VOGONS


First post, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie

Hello all! I should start off and say im new to the site. Though ive been lurking as a guest for a few months following old PC topics while I build a good old Pentium 3 Slot 1 system.

Anyway to my issue, I finally managed to get my hands on an SD card to IDE Adapter to get the system a somewhat proper modern HDD. Followed with a 64 GB Kingston SD Card.
I formatted the SD card with msdos and a fat32 partition of 40 GBs (Leaving 22 GBs so that the SD card never thinks its full and causing slow downs), Making sure the sector size was 512 mbs. Copied the Heads, Sector, Tracks data from Gparted to enter into the motherboards BIOS (Auto-detect would cause the system to freeze up).

However on every restart, The motherboard reports a "Primary Master Hard Disk Failure" and Windows 98 Setup reporting that there are no hard disks installed in the system.

The system specs are as followed:
FIC VB-601-V REV 1.4 (VIA 440BX Chipset, Award Modular BIOS v4.51PG SV601V14 dated 19.07.1999)
512 MBs of RAM (2 * 256 mb PC133U Micron sticks)
Pentium 3 50mhz 100mhz FSB
ASUS V7100 (Based on the GeForce2 MX 400 chipset)
NEC USB card (Seems generic, No model numbers)
Soundblaster Live! Value (CB 4830)
Netgear Gigabit Networking card (GA311 REV A1)
Disk Drive, Floppy Drive and PSU are Generic (PSU Will be swapped to something more modern)

Troubleshooting attempts:
- Using a Quantum Bigfoot, System installed and booted fine
- Using a 80 GB Seagate SATA HDD with an adapter, Freezed on BIOS
- Using a 80 GB Seagate SATA HDD and limiting it to 64 GBs using Seagates Diskwizard program (Was a while back, Forgot the name), System installed and booted fine
- Using a 2 GB Generic SD card in the adaptor. System installed and booted fine.
- Using the 64 GB Kingston SD card in another PC. Windows xp reads and writes from it like a normal HDD with no issues.
- Attempting to use FreeDOS and 3 different DOS programs to detect the CHS configuration of the drive. Then using it in the BIOS. Freezed on BIOS, As well as reported the drive as only 8 GBs

I do know some 440BX chipsets limit the max HDD size to 32 gigabytes, However. I dont believe this is the case as a 64 GB hdd was detected fine by the system. The bios also supports LBA.

The only idea I have left is to get a 32 GB sd card. Sadly I cant return the 64 GB sd card as its already been opened from packaging 🙁. Anyone have any clue on what I could be doing wrong? Happy to answer any questions if ive missed something.

EDIT:
Completely forgot to list the SD card adaptor! Its a "FC-1307 SD to CF Adapter V1.4" according to BIOS and Windows XP device manager (When testing the 64 GB card on another machine). With a model number SD35VC0

Reply 3 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
douglar wrote on 2020-11-24, 04:57:

Try connecting using a 40 conductor ide cable to encourage the negotiation of a slower transfer rate. Those adapters are limited to 25mbs, so the performance drop should not be large.

Just tried a 40 conductor. Still no luck. I know the transfer speeds are low. I mainly want to use the SD card for easy file transfer as well as making small backups of it in-case I break the install in a future.

Warlord wrote on 2020-11-24, 05:10:

I have no experience with CF to SD. I only know about IDE to CF. I can see how these problems can happen cuz a SD card isn't conform to same standards as a CF card or even a SSD or a HDD.

I knew I was going to be entering a real "hit or miss" going with an SD card, As I know there are cases where the motherboards really dont like the adaptors. CF was going to be my original choice. However the prices here (New Zealand) are quite a bit even for very small sizes (Though if I do start to play in the realm of 486 or lower DOS machines. I would pay for a small CF card knowing it would have a way better chance with compatibility)

Reply 4 of 20, by jmarsh

User metadata
Rank Oldbie
Rank
Oldbie

The FC-1307 adapters are notorious for acting like this. They rely on detecting a specific type of FAT32 or NTFS partition as the first on the disk to initialize their emulated C/H/S values, or else they won't work correctly until sector 0 is rewritten.

Reply 5 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
jmarsh wrote on 2020-11-24, 07:11:

The FC-1307 adapters are notorious for acting like this. They rely on detecting a specific type of FAT32 or NTFS partition as the first on the disk to initialize their emulated C/H/S values, or else they won't work correctly until sector 0 is rewritten.

Ahh, Looks like i didn't do enough research on the adapter then. What would be my best option in this case? Ive done a few formats with gparted and reinitialized it with ms-dos with no luck (Keep in mind this was outside of the adapter using a laptop with a native SD card slot)

Though, The only thing that doesn't make sense is how could my Windows XP machine see the adapter with card with no issues? Something about the updated bios in it can see through the adapter straight to the card with magic?

Reply 7 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
jmarsh wrote on 2020-11-24, 08:58:

Partition and format it on the machine you intend using it with, via the adapter. Make sure the first partition is FAT32.

I used Hirens Boot CD v15 on the machine I intended to use it with. I tried Gparted (In low ram mode) first. It found it. Reinitialized ms-dos partition table and did a full format to fat32. Restarted and no luck, Still says "Primary Master Hard Disk Fail" on boot up.
I then tried all the tools under dos programs. All the different versions of FDisk reported that no physical disks were found. Weird however, Some of the hard disk testing tools saw it just fine and at full size.

Though in the end still no luck.

Reply 8 of 20, by darry

User metadata
Rank l33t++
Rank
l33t++

I agree 100% with jmarsh about the FC1307 having issues with FAT16 .

For the current issue, however, If it's the system BIOS that's complaining about "Primary Master Hard Disk Fail", my guess is that it does not like the parameters entered for SD card .

Specifically, my guess is that the 80GB drive limited to 64GB worked because it is ever so slightly smaller than the 64GB SD card .

Was the 80GB drive that was limited to 64GB detected by BIOS or did you need to enter parameters manually ?

Reply 9 of 20, by digistorm

User metadata
Rank Member
Rank
Member

What you say might be plausible. I have an IBM disk with a 2 GB jumper that limits it to 1024 cylinders. But that is still 1 cylinder too much for the BIOS of my 486. Very sneaky bug 😉

Reply 10 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
darry wrote on 2020-11-24, 09:53:
I agree 100% with jmarsh about the FC1307 having issues with FAT16 . […]
Show full quote

I agree 100% with jmarsh about the FC1307 having issues with FAT16 .

For the current issue, however, If it's the system BIOS that's complaining about "Primary Master Hard Disk Fail", my guess is that it does not like the parameters entered for SD card .

Specifically, my guess is that the 80GB drive limited to 64GB worked because it is ever so slightly smaller than the 64GB SD card .

Was the 80GB drive that was limited to 64GB detected by BIOS or did you need to enter parameters manually ?

The 64 gb drive was automatically picked up by the system. So it entered all the parameters when it auto detected. But now that you say that I don't believe the capacities were the same when they showed in BIOS (I will have to look back at it). So that could be the issue.

Worse comes to worse Ill just have to get a 32 GB SD card. I can probably have the 64 GB card on hand for some future project.

Reply 11 of 20, by skideric

User metadata
Rank Newbie
Rank
Newbie

I am pretty sure my adapter is that same one.Had hit & miss to start Finally formatted my NEW 16GB SD Card using Win 7 to Fat 32.Then booted Hiren's Boot CD,used Acronis to make 1St part Fat 16 & Active,then rest of card Fat 32.
Worked after that to install Win98se on my XPS 400...

Unisys PW2 3256 386 DX25 & Co-Proc (x4) 16mb,SCSI) Dimension XPS R400 PII 400mz 384mb Voodoo 3 2000,16gb MSD) Towers with Intel D850EMV2 P4 2.4gz 256mb PC800 250mb Zip) ECS UC-4913 486 DX-2 66

Reply 12 of 20, by johnnycontrario

User metadata
Rank Newbie
Rank
Newbie

I have two of these SD to IDE adapters. I have not had any problems with FAT16 partitions, however, wiping the first couple of blocks with the linux dd command before formatting the disk in the target computer has become a habit of mine. I believe there are some Windows utilities that can accomplish the same thing. I can't remember the exact technical explanation, but I began doing this because I read that the old partitioning utilities didn't actually erase the other data that might be in the first block, which can cause issues with booting some flash cards (the chip in the SD to IDE is used in SD to CF adapters). This has helped me resolve the occasional issue.

While I haven't tried using an SD card larger than 16GB, I have had some other SD card size related issues on a 486 with VIA chipset and Award 4.51 BIOS. This may not be related to your problem at all, but I thought I would share in case it helps.

8GB SD cards were recognized by the BIOS OK, but no matter what version of DOS I tried (freeDOS, DOS 6.22, Win98SE), FDISK would not see the full size of the disk (I believe it was always 512MB? I don't remember). Larger and smaller SD card sizes tested OK. I could partition and format up to 8GB of a 16GB SD card in DOS 6.22 and I was able to read/write the extended partitions without any issues. I was able to use the full 16GB with Win98SE. I didn't have the tools at the time to boot Win2k or Linux to do any further troubleshooting, so I'm not sure the exact cause. However, I'm fairly certain it was a bug (not a limitation) in the BIOS for the following reasons.
1. After installing the XTIDE boot ROM, 8GB SD cards work correctly.
2. The problem could not be reproduced on a more modern VIA C3 computer using the same SD cards and adapter.

In my case, I think the BIOS had a bug where the 8GB SD cards were hitting some magic CHS combination that caused invalid things to happen. I wonder if it's a similar issue with the 64GB card.

Reply 13 of 20, by debs3759

User metadata
Rank Oldbie
Rank
Oldbie
johnnycontrario wrote on 2020-11-24, 23:18:

I have two of these SD to IDE adapters. I have not had any problems with FAT16 partitions, however, wiping the first couple of blocks with the linux dd command before formatting the disk in the target computer has become a habit of mine. I believe there are some Windows utilities that can accomplish the same thing.

The diskpart app in Windows can do that. Whenever I test a drive, I use that first, then create a single partition before wiping free sectors with PrivaZer (free or make a donation - I make a small donation from time to time, as it is part of my system clean-up routine, and the only app I have that cleans the MFT and other file system features) then testing the drive with hddscan, a free utility. I do that even with new drives, just to be sure there are no faults.

See my graphics card database at www.gpuzoo.com
Constantly being worked on. Feel free to message me with any corrections or details of cards you would like me to research and add.

Reply 14 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
skideric wrote on 2020-11-24, 21:05:

I am pretty sure my adapter is that same one.Had hit & miss to start Finally formatted my NEW 16GB SD Card using Win 7 to Fat 32.Then booted Hiren's Boot CD,used Acronis to make 1St part Fat 16 & Active,then rest of card Fat 32.
Worked after that to install Win98se on my XPS 400...

Ill have a go at formatting the first part of it FAT16. Ive been just trying to use FAT32. So will have an attempt doing the way you did it.

johnnycontrario wrote on 2020-11-24, 23:18:

...In my case, I think the BIOS had a bug where the 8GB SD cards were hitting some magic CHS combination that caused invalid things to happen. I wonder if it's a similar issue with the 64GB card.

It could be, Gparted reports the SD card as having 255 heads. Which the systems BIOS does let me input and accepts the changes. But in reality what drive back in 1999 had 255 heads haha.

UPDATE

Tried skideric's way of making the 1st partition FAT 16 as well as Active. Still nothing sadly, Same error message.

UPDATE 2

Nevermind! Some progress. Skideric's way seems to let the system actually see the card and autodetect it without an error message! May have some luck here!

Reply 15 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie

Update 3

Something interesting happened. So following skideric's advice. I managed to get the system to auto detect the card with a 4 gigabyte FAT16 partition set to "boot" at the start of the card. The system then booted with no errors. I started Gparted however and the it was telling me the card was only 34 MBs in size. Which was weird.

I put the card into my laptop. Using Gparted on it to add a small 10 GB FAT32 partition. Then plugged it back into the machine... However the machine then hanged on boot up.

Using Gparted again on machine with the adapter. I made a new MBR, Made another 4 GB FAT16 partition with a boot flag. And the motherboard wont auto detect it and instead freeze up. So I had it working. But after adding the FAT32 partition it all stopped working? Im confused on whats happened here.

Reply 16 of 20, by Spitz

User metadata
Rank Member
Rank
Member

In my case, I've made the partitioning using FDISK, deleted all garbage (SD card 32GB after Amiga use) created new partition >2000MB (I don't remember if agreed "use large partition ... Bla bla" in fdisk so check it out too) then created 2nd with the rest of remaining space. 2nd was corrupted under Windows 98 so I've installed partition magic, wipe it out remade and it was ok. For some reasons 1st partition had to be under 2GB... Don't know why. It's working like a charm now.

Well... I miss 80/90s ... End of story

Reply 17 of 20, by johnnycontrario

User metadata
Rank Newbie
Rank
Newbie

Before you posted your last update, I was about to ask about this:

BobyTheDragon wrote on 2020-11-24, 04:19:

- Attempting to use FreeDOS and 3 different DOS programs to detect the CHS configuration of the drive. Then using it in the BIOS. Freezed on BIOS, As well as reported the drive as only 8 GBs

This may be a significant clue. On the geometry reported by the DOS utilities: did it correspond to an 8GB drive, or did it correspond to a 64GB drive but was miscalculated by the BIOS as 8GB? Have you checked for BIOS updates or mods?

BobyTheDragon wrote on 2020-11-25, 04:37:
Update 3 […]
Show full quote

Update 3

Something interesting happened. So following skideric's advice. I managed to get the system to auto detect the card with a 4 gigabyte FAT16 partition set to "boot" at the start of the card. The system then booted with no errors. I started Gparted however and the it was telling me the card was only 34 MBs in size. Which was weird.

I put the card into my laptop. Using Gparted on it to add a small 10 GB FAT32 partition. Then plugged it back into the machine... However the machine then hanged on boot up.

Using Gparted again on machine with the adapter. I made a new MBR, Made another 4 GB FAT16 partition with a boot flag. And the motherboard wont auto detect it and instead freeze up. So I had it working. But after adding the FAT32 partition it all stopped working? Im confused on whats happened here.

Did you create a 4GB FAT16 partition or a 2GB FAT16 partition? I don't know what DOS would do with a 4GB FAT16 partition. I stopped using GParted for partitioning for DOS because it was too easy to create partitioning schemes that don't work. My current workflow is to zero out the first 2 sectors with dd and partition the disk in-system with the partitioning utility of the OS I plan to use. If I need something more advanced, but need the disk to be DOS friendly, I use FreeDOS fdisk to prepare the first few FAT16 and 32 partitions. When I don't have access to the target computer, I have had decent success with partitioning within QEMU (I mount the block device directly like a maniac), or PCem (dd your SD card to disk, mount the IMG in PCem, then dd the partitioned IMG back to your disk). The benefit of doing this in PCem is that you can see what a real BIOS thinks of your disk geometry. It's an idiot's approach, but DOS is fussy and it's an easy way to guarantee that your partitions are created correctly.

When I had the problem with the 8GB drives, I didn't immediately discover it because I had prepared my 8GB SD card with DOS and Win 3.1 via a VM. I actually used that SD card in my 486 for a long time as my test card and didn't notice anything wrong until I tried partitioning another 8GB SD card on that system. I believe I didn't notice because I hadn't used any part of the disk above the 512MB that DOS was unable to access. DOS would happily list the volumes because that data is in the area before the first partition. Sure enough, if I tried partitioning ANY 8GB SD card, including that one that was 'working', DOS could only see a 512MB disk. Nothing worked despite the BIOS auto-detecting the correct geometry every time. Your false starts may be due to a similar issue.

I recognize the mystery 34MB disk size, even though it sounds like you discovered it a different way. If you start the system with no SD card in the adapter, the system will report a 34MB disk. It might be worth it to dd the contents and see what's in there. I didn't have a Linux boot disk handy when I noticed it. Could we be so lucky that it's the firmware? 🤣

This SD to IDE adapter has some other funny behaviors that may or may not be related, so I'll list them here.
*If you put 2 of these adapters in the same system, only one is recognized by the BIOS, but the adapters talk to each other somehow to combine both SD cards into that disk. It doesn't matter if they're on different IDE channels. The partition tables on the SD cards were untouched, so it appears that the adapter was presenting its own partition table to the system. This might have something to do with that 34MB disk. Food for thought: the SD to IDE adapter uses the same chip used in those SD to CF card adapters, and I've noticed some available with 2 micro SD slots. I didn't test it because I didn't want to wreck my data, but I bet it's meant to concatenate the FAT32 volumes of both SD cards.
* These adapters can share the same IDE channel with other IDE devices, but the adapter MUST be master.

Reply 18 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie
johnnycontrario wrote on 2020-11-25, 06:19:

This may be a significant clue. On the geometry reported by the DOS utilities: did it correspond to an 8GB drive, or did it correspond to a 64GB drive but was miscalculated by the BIOS as 8GB? Have you checked for BIOS updates or mods?

It seemed to correspond to a 8GB Drive when the BIOS calculated the size after putting in the CHS (When it was given to me by the DOS programs). But some programs reported the the size correctly as 64 gbs (Or close to it + or - 2 gbs) and were giving me the same CHS that correspond to it being 8GBS.

Have attempted to look for BIOS and Mods. Cant find much but dodgy driver websites. Though it seems as I already have the latest BIOS.

johnnycontrario wrote on 2020-11-25, 06:19:

Did you create a 4GB FAT16 partition or a 2GB FAT16 partition?...

...I stopped using GParted for partitioning for DOS because it was too easy to create partitioning schemes that don't work. My current workflow is to zero out the first 2 sectors with dd and partition the disk in-system with the partitioning utility of the OS I plan to use.

I tried a 4GB FAT16 as well as a 256MB FAT16 partition. Both didnt work. Also, I am unable to use fdisk or windows 98 installer to do any formatting of the disk as they both report there being no drive currently in the system.

johnnycontrario wrote on 2020-11-25, 06:19:

I recognize the mystery 34MB disk size, even though it sounds like you discovered it a different way. If you start the system with no SD card in the adapter, the system will report a 34MB disk. It might be worth it to dd the contents and see what's in there. I didn't have a Linux boot disk handy when I noticed it. Could we be so lucky that it's the firmware? 🤣

Funny enough, I did remove the card and it does infact see a 34 mb partition. So maybe it is the firmware haha.

Reply 19 of 20, by BobyTheDragon

User metadata
Rank Newbie
Rank
Newbie

What I might do is if in the future if move to a new motherboard platform (Probably Pentium 3 based socket 370 for that lovely 1GHz fun) I can use the adapter and 64 GB card there (If the BIOS can be modded or natively supports very large HDDs), I'm sure its not the adapter at fault here as it works fine in every other system using IDE and the 64GB card. I think in the end its just the BIOS not liking the weird CHS values. As the motherboard itself doesn't seem well documented.

In all honestly. 64GBs for a Windows 98 system is a bit overkill, I only got it because it was on special and was a few dollars more than a 32GB model. So what I might do is just go with a 16 or 32GB card (Wanting to leave some unpartitioned space so that the card doesn't ever think its full) since i definitely know for a fact that the system should be okay with 32GB disks. As playing around with the CHS to try and go higher than about 65 GBs causes the system to freeze instantly (Not even allow me to save it, Instantly crashes as soon as I put the values in)

I can try any more ideas you guys wanna throw at me. But I think in the end is just the motherboard being weird with the card. I think what darry said above:
"Specifically, my guess is that the 80GB drive limited to 64GB worked because it is ever so slightly smaller than the 64GB SD card ."
is correct. It doesn't like 64GB storage. And since I did some probably terrible calculations (Since Seagates tools doesnt play with GBs or MBs but LBA sizes) My "64 GB" HDD was truly lower than a 64GB hard limit of the BIOS.