XT-IDE tech Support

[Lutiana]

So here is a new thread for anyone that is beta testing the 8 Bit XT-IDE card, either Rev 1 or Rev 2.

Posting the issues here will help keep the development thread focused where it needs to be.

 

[per]

Speaking of my experiences with the rev 1 card:

I generally don't have issues with mine unless I use multipile drives. However the problems I encour when using two drives are noise issues due to not having grounded wall adapters availble, and not using an 80-wire cable. If you face data corruption problems, always try with an 80-wire cable first before reporting here.

By the way, I did not have any problems using this card with the Orchid PCTurbo 286 accelerator card either. The combination works great, and transfer speeds are quite a lot faster. I had to change the I/O port for the card because 286 board used I/O port 300h.

One last note, if anybody finds any bugs in the utilities package, please drop me an PM. I will do wathever I can to fix them ASAP get some reported.

 

[hargle]

Post moved from the development thread...

Interesting. So what happens if you try and do a "dir a:" from your C: after booting fails? Does it just grind and ARF, or do you get a "drive not ready?"
When I boot to the hard drive, I have full access to both of the floppies in DOS.

Now that's weird. You can't boot to it, but once you're booted to the C: drive, both floppies read just fine. The only thing that keeps the system from booting off the A: drive is if there's a read error. How can there be a read error during POST but not after boot? (I'm not expecting you to answer, just thinking out loud)

"not being able to get to the hard drive" means that typing "c:" and it said invalid drive specification? I'm thinking another simple utility to do some diagnostics here, like checking 40:75 for the hard drive count, then trying to read the 1st sector off the drive and displaying the data/any error codes could be good to know.

Yes, says "invalid drive specification"

How about if you type Fdisk at the A: prompt? Does it show that you have a partition, or does it say the drive is empty, or perhaps it says "error reading fixed disk" or maybe "no fixed disks present"?

I'd really like a dump of that ROM. Even just to satisfy my curiosity, but I want to see what it is doing to INT 13 when it hooks in.
OK. Does that previous file for pcjr actually dump the ROM too?

mike's pcjrdump program works on all machines, and dumps every rom it finds in your system.

I've just finished up a small augment to me I13dump utility,which might help us diagnose this. I need to test it tonight on my machine before I release it.

 

[kb2syd]

Did my first XT-IDE build. Effortless. Finished in about an hour. That included finding my cut off snips. I left the "eprom write enable jumper" off due to past experience.

Put it in the 1000sx since that is what is on the bench. Booted and recognized the 3rd drive I tried. The first 2 were less than 500MB each so I wasn't suprised. The 1.7GB drive had windows 98 on it so it obviously couldn't boot.

Booted with 5.25" MS-DOS 3.3 disk, and couldn't kill the partition. Found a debug script on microsoft.com and killed the entire partition table. FDISK then sees it (as having 817 cylinders). Added primary partition and made it active. Tried to add a secondary partition and the machine locked.

Booted from floppy and formatted C:. Copied dos on to it, and rebooted. Everything fine. Then tried to add an extended partition. Locked up again, and primary partition was hosed. More when I know more, but this is real progress for me!

 

[bugman2112]

How about if you type Fdisk at the A: prompt? Does it show that you have a partition, or does it say the drive is empty, or perhaps it says "error reading fixed disk" or maybe "no fixed disks present"?

"no fixed disks present". But findcard and such still detect the card.

 

[bugman2112]

Found the bug, the checksum byte still wasn't written to memory. SO, at the version I just uploaded, the checksum is corectly generated, AND it is also written to memory.

However, it's not tested since I don't got the ability rigth now (I have my fall break this week), so please test it for me

.

When I used your new setcard to change a boot setting (like delay, or firstboot), I still get a verify error. "unexpected byte found at the following address: D000:1FFF 92<>43" Then upon reboot, I get the "D000 rom error, press f1 to continue" and the hard drive is disabled. Also, using setcard with the /B switch results in "D000 AA55Error, Signature not found!"

I tried several different combinations with setcard to no avail. The only way to regain a working system is to reflash from scratch. For clarity, this is my functional setup using the IBM floppy controller. Except for not being able to use setcard, the xtide seems to be working.

 

[Lutiana]

Booted from floppy and formatted C:. Copied dos on to it, and rebooted. Everything fine. Then tried to add an extended partition. Locked up again, and primary partition was hosed. More when I know more, but this is real progress for me!

That may be a DOS 3.3 issue. Try do the same thing with Dos 5 or Dos 6 and see if it does the same thing.

 

[kb2syd]

That may be a DOS 3.3 issue. Try do the same thing with Dos 5 or Dos 6 and see if it does the same thing.

I don't have a Dos 5 or 6 boot disk on 5.25" media. Guess I'll hook up a 486 to do the formating. I would really prefer to use DOS 3.3 because of the memory footprint. Actually, for the Tandy's I'd like to stick with Radio Shack's DOS 2.11, but that is having problems with partioning also.

Does anyone have 3.3 working on the XT-IDE? I'm getting a lot of mixed results after partioning, and am trying to figure out if it is 3.3 or build issues.
FDISK partitions OK.
Format /s/v runs without error.
Reboot, and after Press [Esc] for boot menu I get the drive enumeration again, a string of non-sense characters, then a 6 digit number (598621) repeated 2 times.

Anyone else trying this with DOS 3.3?

 

[hargle]

Reboot, and after Press [Esc] for boot menu I get the drive enumeration again, a string of non-sense characters, then a 6 digit number (598621) repeated 2 times.

The drive ID display shows up properly in the main POST screen, but is a bunch of garbage when you see it in the boot menu? Or is it after the drive ID string that this garbage shows up? It's supposed to look like this:

***************************
*      Boot Menu          *
***************************
* A: floppy disk drive 0: *
* B: floppy disk drive 1: *
* C: ST31240ACE           *
* D: ROM BASIC            *
***************************
*  Press A-D to Select    *
***************************

Where C: would be the same drive model # text string as displayed during POST when it's trying to locate what drives you have available. If you have a D: drive, that'll be in there too, and BASIC will be selection E.

If there's *any* garbage in there, there is something wrong. I would at that point think that you may have some noise invading your signal. Are you using an 80 pin cable? What is the drive model that you're trying to use?
I *strongly* suggest using a drive not much older than about 10 years. I've had really good luck with drives made at least in this century.

 

[kb2syd]

The drive ID display shows up properly in the main POST screen, but is a bunch of garbage when you see it in the boot menu? Or is it after the drive ID string that this garbage shows up?

Guess I wasn't clear. The POST ID is perfect, the Boot menu is perfect. But then the POST ID shows up again after selecting 'C' as the boot device. This shows up correctly, but in the middle of a bunch of garbage characters (spade club heartl, etc).

If I boot from floppy, the C drive is fine, this is only if I try to boot from the drive.

If there's *any* garbage in there, there is something wrong. I would at that point think that you may have some noise invading your signal. Are you using an 80 pin cable? What is the drive model that you're trying to use?
I *strongly* suggest using a drive not much older than about 10 years. I've had really good luck with drives made at least in this century.

I'll try an 80 pin cable. All my 80 pin cables are keyed so I'll have to snip off the offending post on the 40 pin header.

The drives I've tried are around 1.7 to 2.2 GB in size, and may be from last century. They were a mix of WD and Seagate drives.

If snipping the post and going to an 80 conductor cable doesn't work, I'll try something newer. This was my next step anyway.

If that doesn't work, the NEXT step will be to boot with it in my 486, format and fdisk with dos 6.22 and see if that helps the partitioning problem.

 

[hargle]

Guess I wasn't clear. The POST ID is perfect, the Boot menu is perfect. But then the POST ID shows up again after selecting 'C' as the boot device. This shows up correctly, but in the middle of a bunch of garbage characters (spade club heartl, etc).

No idea what that could be. The boot menu is the very last thing that the XTIDE BIOS is in control of, so things that show up after that are outside of our control. My guess as to what is happening here is that the the XTIDE starts to boot your C: drive, something somewhere there is corrupt, and it is causing something to spew garbage characters to the screen.

It may be wise for me to add a clear screen as very last thing that the XTIDE does before starting to boot, thus making a better distinction between what is "us" and what is "them".

If that doesn't work, the NEXT step will be to boot with it in my 486, format and fdisk with dos 6.22 and see if that helps the partitioning problem.

Good plan. A more modern drive, DOS 6.22 (or 5.0) and a wiping of the drive before fdisk/format should help you get to the same spot as the rest of us. That's a good starting point anyway, so you can build up some trust of the controller before doing anything more unique. I think adding a drive wipe utility to our collection of utilities might be a good idea; that would wipe out all those partition manager type softwares, since they are no longer required.

I would also suggest an 8+ gig drive and an 80pin cable. There's a bug report that a 2g drive wasn't sizing properly, so there is something fishy somewhere, but larger drives should be ok.

 

[kb2syd]

OK, same drive, same controller, same cable. Works perfectly under dos 6.22, not dos 3.3 or 2.11. That's on a 1000sx with 640k RAM.

Is anyone running this with earlier dos versions? What could be the problem with the BIOS and the partiton table? Or, could it be drive geometry?

 

[Agent Orange]

OK, same drive, same controller, same cable. Works perfectly under dos 6.22, not dos 3.3 or 2.11. That's on a 1000sx with 640k RAM.

Is anyone running this with earlier dos versions? What could be the problem with the BIOS and the partiton table? Or, could it be drive geometry?

I'll be running Tandy MS-DOS 3.2 (oem) on my 1000SX as soon as I can get the board soldered up. Should be Wednesday or Thursday if I don't have problems. I'll be using a Maxtor 7546AT (550MB) and a Maxtor 90845DA (10GB) for the 1st go-around. The next step will be a 80GB Western Digital. I'll post the results ASAP.

P.S. I'm excited about the whole thing.

 

[Lutiana]

OK, same drive, same controller, same cable. Works perfectly under dos 6.22, not dos 3.3 or 2.11. That's on a 1000sx with 640k RAM.

Is anyone running this with earlier dos versions? What could be the problem with the BIOS and the partition table? Or, could it be drive geometry?

I think that you are running into a limitation in DOS 3.3 or an incompatibility, try looking up what DOS 3.3 can and cannot support.

 

[kb2syd]

I think that you are running into a limitation in DOS 3.3 or an incompatibility, try looking up what DOS 3.3 can and cannot support.

Should it matter what dos can support? I'm using 3.3 FDISK, booting fine, making a 32 MB partition, formatting it with system, and rebooting.

Everything works up until the reboot. On reboot, instead of reading the partition, the machine barfs up junk.

I thought 3.3 would work on a drive regarless of geometry if you only told it about the 32MB limits of it.

I'll try 3.3 on my 486 with a 2 GB drive. If it doesn't work there then I know what's going on and I'll have to live with generic dos 6.22 on my Tandy machines.

 

[hargle]

OK, same drive, same controller, same cable. Works perfectly under dos 6.22, not dos 3.3 or 2.11. That's on a 1000sx with 640k RAM.

whew. I was really starting to suspect your house was built on top of an old cemetery or something. you've had such a string of bad luck with these cards that no one else seemed to be having...

Anyway, at least we've eliminated a majority of the issues; hardware seems to be working just fine, now we need to focus on what is causing old DOS versions to freak out. That may be a good thing for the wiki, sort of a "getting started" section. After building the card, first thing you should do is try a 2Gig or higher hard drive, with an 80pin cable, with DOS 6.22 to make sure you're at the same level of functionality as everyone else. then, and only then, start exploring new options, and report back...

DOS 3.31 might also be another option to explore. Don't I recall somewhere that they did a quick hard drive upgrade/fix in there?

 

[Lutiana]

Should it matter what dos can support? I'm using 3.3 FDISK, booting fine, making a 32 MB partition, formatting it with system, and rebooting.

Everything works up until the reboot. On reboot, instead of reading the partition, the machine barfs up junk.

I thought 3.3 would work on a drive regarless of geometry if you only told it about the 32MB limits of it.

I'll try 3.3 on my 486 with a 2 GB drive. If it doesn't work there then I know what's going on and I'll have to live with generic dos 6.22 on my Tandy machines.

My apologies, I did not realize the issue was deeper than just a partitioning thing.

It does sound like it is something to do with DOS 3.3 since it works on DOS 6.2.

Try creating a 32Mb partition with DOS 6.2 and then formatting and installing DOS 3.3 onto it.

 

[kb2syd]

whew. I was really starting to suspect your house was built on top of an old cemetery or something. you've had such a string of bad luck with these cards that no one else seemed to be having...

Ha! I was thinking the same thing for a while. Rev 2 is a much better board. Build time was NIL.

There is a definite compatability issue somewhere. I installed a WDAC21700-40H (1.7 GB) drive in my 486 and booted with a generic dos 3.3.

Partitioned it into C:\ Primary and D, E, F, G in an extended partition. All 32 MB.

Moved the drive over the 1000sx, and it booted perfectly. Installed 3.3 and off we jolly well go!

So, with the XT-IDE in the 1000sx, partitioning under 2.11 and 3.3 does not work. It does recognize and boots correctly once everything is created on a newer machine (with 3.3). I'm going to try 2.11 next if I can find something on 3.5 that'll boot to that version.

UPDATE: Tandy MS-DOS 3.02.2 for the 1000sx does not seem to handle large drives at all. I can live with 3.3 if I can't figure this out. Partitioning seems to work, but then it all falls apart when I try to boot. This is even from the 486 machine.

I did get a 3.5" drive working to boot the 1000sx. Get a lot more data on a single drive.

 

[Lutiana]

So we need to get someone to try the same thing on a true blue IBM XT with DOS 3.3 to see if it is specifically a Tandy issue.

 

[nathan]

I will be able to do so as soon as my kit comes in and I get it soldered up. 5150 here with DOS 3.3.