but rather actually figure out the geometry that BlueSCSI is reporting.

So the controller will support 4GiB drives, and I will push the limits on OS/2 and test if everything is accessible/stable. I'm hoping to be able to get a pair of 16G drives working under OS2 and being able to keep a collection of low-end DOS and Windows games on the system.

Ye, in my standard partition scheme for retro computers I place my legacy partitions to end no later than at cylinder 1021.
Using my larger disks I use this rule:
My FAT32 Windows 98 SE partitions start at cylinder 1022 (not at 1023, because some BIOSes report 1022 as last cylinder instead of 1023 - DOS' FDISK shows 8025 MB instead of 8033 MB then).
If you can place the start of your OS/2 boot partition at cylinder 1022 or before, you may try to go beyond that limit with that HPFS partition. Does it work?