Did you tested the Cf ide adapter on more "modern" machine with onboard IDE controller? Like socket 7/slot1.
The quality of adapter or the I/o Card is a possiblity.

sineofsine wrote:

Any other ideas?

Your overclocking. Bring every speed speed down to default/normal and try again, corruption due to increased bus speed is not uncommon.

konc wrote:

sineofsine wrote:

Any other ideas?

Your overclocking. Bring every speed speed down to default/normal and try again, corruption due to increased bus speed is not uncommon.

I've tried every bus clock from clk2/1 on down to clk2/12, as well as fixed a 7.19MHz. None of them seem to help with this problem.

Try new cable for IDE and another different multi-i/o ISA card. Acer is not my favorite.

Cheers, pentiumspeed

pentiumspeed wrote:

Try new cable for IDE and another different multi-i/o ISA card. Acer is not my favorite.

Cheers, pentiumspeed

Thanks, will do! I'm waiting on a Goldstar and UMC IO card, will update when I try them out.

I would prefer using winbond chipset for the multi-i/o, and properly buffered IDE port.

Cheers,

See what the CF card settings detect as on a much newer system.

I have some CF cards that are really flakey with a 486 setup but seem to work fine with a newer system. They are Monster Digital brand. They also do not detect the settings properly on the 486 system. To get them to sort of work reliably on the 486 system I have to detect the settings with a newer system and then manually enter those settings into the BIOS on the 486 system.

cyclone3d wrote:

To get them to sort of work reliably on the 486 system I have to detect the settings with a newer system and then manually enter those settings into the BIOS on the 486 system.

Great tip, thanks! So far, I've checked on several machines as well as the `whatide` tool and they all agree on the drive geometry. I've now tried on several motherboards.

I did find something that does work with all three IO cards in my original motherboard: a real IDE hard drive. It's rock solid. Go figure!

Although I wouldn't necessarily be surprised, and I don't know if I believe that CF cards really have the correct signal timings and behavior for IDE in fixed disk mode, it's hard for me to believe that every single one of these CF cards are that unreliable. People seem to just be throwing random CF cards into their machines and it always seems to work... except in mine!

The only thing left for me to consider is the CF-IDE adapter, which I have seen some few complaints about online. Maybe I'm unlucky. I examined all of the solder joints on this thing and none of them look obviously broken. Maybe the fine-pitch SMD connections on the CF socket aren't solid. Rather than go over it all with an iron, I'm going to return it and try another.

What a hassle.

Have you tried different power cables to the adapter? When you use a real HDD it would be using a different plug...
Maybe also replace the jumpers as well, there might be a poor connection in there somewhere.

Another bunch of ideas, if the CF cards are really the culprit.. 😀

* See if FreeDOS behaves the same (more modern; it can handle FAT32 and LBA)
* Format your CF cards in one or two old camcorders or digital cameras.
They often can fix/overwrite some broken filesystems. And remove/replace existing partitions, too.
* Try an industrial CF card with up to 512MB of capacity (avoids trouble with ~500MB barrier).
* Clear your CF cards starting sectors (esp. track 0), too.
There are a related few tools for *nix (dd etc) and DOS (s0kill etc).

I didn't ready all the replies properly just skimmed them, so it might already been said but most likely what is happening is you are using a card over 501mb size. The 486 dx2 most likely is limited to 501mb. Even tho the hard disk controller can see 1 or 2 gb etc and shows that in the bios, the bios can't see any sectors past the 501mb mark so it reads and writes files under 501mb fine and anything over that has seek errors

I've got the same problem with a NEATSX 386 - I can read, format and write to a CF card, but running anything off it I get a NOT READY error. Card works fine in a 486 and MMX system.

If that's the case, it might be a BIOS issue.
286/386 systems usually predate HDDs bigger than a few hundred MBs.
Also, they are not aware of the subtle changes introduced with ATA-2..

Using XTIDE Universal BIOS or one of these old EIDE BIOSes might fix these compatibility issues.
Re: A special BIOS LBA EIDE Multi-IO Controller: DTC EIDE Ultima Pro

Jo22 wrote on 2021-01-09, 12:59:

If that's the case, it might be a BIOS issue. 286/386 systems usually predate HDDs bigger than a few hundred MBs. Also, they are […]
Show full quote

If that's the case, it might be a BIOS issue.
286/386 systems usually predate HDDs bigger than a few hundred MBs.
Also, they are not aware of the subtle changes introduced with ATA-2..

Using XTIDE Universal BIOS or one of these old EIDE BIOSes might fix these compatibility issues.
Re: A special BIOS LBA EIDE Multi-IO Controller: DTC EIDE Ultima Pro

A CHS 1060/16/63 540 MB hard drive was able to be booted off on the same machine that has problems reading a 933/16/63 CF card - more details here:

(SOLVED) CF card woes (this time in a 386 and ALSO a Pentium MMX with SCSI).

DX2 100? Are you running 50MHz? This could be the problem. Try 40MHz which will be DX2-80. Might be more reliable.

Cheers,

^ I second this. In 286-486 systems, the HDD often is directly attached to the ISA bus, which is the front side bus, essentially.
If the motherboard gets overclocked, so does ISA and IDE.

Also, some BIOSes may not support more than 1023/15/63 or 1023/16/63, even though the CMOS Setup Utility accepts any value.
Re: Dos floppy installation problem