What kind of disk controller (hard disk) is present?

MFM/RLL/ESDI controllers can have a ROM based lowlevel routine that lives in the adapter rom region. Sometimes these cards configure the location of this bios in the C000 segment.

It is also possible for XTIDE bios to be mapped in this segment.

VGA video bioses *also* live in the C000 segment, and some are larger than others.

If the disk type is MFM/RLL with a bios, or XTIDE is present, and it is in the C000 segment, this could be your problem.

Int18 gets called when no boot devices are found.

wierd_w wrote on 2026-08-04, 14:24:

What kind of disk controller (hard disk) is present? […]
Show full quote

What kind of disk controller (hard disk) is present?

MFM/RLL/ESDI controllers can have a ROM based lowlevel routine that lives in the adapter rom region. Sometimes these cards configure the location of this bios in the C000 segment.

It is also possible for XTIDE bios to be mapped in this segment.

VGA video bioses *also* live in the C000 segment, and some are larger than others.

If the disk type is MFM/RLL with a bios, or XTIDE is present, and it is in the C000 segment, this could be your problem.

Int18 gets called when no boot devices are found.

Hi!
I use a Prime 2C controller, https://theretroweb.com/chips/6407
I'll try a bunch of others then, I have a lot of ISAs. Video card I have only 2 ISA with VGA, the one that works must stay.Cl-GD5420-750C-C

Ok, one, UMC, works.
Not loading 6.22 though, 640K only is the issue?

Error is "division overflow"

Edit: so far Phil's fast DOS 3.5 disk worked; 6.22 crashes as above.

All the controllers fail to iitialise the HDD, either the video card stops working or prevents posting at all

https://theretroweb.com/chips/6236 is the only one loading floppy drive

Leaving or removing the FPU doesn't change a thing.

Sounds like a signal integrity issue - as if the data read doesn't quite make it intact up to the point where the uops enter the cpu and cause strange errors and/or hangs.

640K is more than enough for 6.22.

Prime 2C has no bios in c000 that could bother anything, only those 8 bit XT MFM controllers did.

st31276a wrote on 2026-08-04, 17:34:

Sounds like a signal integrity issue - as if the data read doesn't quite make it intact up to the point where the uops enter the cpu and cause strange errors and/or hangs.

640K is more than enough for 6.22.

Prime 2C has no bios in c000 that could bother anything, only those 8 bit XT MFM controllers did.

I noticed 2 things:

1. my floppy with 6.22 disk 1 is dead: I should have used the gotek instead, shame on me
2. sometimes the board hangs: if I press down the cpu it starts working again, probably I need to put some extra layer to have the bracket push it more
3. i have to figure out which hdd tyoe is a 980 cyl * 10heads * 17 sectors... and try to see if it works like this 😀

As for the HDD I have to go the boring check all connections.

Are there bad joins on the back of the cpu socket?

wierd_w wrote on 2026-08-04, 18:04:

Are there bad joins on the back of the cpu socket?

At this point I have to wash it and deoxit, check it thoroughly.
Board was boxed an eternity ago (90s) and then sold to me days ago.

Edit: it now uses the HDD but the HDD isn't picked p by the POST screen (failed test) but it works if defined correctly in bios.
Strange 😀

Is there a universal bios setting program?
I could put Type 47 and then set it via such program.

When I put 47 I can't enter data manually.

Nexxen wrote on 2026-08-04, 18:10:

Edit: it now uses the HDD but the HDD isn't picked p by the POST screen (failed test) but it works if defined correctly in bios.
Strange 😀

It's an IDE drive? If it's smaller than about 81MB (980 * 10 * 17 * 512 / 1048576) it will fail when the POST tries to seek to the last cylinder but will work otherwise.

Nexxen wrote on 2026-08-04, 18:56:

Is there a universal bios setting program?
I could put Type 47 and then set it via such program.

When I put 47 I can't enter data manually.

While GSETUP and friends may be able to load type 47 data into the expected location in CMOS RAM for various BIOSes, it can't add type 47 support to a BIOS that doesn't have it.
Also, see https://forum.theretroweb.com/t/286-discussio … 56/2?u=makefile before assuming Multi-I/O card IDE will work with your existing 286 BIOS.
My recommendation is try out XUBDisk XUBDisk - floppy disk XTIDE Universal BIOS booter so you can "try before you buy" (an EEPROM) to ensure your IDE issues are purely software issues.

jakethompson1 wrote on 2026-08-04, 19:03:

It's an IDE drive? If it's smaller than about 81MB (980 * 10 * 17 * 512 / 1048576) it will fail when the POST tries to seek to t […]
Show full quote

Nexxen wrote on 2026-08-04, 18:10:

Edit: it now uses the HDD but the HDD isn't picked p by the POST screen (failed test) but it works if defined correctly in bios.
Strange 😀

It's an IDE drive? If it's smaller than about 81MB (980 * 10 * 17 * 512 / 1048576) it will fail when the POST tries to seek to the last cylinder but will work otherwise.

Nexxen wrote on 2026-08-04, 18:56:

Is there a universal bios setting program?
I could put Type 47 and then set it via such program.

When I put 47 I can't enter data manually.

While GSETUP and friends may be able to load type 47 data into the expected location in CMOS RAM for various BIOSes, it can't add type 47 support to a BIOS that doesn't have it.
Also, see https://forum.theretroweb.com/t/286-discussio … 56/2?u=makefile before assuming Multi-I/O card IDE will work with your existing 286 BIOS.
My recommendation is try out XUBDisk XUBDisk - floppy disk XTIDE Universal BIOS booter so you can "try before you buy" (an EEPROM) to ensure your IDE issues are purely software issues.

I'll look into it 😀
Thanks!

Does anyone know how to open the socket?

Nexxen wrote on 2026-08-04, 21:51:

Does anyone know how to open the socket?

The metal clip hooks around the corners of the socket. You need to pry it away from the socket at the corner, but be careful, usually it's really tight and takes a bit of force, so try to use a plastic tool (it has to be quite thin to fit between the socket and the clip). Once you release the first corner the second one is easier, and then you can lift it to release the CPU.

TheMobRules wrote on 2026-08-05, 04:13:

Nexxen wrote on 2026-08-04, 21:51:

Does anyone know how to open the socket?

The metal clip hooks around the corners of the socket. You need to pry it away from the socket at the corner, but be careful, usually it's really tight and takes a bit of force, so try to use a plastic tool (it has to be quite thin to fit between the socket and the clip). Once you release the first corner the second one is easier, and then you can lift it to release the CPU.

Ok, it does not look "that" difficult but definitely not easy either.
I have 20 of those cpus to test, is a heavy heatsink enough or it needs to be tight to work?

It wasn't that difficult but yes, you need to be very careful to avoid damaging anything around.

The bios images are the same as the ones on TRW, I'm gonna add them here just to have a backup.

I was wondering, could any other BIOS work? Or this board has specific needs?
Like any other clone could it just work? I'll test it with some other's chips. Worstcase it won't work 😀

I couldn't locate a soft reset button header.
Could it be that there's none?