Jed118 wrote:

So no matter what I try, I can't get the SD card to make a 2Gb partition using DOS. I tried Windows 98 boot disk and DRDOS 7.03, and still nothing...

Fast seach in Google:
http://www.vcfed.org/forum/showthread.php?280 … -scsi-card-bios

TL;DR: Your SCSI2SD is fine. Is your card who exhibit certain onboard program bugs.

Remeber SCSI isn't IDE. These things do certain stuff as drive tranlation different than their ATA counterparts, as part of the logic is in the HBA. Certain older cards, in this case Adaptec, exibit troubles in the SCSI drive translation onboard microcode or BIOS. The only way to fix is is searching for the EEPROM versions which don't have the bug, and update your ASPI drivers for non bootable secondary disks.

I was actually going to go through my drivers, as the one I used is causing some issues (it was on a double density diskette... quite old) but some more digging yielded this:

http://www.walshcomptech.com/ohlandl/SCSI/AHA-1640.html

The Adaptec Boot BIOS is
"old fashioned" and has a 1GB-limit.
It is principially the same as used on the ISA AHA-1540B (not -C or -CF) and has no CTRL-A support and uses the DOS-compatible HD translation, which is a bit "register-width challenged".
You need to LL-Format the drive with Adaptecs SCSIFMT ... which requires a boot disk with the ASPI4DOS.SYS to have the drive under ASPI-Manager control. The ASPI Manager is capable to handle drives over 1GB - the Boot-Bios isn't.

Ding ding ding - It appears it is the controller. And it appears I can use the second option mentioned since the drive is not bootable to begin with.

Thing is, I have both a 1540CF and an older 1542...

Here's some more conversation with the OEM:

I played around with it some more and found an interesting bug. Turns out that if you have the downloaded monitor software worki […]
Show full quote

I played around with it some more and found an interesting bug. Turns out that if you have the downloaded monitor software working alongside the utility, it will almost always fail a write and only sometimes read the information off the host (SCSI2SD unit), which much of the time is garbled.

21:37:31: Save Complete. Monitoring not running - Several tests were performed, SD card in and out (various SD cards)
21:37:33: SCSI2SD Ready, firmware version 4.7.1
21:37:33: SD Capacity (512-byte sectors): 0
SD CSD Register: 00000000000000000000000000000000
SD CID Register: 00000000000000000000000000000000
21:37:36: Saving configuration Monitoring on
21:37:36: Programming flash array 1 row 187
21:37:36: Programming flash array 1 row 188
21:37:37: Programming flash array 1 row 189
21:37:37: Programming flash array 1 row 190
21:37:37: Error writing flash 1/190
21:37:37: Save failed

Then I tried it some more with the card inserted and no monitor software running, and occasionally I would get errors (the log got corrupted but it had a whole mess of 00 addresses then the occasional ff would show up.) I tried to duplicate it a few times and it did not come back. I turned on the monitoring software, and it did a load but I got literally nothing in any of the fields in the SCSI2SD-UTIL program (looks like everything had a zero value, similar to the loadpass(but garbled).jpg but this time with blanks almost everywhere). In the monitoring software, this kept repeating:

SD CSD Register: 00000000000000000000000000000000
SD CID Register: 00000000000000000000000000000000

Stopping the software (UTIL and MONITOR) and restarting only UTIL I was able to get the attached XML (I'm not too sure I want to have SCSI2 on - I have internal peripherals such as a Syquest 88Mb media reader, a 4X CDROM and a ZIP100 disk, all on SCSI - having said that when tested on my 486/66 it was just the controller and the SCSI2SD unit and the results were exactly the same) and mostly trouble free loading.

I'm using three controllers to test this:

ISA Adaptec AHA 1540B (BIOS BC00) 1990 (under DOS 6.22)
ISA Adaptec AHA 1540CF (BIOS C380) 1993 (DOS 6.22)
PCI Adaptec 2940UW BIOS 8800 REV C 1996 (Windows XP)

... and I'll be Goddamned if the PCI controller doesn't see the 2Gb partition under MMC -> Disk Management snap-in.

And after some mild annoyances in 486DLC configuration land, I was able to boot off a primary IDE drive with the D:\ drive SCSI2SD reporting 2 gigs in DOS 6.22 (Both AHA 1540s worked) - It's doing a sector-by-sector analysis now.

I'm going to run more tests here on this end, but for now this was either a BIOS or OS issue, or both.

So the solution for me was to recall that I had a pair of 2940UW PCI cards lying around, to plug one into an XP machine, and run Disk Management and forcing a FAT partition (2gb). I tried this on two SDHC cards I had lying around and it worked on the AHA 1540XX controllers in the 486DLC machine I have. If you try to make a larger partition in the SCSI2SD integrated controller, and then format the card in FAT32, not surprisingly DOS will not even see the partition. If you FDISK in and delete/create a new FAT partition, it will show as.... 1 gigabyte. Both of these AHA 1540s will see and work in a 2 gig partition, but will not make one. The 486DLC has absolutely no drivers loaded, ASPIxxx.SYS is not present, DOS loads strictly the HIMEM.SYS driver and that's all, so I'm not sure if (how?) it could be loading Extended BIOS Translation. Supposedly the 1540B is REV B and the 1540CF is REV C BIOS. The former isn't supposed to see past 1 gigabyte (no Extended BIOS), and the latter is purported to have EXT BIOS, yet both can't create a 2 Gb partition natively. Both can, however, see one prepared by way of another card (or O/S if I'm honest).

I'm curious what other DOS based software I could try on my 486 that might have some success in creating the full 2gb partition. Or perhaps I will try booting DOS into the XP machine and running FDISK. Maybe I'll even open up my P166 or when I test Gahhhrrrlic's RTC mod I'll play around with both cards on an ISA/PCI board: I want to figure out if this is a card limitation.

Strange, either way.

Cards:

There's gotta be an easier way:

It works though:

A better way to fix this issue is to CTRL+A on the SCSI BIOS prompt and enable >1 GB disk reading. This works only on the REV C BIOS. (C) 1993 card I have. If you put REV B in there, it will only accurately read <1 gigabyte partitions.

On the 1540B card, what happens is you fill up the first 1 GB then try to write to the second GB of drive real-estate? Im not sure, but I have a feeling the data will get garbled up or you will receive errors unless you use the C BIOS.

It's hit or miss - I tried this by first booting a 125Mb drive (IDE) with DOS 6.22 on it and copying some files onto the D:\ SD "disk".

I then tried to run those files off the SD "disk" - DOS files ran fine, Windows 3.1 crashed.

I then powered down and put a 528Mb drive with DR DOS 7.03 on it and.... there was ONE file readable on the SD "disk"

I formatted it (it accepted the 2.1Gb limit imposed by the FAT), it stopped at 100% and said there was a general failure. From the C:\ prompt I got into D:, but a DIR listing showed "Sector Not Found" error.

I then FDISK the drive and recreated the partition at... 1GB.

Basically it's not reliable if I force a 2 Gb partition created on a C rev onto the B rev.

Essentially, hyoenmadan was correct 😉

How about buying a 1540C or 1542C ? I bought one a few months ago. The price was OK, cheaper even so if you live in the US.

I just got a 1540CF for $15 off eBay, so I was able to test it. REV C FTW!

Honestly, new computers are SO much easier to configure. The last time I had SCSI was on a PIII 1.266 (Or 1.3? I can't remember) and it was an Adaptec 29XXUW (I recall using the 68 pin variant) so jumping back into AT SCSI from the early 90s seems to be creating a steep re-learning curve.

I'm going to be beta testing the new SCSI2SD card, so my efforts and tinkering (and logging) have paid off... Not that I require payment, it's just good to give myself a challenge once in a while 😉

Jed118 wrote:

I just got a 1540CF for $15 off eBay, so I was able to test it.

Maybe I missed something in the transcript. What was the outcome of your testing it? 2 GB no problem now?

So far it works fine - I haven't really been using it all that much due to the everproject nature of the computer it is in, but it's gotten a couple hours of use without issues.

Are you daring enough to try the C BIOS in the B card? I recall someone may have mentioned that the hardware was slightly different.

You mean, swap out the BIOS chips? Actually, yes, if that works, then I can have them flashed somewhere. Good idea!

Let me enjoy it first for a little - I've made tremendous progress with the solid-stating of my drives in the 486DLC, I'd like to document it.

Jed,

What was the outcome of this?

I have an Adaptec AHA-1542C and the SCSI2SD. When I attempt to boot from the SCSI2SD, even though dos is installed correctly, it hangs unless I enable the floppy and boot from a floppy disk. Then, the Hard drive shows up fine. See photos.

Any hints?

Jed118 wrote on 2018-03-30, 15:12:

Let me enjoy it first for a little - I've made tremendous progress with the solid-stating of my drives in the 486DLC, I'd like to document it.

Update, in my case, I had to run FDISK /mbr after booting from a dos disk. Even though I had partitioned the disk before, there was no master boot record. I'm going to write an article about my process for posterity sake. Thank you all for your help.

Sometimes you need to initialize the drive with the SCSI card (I found at least that to be partially true with 68 pin SCSI) but I have been able to move drives over among the same controllers without too many troubles.