weedeewee wrote on 2024-02-17, 22:07:

got any IDE block mode setting in the bios ?

Yes, I tried it enabled and disabled.

Shponglefan wrote on 2024-02-17, 21:59:

steevf wrote on 2024-02-17, 21:46:

The 286 and 386 have XTIDEUB.
I'm just using the motherboard BIOS as it has LBA mode and can detect the CF cards up to 8BG. No overlay software.

Try using XTIDE with the 486 and see if that makes any difference.

If you can recreate the conditions under which the cards work with the 286/386, it might narrow it down to a specific issue with the BIOS in your 486.

I'm going to dig out my XTIDE Cards/chips and try them but that won't be until tomorrow. I do wonder about the BIOS and the BIOS version. Like I said earlier, there is an update to the MB bios but I first need to source a compatible PROM to implement that.

An EPROM or EEPROM would be better than a PROM. A PROM would work, but if the write fails, it can't be erased.

Here is another thing to try: if you have, use a small(128 or 256 MB) CF card, just for a try. I have several DOS systems (also with LBA option un the bios) which just dont work with larger cards.
here is an example
Dec 466d2 CF Card trouble

my bet would be 85% cgance that a small (and better older) cf card works

steevf wrote on 2024-02-17, 21:53:

All my other systems are stable and I don't want to mess with them. This is the only 486 motherboard I have. I guess I could […]
Show full quote

pentiumspeed wrote on 2024-02-17, 21:35:

Have another computer? Or different motherboard. BTW, please no EDO memory.

Cheers,

All my other systems are stable and I don't want to mess with them.
This is the only 486 motherboard I have. I guess I could try finding a PCI 486 motherboard and abandon the VLB stuff altogether.

Yeah that EDO stick is out. It would work better in my Pentium-200 system anyway if I didn't already have that computer maxed out at 128 MB.

I agree with Shponglefan.
I use XTIDE on a network card on an ASUS SV2GX4 board.
This board has no problem without XTIDE, but XTIDE removes the 8GB limit, so I can use a 16GB CF flash card.
Setting in Bios: harddisk => None (so no fooling around with LBA, alll done by XTIDE) and Block mode On. (disk access seems to be a tiny bit faster)

Maybe I've missed this but I've had the same thing a few times. For me the solution was:
- get the cf into a modern computer, clean the partition table (diskpart ie). Do not create any new partition yet.
- put the cf back into the old computer, there partition it. Reboot. Run fdisk /mbr.
- then high level format it, and get the software back on it.

It sounds to me there's a mismatch between the partition table and the layout the computer detects.

waterbeesje wrote on 2024-02-18, 16:34:

Maybe I've missed this but I've had the same thing a few times. For me the solution was: - get the cf into a modern computer, cl […]
Show full quote

Maybe I've missed this but I've had the same thing a few times. For me the solution was:
- get the cf into a modern computer, clean the partition table (diskpart ie). Do not create any new partition yet.
- put the cf back into the old computer, there partition it. Reboot. Run fdisk /mbr.
- then high level format it, and get the software back on it.

It sounds to me there's a mismatch between the partition table and the layout the computer detects.

Interesting. So removing all partitions with MS DOS 6.2 fdisk is not enough?
But I also do not know exactly what the OP did.

Ok, time for an update.

I think the general consensus of some folks here were pushing toward using XTIDEUB and in the back of my mind I was starting to think the same way which would imply that the Motherboard BIOS is not up to the task. And I'm going to agree with that.

To validate that direction I configured an ide_386l.bin bios for this machine using the VL 400GG GoldStar Prime 2 IDE Controller Card as it seemed to be the best one for this system and the XTIDECFG utility actually detected it.
Then burned it to a chip and stuck it in using one of my Lo-tech ISA ROM board. (see image for reference)

And it just worked! It was stable! Part of me is not surprised but I'm also relieved that something was working . Going forward I'm going to trust my gut about making BIOS updates first(which I failed to do this time).
So, the XUB is just as fast as long as you enable the shadow BIOS for it. In my case address C800. It's especially important because fast EEPROM chips are impossible to find and all I could find was slower 150ns ones. (the MB bios is on a 45ns PROM)

At this point I would normally say I'm done but I was really hoping the XUB option wasn't going to be necessary so I'm also going to try the BIOS update.

I was able to source some 45ns 64K proms. (I can't find any 64K EEPROMS anywhere, the listing states one thing but when looking at the data sheet they contradict themselves in the same sentence by saying something like "One Time Programmable -EPROMS". My sources for looking have been (Digikey, Mouser, Newark and Jameco) It's a mess and a topic for another time.)

Anyway, I just want to give the BIOS upgrade a shot before making a final decision on this as there is precedent for using 2GB partitions on a 4GB CF using the V2.3 BIOS for this motherboard. If it doesn't make a difference I've still got the XUB option and I can put that on an Ethernet card option ROM.

When I try the BIOS upgrade I will update this thread on my result (estimated time to complete that test is about 1.5 weeks.)

Ok, I finally got a PROM programmed with the 2.3V of the BIOS and did my full 2GB CF card test

TL;DR the BIOS upgrade did not fix anything. I will continue to use XUB.

I should say, my testing methodology during all of this has been the following activity:

1. start off with a cleaned 2GB CF card
2.boot from DOS 6.21 install disk
3. manually re-partition to four evenly sized partitions,
4. fdisk /MRB the thing (usually only need to do this once but I redo it every time in case.)
5. format all partitions
6. install DOS 6.21
7. boot from C:
8. install DOS 6.21 supplements
9. xcopy the entire contents of C: drive to the D: drive
10. xcopy again all of C: to E: drive
11. xcopy again all of C: to F: drive.

If I can make it through all of that without any write errors I consider it stable.
When it fails I usually see failures in step 6. or step 8. If it fails in either of those two steps it will 100% fail in steps 9,10, and 11.
So far the XUB is the only thing that made it through all the way with several different partition sizes and CF card sizes larger than 2GB.
I will be burning the XUB image to a prom and sticking it on a network card.

I would like to thank everyone who chipped in with comments and suggestions.
It was all a huge help to get my thinking towards a solution and I learned a lot.
Even if I didn't reply to everyone I certainly noted your advice.

This sure is a great forum!