• Please review our updated Terms and Rules here

XT-IDE via floppy boot instead of EPROM?

AlexC said:
But if I don't have the XT-IDE card yet I do have an old 16-bit ISA IDE controller lying around, can I get the same effect through software?
Click to expand...

Yes, ATA-2 8-bit on 16-bit controller in 8-bit slot is a supported mode :) Not all controllers work but there is a post on here somewhere detailing why and how to make them work (covering a pad, I forget which). Many controllers will work.

The key however is that only 8-bit transfer mode supporting devices can be used, i.e. ATA2 HDDs, Microdrivers, CompactFlash cards.

HTH
 
AlexC said:
What system are you running it on? What types of drive? NASM will fail silently but PARTCOPY should throw errors if it can't find the file or the destination device.
Click to expand...
I ran it on a PII and a 486DX2-66 with the same results from both. There were no visible errors. But I don't think there were any invisible errors, aka, NASM, because when I booted the floppies on my PII it showed the XTIDE BIOS message and searched for Master and Slave drives... unsuccessfully. It then continued to boot normally.

AlexC said:
Try doing the NASM part on a 386 but then moving the relevant files to an XT to finish the job.
Click to expand...
I'm sure it worked since the PII did post the XTIDE BIOS message.

AlexC said:
Try putting break points (i.e. 'pause' statements) in the batch file and watching which drive LEDs light up while its running. That'll tell you if it's using the correct drive.
Click to expand...
That was ok as I saw the correct LED as the boot sector was being written to the floppy and, again, the PII showed the XTIDE BIOS message.
 
Stone said:
I ran it on a PII and a 486DX2-66 with the same results from both. There were no visible errors. But I don't think there were any invisible errors, aka, NASM, because when I booted the floppies on my PII it showed the XTIDE BIOS message and searched for Master and Slave drives... unsuccessfully. It then continued to boot normally.

I'm sure it worked since the PII did post the XTIDE BIOS message.

That was ok as I saw the correct LED as the boot sector was being written to the floppy and, again, the PII showed the XTIDE BIOS message.
Click to expand...

So the floppy works on the PII but not on your XT?

Once the files have been written, run xtidecfg.com from the floppy *on the target system* then load the IDE_XT.BIN file and try auto-configure. If that doesn't work you may have to tweak things manually for your controller, but it worked for mine.
 
Also, I definitely suggest running the batch file on the target machine if you can. That should eliminate any floppy drive geometry issues.
 
Whatever I do the results remain the same. Nothing on the 486 or XT. I can't really be sure that it's working on the PII either since that machine runs the drive in it normally anyway. :) But the four choice turquoise menu appears at the top so I think it's responding to the 'C' keypress.
 
Stone said:
Whatever I do the results remain the same. Nothing on the 486 or XT. I can't really be sure that it's working on the PII either since that machine runs the drive in it normally anyway. :) But the four choice turquoise menu appears at the top so I think it's responding to the 'C' keypress.
Click to expand...

I don't know what else to suggest. Bear in mind that yesterday I didn't have a clue about any of this. ;)

If anyone else has successes or failures, please post them here. Perhaps we can identify what works and what doesn't.
 
AlexC said:
Yes, ide_tiny.bin is 4KB but that lacks CF support and possibly some other features.

Different BIOS files use different amounts of RAM when loaded this way, so BOOT12.ASM appears to have a fixed overhead of around 4KB. That may or may not be necessary. When I have time I'll investigate it.
Click to expand...

Done. Changing IMAGE_KB in BOOT12.ASM from 12 to 8 before compiling it allows the 8KB IDE_XT.BIN to fit exactly, so only 8KB of RAM is used. It seems to work OK.
 
Thought i'd have a play but it's not working for me, I get the same as 'Stone' 'Invalid Disk' every time, I tried 2 different 16-bit IDE / Floppy controlers which i know work in an 8-bit slot using the XUB and CF, Tried the old Beta 3 download and the most recent R588, Ran everything from my IBM XT 5160 except for NASM but no joy.
 
Malc said:
Thought i'd have a play but it's not working for me, I get the same as 'Stone' 'Invalid Disk' every time, I tried 2 different 16-bit IDE / Floppy controlers which i know work in an 8-bit slot using the XUB and CF, Tried the old Beta 3 download and the most recent R588, Ran everything from my IBM XT 5160 except for NASM but no joy.
Click to expand...

Hmm. Which version of DOS? I'm using IBM DOS 5.00.
 
I just tested this on my Panasonic Sr Partner. MAKEBOOT.BAT fails because PARTCOPY throws an Abnormal program termination.

Testing this machine with the boot floppy I created on my other XT-class box, I get the same error as you do. This machine doesn't have an IDE controller in it, but I think the ROM should still load. Then again, even a V20 won't work in this Panasonic. It's a bit quirky.

Hmm. This works fine on my no-name turbo XT 8088 board. Odd.
 
Malc said:
Thought i'd have a play but it's not working for me, I get the same as 'Stone' 'Invalid Disk' every time, I tried 2 different 16-bit IDE / Floppy controlers which i know work in an 8-bit slot using the XUB and CF, Tried the old Beta 3 download and the most recent R588, Ran everything from my IBM XT 5160 except for NASM but no joy.
Click to expand...
Hey... I got that error with a 16-bit controller in a 16-bit slot on a 486!!! :) This seems to be really picky. Compared to this, getting a USB flash drive to work in DOS was a snap!
 
Stone said:
Hey... I got that error with a 16-bit controller in a 16-bit slot on a 486!!! :) This seems to be really picky. Compared to this, getting a USB flash drive to work in DOS was a snap!
Click to expand...

I had been using the floppy port of the 16-bit IDE controller in my 8088, but swapping back to the original 8-bit multi-IO card still works fine, so it's probably not a floppy controller issue (though the 8bit controller is quite 'new' as these things go: it has an RTC on it).

The only other thing I can think of is some floppy drives work with this boot sector and some don't, but that's only a guess. At the moment it's working for two people (me and the VOGONS guy) and not for two others.

So take your chances, folks: 50/50...
 
Malc said:
MS DOS 6.22, I'll Have another go later.
Click to expand...

I now suspect it won't make a difference, since I can't get the working floppy to boot on my other XT-class machine.

Someone with a better knowledge of FAT12 boot sectors might be able to work out what to change in BOOT12.ASM, as I suspect the problem is there. The VOGONS poster couldn't get it to boot in a PC emulator, though it did work on his 8088 and mine.
 
Chuck(G) said:
I think it's worth noting that a "bare" BIOS for this shouldn't be more than 4KB--the WD MFM controllers initially used 32Kbit (2732 type) ROMs.
Click to expand...

I think you're comparing apples with oranges, Chuck. FWIW, it is possible to create a BIOS with a size anywhere between ca 3500 to 9127+ bytes depending on build options (with my current work-in-progress-soon-to-be-r589 code). But the great thing about this (being able to load the BIOS from floppy) is that people can test and evaluate XTIDE Universal BIOS with whatever features they want before acquiring any hardware (EEPROM programmer, EEPROMs etc). It could also be useful for debugging purposes. I doubt anyone would want to run it from RAM like this on a permanent basis though, regardless of BIOS size.
 
AlexC said:
This version doesn't work on my system. The BIOS loads and finds the CF card, but then says "Error loading OS" and the system hangs. I'll stick with 2.0.0 beta 3.
Click to expand...

You're getting that error because the drive geometry translation has changed between these versions. If you're happy with beta 3 then that's great. Anyone starting out with an unpartitioned drive should use the newer version though.
 
Krille said:
You're getting that error because the drive geometry translation has changed between these versions. If you're happy with beta 3 then that's great. Anyone starting out with an unpartitioned drive should use the newer version though.
Click to expand...

Ah, interesting. That makes sense. Beta 3 worked straight away with a DOS-bootable CF card that I'd already prepared on a different machine, though right now I can't recall which machine that was. So to use 588 I'd need to re-FDISK? Easy enough to do, then.
 
Krille said:
But the great thing about this (being able to load the BIOS from floppy) is that people can test and evaluate XTIDE Universal BIOS with whatever features they want before acquiring any hardware (EEPROM programmer, EEPROMs etc). It could also be useful for debugging purposes.
Click to expand...

I agree, it would - if it was reliable across machines. Why it's not is beyond me, unfortunately.

Krille said:
I doubt anyone would want to run it from RAM like this on a permanent basis though, regardless of BIOS size.
Click to expand...

I might. I've gone from a 21MB ST-225 to a 64MB CF-IDE and all I had to do was add an old ISA 16-bit IO card and leave a floppy in drive A:. Since this machine always checks the floppy drive on boot anyway, there's no time overhead: all I've really 'lost' is 8KB of conventional RAM.

I realise this has fewer features than the proper XT-IDE solution and I may go down that route for one of my other machines, but for this one it's fine.
 
You must log in or register to reply here.
Back
Top