This is very interesting. The BIOS detects my 80GB Western Digital IDE drive just fine.

Out of curiosity, I tried one of my Crucial 128GB SSDs with a SATA-IDE adapter, and the entire disk capacity was detected.

It seems like the issue is only with the CF card. I know that this disk isn't counterfeit and that the full capacity works just fine as a removable disk.

I know that this card has the "removable" bit set, so maybe this is triggering a different disk detection subroutine that does not have large disk support.

I also confirmed that my 3Ghz P4 533Mhz bus (SL6PG) is a Northwood core works just fine. The other processor I mentioned is a Prescott core, so it's not the bus speed that is the limiting factor, specifically.

This further supports the theory that the board labeled as "rev 1.3" is the newer board.

Problem solved 😀.

I have picked up one of these boards, because it had 3 ISA slots and PS/2 mouse and keyboard connectivity and some 'modern' features. However, I did not see this thread though initially. My board is a Rev B1, so far I have been very lucky it would seem, I did find a manual for it though before getting it, so was aware of the AGP socket and CF card slot being linked to the IDE 2 Master channel. I am running it mostly as a crazy MS-DOS/Windows 98SE monster. SB16 works like a charm without issues, but I am only using it with 16gb SSD's.

However, not deal breakers but did have a few issues/quirks:

1) The only thing I could not get working is my Gus PNP, which in an old setup I had working alongside my SB16 and HardMPU (i.e. everything disabled and epprom re-programmed, etc.), but with this board whatever I try it would not work, digging into the ports, port 260 (which I usually use for the Gus) seems to be used by something on the motherboard which I cannot change, so changed it on the Gus, but it would still not work, in fact every time I try to run a game with Gus music it will either crash, or I cannot hear the music at all, so all I can assess is there is a conflict somewhere

2) Regarding graphics cards, my 6200 PCI card works fine in Windows 98 and the on board graphics does not take over. However, like previously mentioned, despite being disabled in the BIOS it still shows up as a standard PCI graphics card in device manager (I did not install the Intel drivers as I had the 6200), along side the Nvidia device.

3) I run my computers through a KVM so can use the same keyboard, mouse and monitors and keep my desk/setup tidy(ish), the keyboard works fine but in Windows 98 one touch of the mouse and the pointer goes crazy and then the machine freezes. plug the same mouse directly into the PS/2 port on the machine instead of going through the KVM and it works find, which is odd never had that before. However, I can connect the KVM via USB instead of PS/2 and that works, so doing it that way at the moment, but DOS is not a fan of the USB mouse driver.

Overall though having a P4 3.06GHz running Windows 98 is rather cool, next is to upgrade it from 512mb to 2gb so time to find those software patches!

just got two of the same boards, rev b1.

1. pci video work fine, geforce 5200 , radeon 9250.
2. sound blaster 32 does not detects at all, crystal sound card magic s23a works just fine in dos, but immediately gets blue screen in windows 98 and no sound in XP.
Another crystal based card works just fine. I belive old non pnp cards are the issue for this board.

What if just replace 82845GV with 82845G? any chance to get agp working ? -)

It's probably not even wired up properly for normal AGP operation, even if that was feasible.

dm- wrote on 2024-05-20, 07:03:

just got two of the same boards, rev b1. […]
Show full quote

just got two of the same boards, rev b1.

1. pci video work fine, geforce 5200 , radeon 9250.
2. sound blaster 32 does not detects at all, crystal sound card magic s23a works just fine in dos, but immediately gets blue screen in windows 98 and no sound in XP.
Another crystal based card works just fine. I belive old non pnp cards are the issue for this board.

What if just replace 82845GV with 82845G? any chance to get agp working ? -)

REV. B00 - all my soundcards non-PNP, semi-PNP and PNP works just fine

Looking up the various DFI industrial boards, it looks like the DFI ITOX G4E620-N would be the better model to get. It has a proper 1.5V AGP slot and can support toggling between AGP and PCI graphics support.

Kahenraz wrote on 2024-05-20, 12:23:

It's probably not even wired up properly for normal AGP operation, even if that was feasible.

according to intel documentation for chipset 82845gv the signals for agp/ digital daughter board are mixed, and agp just disabled in this chipset.
so probably replacing the chipset to normal 82845g may unlock agp

dm- wrote on 2024-05-21, 04:18:

Kahenraz wrote on 2024-05-20, 12:23:

It's probably not even wired up properly for normal AGP operation, even if that was feasible.

according to intel documentation for chipset 82845gv the signals for agp/ digital daughter board are mixed, and agp just disabled in this chipset.
so probably replacing the chipset to normal 82845g may unlock agp

As well as the ability to disable onboard video, since the register controlling the toggle for onboard video is read-only and always enabled for 845GV/865GV.

I've been wondering why the GV variants were designed this way as leaving onboard video always enabled certainly would interfere with scenarios where a PCI video card is being used instead.

EDIT: From the datasheet it appears to be that AGP-related stuffs were removed/reserved for the GV variants, with certain AGP and onboard video related registers hardwired instead of configurable... though I'm still not sure about how things were implemented under the hood...

Wouldn't this also require modifying the BIOS? I suspect that this kind of thing would be very integrated there.

The onboard video is already connected, and the AGP "bus" was never designed to have more than one card attached to it.

Kahenraz wrote on 2024-05-21, 13:48:

Wouldn't this also require modifying the BIOS? I suspect that this kind of thing would be very integrated there.

It depends. If G4V620 (845GV) and G4E620 (845G) are identical in terms of BIOS and PCB design then the BIOS should in theory be able to deduct whether it's using a 845G or 845GV and offer options accordingly. If they have completely different BIOS designs then there's no guarantee.

That was the case with AIMB-865 between 1.0s1.2 and 1.0s1.3 (the former is 865G and the latter 865GV). The BIOS I dumped from 865GV revision can work on the 865G one, with options specific to 865G accessible and functional there.

Kahenraz wrote on 2024-05-21, 13:48:

The onboard video is already connected, and the AGP "bus" was never designed to have more than one card attached to it.

For GV variants those registers controlling AGP-related stuffs were hardwired to onboard video so it is always enabled, even when there's a PCI video card present. There were no middle ground (leaving neither enabled). Maybe 8x5GV were chip-harvested 8x5G under the hood...

BIOS may offer the option to "disable" onboard video but from my experience with AIMB-865 1.0s1.3 all it does is disable framebuffer allocation. Only on a real 865G board (same applies to 845G also) can the onboard video be fully disabled by that option.

i switched the chipset IC to RG82845G yesterday and... it's showing no signs of life, board is starting but on the POST card only dashes.
i got the IC from aliexpress (looks like new) and changed in local laptop service center (30$ to replace IC) to avoid any issues. But it is not showing any post code -(

luckily, last time my scrap yard got a lot of this boards, so i can just grab another