• Please review our updated Terms and Rules here

XT-IDE not recognizing drives correctly

C

Chris Hafner

Member
Joined
Jan 4, 2012
Messages
42
Location
Seattle, WA, USA
So, I picked up a 16-bit ISA NIC containing the XT-IDE BIOS and a 16-bit IDE card from dabone, who was kind enough to make them available to me. I installed both into my machine, but am having issues with XT-IDE recognizing my drives correctly. Here are the issues:

- My CF card and adapter have been formatted as bootable in MS-DOS 5.0, and I used it to successfully boot my Win98 machine. However, the first two times I started up the 5150 with it connected to the IDE card, it was recognized as [diamond][heart][heart][diamond] and XT-IDE didn't see it as bootable. The next two times, it didn't recognize that it was present at all.

- XT-IDE sees my A: drive (a typical 360KB 5.25" floppy) as a 1.44MB drive, and it can't boot off A: either. In fact, the drive just spins forever and the machine goes unresponsive.

I feel like I'm *so close* but I'm not sure how to get over the hump and to get these drives recognized correctly. A search through the forum didn't turn up any obvious answers. Any suggestions?

Cheers,
Chris Hafner
 
are we correct in assuming this combo of cards worked properly when dabone had them? what machine did he have them in?

My 1st suspicion is that dabone was using them in a 286 or better, and thus had the xtide bios configured for 16 bit addressing, which is not going to work so well in a 5150.
 
hargle said:
are we correct in assuming this combo of cards worked properly when dabone had them? what machine did he have them in?

My 1st suspicion is that dabone was using them in a 286 or better, and thus had the xtide bios configured for 16 bit addressing, which is not going to work so well in a 5150.
Click to expand...

Thanks for the help, all.

It's my understanding that the combo of cards was working for dabone. In PMs he's saying he may have put the XT version of XT-IDE BIOS in there, which wouldn't be 16-bit, but may still be driving the weirdness. But there may be more than one problem here.

Here's the full rundown of what I'm seeing upon powerup:

1. Fast cursor
2. Screen says "1040 201", which is not an error I was seeing beforehand
3. Checks A: drive and then B: drive
4. Then this screen comes up:

---

A>>FDD [A] C>>HDD [C] F8 RomBoot

A: ==> 1.44 M DISKETTE DRIVE.
B: ==> 360 K DISKETTE DRIVE.

-=XTIDE Universal BIOS (XT)=- @ D800h
Released under GNU GPL v2

Master at 170h: Not found (sometimes finds the CF as master, but IDs it as [diamond][heart][heart][diamond] and can't find the boot sector)
Slave at 170h: Not found
Booting C>C
Error 1h!
Booting A>>A
Error 80h!

--

Then Drive A just spins, and the computer doesn't respond to the keyboard if I try to select ROM boot, or Ctrl-Alt-Del

But it's not consistent, and weird stuff happens. Like I mention above, sometimes it does see the CF (but IDs it weird and doesn't see it as bootable). Once the drive light on A: came on, but it was B: that was spinning. And in no case is the machine responsive to the keyboard.

There may be two problems here. The fact that this seems to be a 5160 version of the XT-IDE BIOS instead of 5150 might be horking up the A: identification, leading to non-floppy boot and endless spin. But the sometimes-missing CF, the odd identification (diamonds and hearts), and the inability to boot may just indicate that the CF card isn't compatible with XT-IDE. The XT-IDE "manual" appears to indicate that mis-identification of the CF is one symptom of incompatible media.

I'll continue to run down the XT BIOS angle with dabone, but any additional wisdom anybody would be willing to share would be eagerly and appreciatively consumed.

Cheers,
Chris Hafner
 
pearce_jj said:
You following this, presumably?
Click to expand...

I just read through, but am not sure I'm grasping the significance. As I understand it, that describes the ability to place a custom BIOS on a NIC (or an all-in-one card, or an IDE card with the ability to carry its own BIOS), and that BIOS will allow the use of CF on a 16-bit ISA card.

But at the most basic dunderheaded level (which, unfortunately, is where I'm at), isn't that the same as XT-IDE? Right now I have a custom BIOS on a NIC which in theory will allow the use of CF on a 16-bit ISA card. I'm excited about the excitement, as it were, but I'm not totally grasping it.

I am rooting for your Lo-Tech V3 card, though, since having everything in one card with the ability to access the CF without popping the case seems *really* exciting, and I'm hoping I can bribe somebody to build one for me once it's ready. :)

Cheers,
Chris Hafner
 
Not all CF-to-IDE adapters work in 8-bit mode so that might be the problem why the drive does not identify properly (unless the adapter was tested as well). I don't know why floppy drive is identified wrong. It should be identified as 360k or 720k when there is no AT BIOS functions available to properly identify the drive.
 
ok, lots of things to consider.
I don't know what a "1040 201" error is. I'm not finding it in google searches, and i don't have a tech spec handy.
Whatever it is, this is most certainly a good thing to start with, since it will be obvious when we fix it.

I'd do the following:
1) remove all unnecessary cards from the machine: you need only video+floppy.
2) make sure the machine has no error codes and boots to a floppy
3) put the IDE controller in the machine (not the NIC) You do not need to attach a CF or other drive.
3a) does the machine report errors now? Will it still boot to floppy and operate ok? If not, stop and report results here.
4) put the NIC in the machine. This is likely where the fails start adding up because the BIOS on the NIC is now trying to use the IDE controller.

Currently, we do not know what BIOS type is installed in the NIC (XT vs AT). Unfortunately, if the NIC is what is keeping your floppy from operating, we can't even boot it and load up xtidecfg to examine what the settings are. Unless you have an eeprom burner, we may be at an impasse at getting this to work on a 5150.
 
hargle said:
I don't know what a "1040 201" error is. I'm not finding it in google searches, and i don't have a tech spec handy.
Whatever it is, this is most certainly a good thing to start with, since it will be obvious when we fix it.
Click to expand...
It's a Bank 1, bit 6 (the 7th chip) memory failure.
 
Thanks for all the help, everybody - hargle, that seems a sensible set of steps, and I'll follow it when I get home.

You're all right that a 201 is a RAM error, but:
- I never saw it before plugging in the NIC
- a four-digit RAM error code beginning with 1 shouldn't be possible for my 16-64KB machine, as all of those codes begin with 0 (http://www.vintage-computer.com/vcforum/archive/index.php/t-12856.html) - unless I'm misreading it

Is it possible that the error is thrown by the possible wrong BIOS on the NIC? That BIOS (possibly for a 5160) may be looking for RAM in places where it won't be found on a 5150.

Anyway, I'll go through the systematic check when I get home. Thanks again to everybody for their help.
 
Chris Hafner said:
Is it possible that the error is thrown by the possible wrong BIOS on the NIC? That BIOS (possibly for a 5160) may be looking for RAM in places where it won't be found on a 5150.
Click to expand...
nope. Unless the XTIDE Universal BIOS has gone waaaaay beyond just supporting hard drives, it shouldn't be touching or caring about any of your memory.
There is also not a 5150 and a 5160 version of the xtide universal BIOS- it's either AT (16 bit) or PC/XT (8 bit) modes. It's actually the same BIOS, just settings are changed for 8/16 bit modes.
 
Chris Hafner said:
So, I picked up a 16-bit ISA NIC containing the XT-IDE BIOS and a 16-bit IDE card from dabone, who was kind enough to make them available to me. I installed both into my machine...the first two times I started up the 5150 with it connected to the IDE card, it was recognized as [diamond][heart][heart][diamond] and XT-IDE didn't see it as bootable.
Click to expand...

pearce_jj said:
You following this, presumably?
Click to expand...

Chris Hafner said:
As I understand it, that describes the ability to place a custom BIOS on a NIC (or an all-in-one card, or an IDE card with the ability to carry its own BIOS), and that BIOS will allow the use of CF on a 16-bit ISA card....isn't that the same as XT-IDE? Right now I have a custom BIOS on a NIC which in theory will allow the use of CF on a 16-bit ISA card. I'm excited about the excitement, as it were, but I'm not totally grasping it.
Click to expand...

As I read what's been said, you've purchase two 16-bit ISA cards and are attempting to run a CompactFlash connected to one of them, all in an IBM 5150?

Some 16-bit IDE controllers will work in 8-bit ISA slots (i.e. a 5150) but as described in that thread, it needs a special BIOS that tells the CompactFlash media itself to flip to 8-bit transfer mode. Normal PATA hard drives won't work with this trick, and it needs the BIOS specifically tweaked to make this work. It's not the same as the original XT-IDE and the currently available XT-IDEv2, since both of those are designed to allow a 16-bit PATA drive to be used in an 8-bit ISA slot, which works by storing the half the bits that the bus can't transfer in a buffer, then sticking things back together in the host processor.

The current XT-IDE BIOS supports a wide variety of adapters and transfer modes, one of which is the 8-bit configuration for CompactFlash. The BIOS file I released via the other thread was a manually hacked-up proof-of-concept, unfortunately I didn't update the displayed info mainly as I don't know how to (!). So I would check with the seller exactly what BIOS is on the card.

As said there is another issue which is that not all CF adapters seem to work. IIRC cabled adapters (i.e. where the CompactFlash adapter is connected to the IDE controller card by a normal 40- or 80-pin IDE cable) tend to work more of the time.

Hope that helps.
 
Chris Hafner said:
- a four-digit RAM error code beginning with 1 shouldn't be possible for my 16-64KB machine
Click to expand...

Exactly how much RAM do you have. 64k shouldn't be a problem but any less won't work. I've never tested with anything less than 256k.
 
Chris Hafner said:
2. Screen says "1040 201", which is not an error I was seeing beforehand
Click to expand...
Chris Hafner said:
I have 640K of RAM - I just have the 16-64KB early motherboard.
Click to expand...

Reference: http://www.minuszerodegrees.net/5150/ram/5150_ram.htm

Irrespective of the type of 5150 motherboard ('16KB-64KB' or '64KB-256KB'), the "10" in "1024" means error at the 64KB mark. The "40" portion indicates failure of bit 6.

So either a RAM chip failure has occurred coincident with the 16-bit card addition, or the 16-bit card addition is the error cause. How about you undo the '16-bit card addition' to discover which it is.
 
pearce_jj said:
As I read what's been said, you've purchase two 16-bit ISA cards and are attempting to run a CompactFlash connected to one of them, all in an IBM 5150?
Click to expand...

Correct. To expand on it a bit - one 16-bit ISA NIC, and one 16-bit ISA IDE card. The 5150 is a 16-64KB with the third revision BIOS, two 360KB floppies, and 640K of RAM, and it worked fine prior to installation of the two 16-bit ISA cards

The CF card is a SanDisk Ultra 2GB in one partition, formatted as bootable (format /s) in MS-DOS 5.0.

The IDE/CF adapter is this Syba adapter, plugged directly into the IDE card: http://www.amazon.com/gp/product/B000YUMLPI/

The combo, in that configuration (plugged directly into controller, with Win98 hard drive unplugged), booted my Win98 machine successfully into MS-DOS 5.0.

pearce_jj said:
Some 16-bit IDE controllers will work in 8-bit ISA slots (i.e. a 5150) but as described in that thread, it needs a special BIOS that tells the CompactFlash media itself to flip to 8-bit transfer mode. Normal PATA hard drives won't work with this trick, and it needs the BIOS specifically tweaked to make this work. It's not the same as the original XT-IDE and the currently available XT-IDEv2, since both of those are designed to allow a 16-bit PATA drive to be used in an 8-bit ISA slot, which works by storing the half the bits that the bus can't transfer in a buffer, then sticking things back together in the host processor.

The current XT-IDE BIOS supports a wide variety of adapters and transfer modes, one of which is the 8-bit configuration for CompactFlash. The BIOS file I released via the other thread was a manually hacked-up proof-of-concept, unfortunately I didn't update the displayed info mainly as I don't know how to (!). So I would check with the seller exactly what BIOS is on the card.
Click to expand...

Ah - got it, I think. So XT-IDE is more full-featured and will allow 16-bit drives to communicate through an 8-bit ISA slot. The new development flips CF to 8-bit mode so that no fancy footwork is necessary to translate 16-bit data to 8-bit. Right?

pearce_jj said:
As said there is another issue which is that not all CF adapters seem to work. IIRC cabled adapters (i.e. where the CompactFlash adapter is connected to the IDE controller card by a normal 40- or 80-pin IDE cable) tend to work more of the time.
Click to expand...

Ah, bummer - mine is plugged in directly. I'm not sure if that would impact the floppy identification problem, but it could very well be impacting if (and how) the CF is being identified. I guess I'll need a male CF adapter now to plug into an IDE cable.

pearce_jj said:
Hope that helps.
Click to expand...

It does. Thanks very much for taking the time to explain it to me!

Cheers,
Chris Hafner
 
Might have one of the HD floppy Bios's appended to the XT-IDE Bios ? which may account for the floppy identification problem and possibly the wrong version of the XT-IDE bios, I use CF adapters that plug straight in and not had any problems, do you have the gear to read / burn eproms ?

Malc

Chris Hafner said:
Ah, bummer - mine is plugged in directly. I'm not sure if that would impact the floppy identification problem, but it could very well be impacting if (and how) the CF is being identified. I guess I'll need a male CF adapter now to plug into an IDE cable.

Cheers,
Chris Hafner
Click to expand...
 
Okay - I've returned home and have had a chance to shuffle some cards around. Here's what I've found.
1: The memory error appears to be based purely on card placement. Removing the two 16-bit cards didn't solve the memory problem (to my consternation), but moving the floppy/IO/RTC card to its original location did. I'd had to shift it before to make room for the IDE card with a compact flash sticking off to the side. With that card back to its original spot, everything worked fine - no memory error, keyboard input worked fine, my a: booted normally. So, while I'm not entirely sure why slot placement should matter to that degree, at least it solved the problem.

2. The memory error doesn't completely solve the floppy problem (though it does help). With the NIC added in (and then tested separately with the IDE card added in), the XTIDE BIOS still identifies A: as a 1.44MB floppy and does not boot off it. However, instead of locking up with the drive spinning perpetually, it tries to boot but then boots up IBM BASIC, as you'd expect with an unbootable disk.

3. The CF recognition problem isn't solved either. In the new card configuration, I couldn't stick the adapter directly into the IDE card anymore, so I tried another of the suggestions and connected a cable. I found a male CF/IDE adapter and connected it (blue connector to the IDE card, black connector to the CF/IDE adapter) and tried to boot. No luck. In a few reboots XTIDE didn't recognize the CF card at all, even as diamonds and hearts.

While the good news is that after removing the two 16-bit cards, my machine is back to being a nice, sweet, A:-booting machine again, I'm left scratching my head. Maybe Malc's suggestion is correct, that there's an HD floppy BIOS appended somehow that's short-circuiting things. Unfortunately I don't have the gear to read/burn eproms.

Thanks for all the help so far!

Cheers,
Chris Hafner
 
Chris Hafner said:
1: The memory error appears to be based purely on card placement. Removing the two 16-bit cards didn't solve the memory problem (to my consternation), but moving the floppy/IO/RTC card to its original location did. I'd had to shift it before to make room for the IDE card with a compact flash sticking off to the side. With that card back to its original spot, everything worked fine - no memory error, keyboard input worked fine, my a: booted normally. So, while I'm not entirely sure why slot placement should matter to that degree, at least it solved the problem.
Click to expand...
Does your "floppy/IO/RTC card" also have RAM (i.e. RAM addressed past the 64K mark)? If so, then a possibility is a RAM chip on it that is not making proper contact with its socket (alternating between good/bad contact on physical provoction).
 
Ok, I've been away due to family issues.

The good.

The bios on the card was tested with a Tandy 1000sx and a DTK 8088 clone, and booted my cards fine.

The bad,

I didn't test the floppies, and I was playing with a version of the bios with the XT floppy bios appended to it.
So I'm going to send him a new bios chip with just the XT-IDE bios on it.
I've still got another of the IDE controller I sent him, so the bios will be tested here with floppy access before it leaves.

Try another IDE cable. I think I put a 80 pin cable in the box, try finding a 40pin cable instead.

Can you post a pic of the screen where it's showing the 1.44 floppy message?

Thanks.
Later,
dabone
 
You must log in or register to reply here.
Back
Top