Duffman wrote on 2022-10-26, 09:12:

@rasteri

Have you tried adjusting these settings to see if it helps?

Sorry I haven't been home for the last few days, I'll try to get to it soon.

I've fiddled with the "Standard IDE Compatible" settings before but it made no difference. Although that was before we had your driver.

Not tried the DMA/PIO settings yet though.

@rasteri

Some ideas for your hypothecical WeeCee Max (Ultimate Win98 machine) -

SOM - SOM304D3-74EINE1-HS (Although a Vortex86EX with MMX if possible)

PCIe 1.1 x2 -> 2x ASMedia ASM1083 PCIe-to-PCI bridges - 4x PCI lanes
- 1x NV17 - https://www.utsource.net/au/itm/p/8929685.html - Nvidia Quadro 4
- 1x VIA VT6241A - https://www.utsource.net/au/itm/p/6389078.html - Universally compatible IDE/SATA controller
- 1x Aureal AU8830 - https://www.utsource.net/au/itm/p/446103.html - Aureal Vortex 2
- 1x ??? - anything else we might need?

We'll keep the CS5237 for some additional sound compatibility in DOS too.

What do you think? wouldn't this be awesome??

Duffman wrote on 2022-10-26, 13:23:

Some ideas for your hypothecical WeeCee Max (Ultimate Win98 machine) -

The way I see it, weeCee is and will always be a DOS gaming machine, not a Windows one. DOS gaming PCs have totally different requirements to Windows PCs.

The Vortex86 just isn't a great platform to run Windows 98 games on - it's really just an overclocked 486 with some Pentium instructions bolted on, and has compatibility issues with Windows games. Even the high-end DX3 doesn't support stuff like SSE. There are other SBCs/thin clients that do a much better job of running 98 than the WeeCee ever will. I'm only using Vortex86 because of its ISA bus support.

If I were to design a Windows 98 gaming PC it would use a completely different architecture - probably ditch ISA entirely and use a more contemporary processor/sound chip/graphics chip. But I don't have the EE skill neccesary to do high-speed memory routing (like would be required for the NV17) so don't hold your breath.

The weeCee max, if I ever do it, will be another DOS gaming machine, just with more ports.

@Rasteri

Ah OK, no worries then.

An imaginary windows 98 build I thought up browsing PC parts on ebay.

MOBO: https://www.ebay.com/itm/195198491427
Asrock QC6000M Socket FT3 micro ATX
Has Embedded CPU
and ACPI 1.1 so 98's ACPI driver should work.

GPU: https://www.ebay.com.au/itm/314202700347
NIB - Gigabyte X300 SE PCI-E 128MB
Not the fastest, but has 98 drivers

SOUND: https://www.ebay.com.au/itm/172740016380
AUDIGY 4 on a PCI-E bridge chip.
I wounder if the 98 Audigy 2 drivers will work with it if I just add the Audigy RX hardware ID to the INF file?
The chip is an emu10k from what I can see, so maybe?

USB 2.0: https://www.ebay.com.au/itm/170641507292
PCI-Express 4 Ports USB2.0 Expansion Card PCI 4xUSB2.0 MOSCHIP MCS9990
I've used this in windows 98 before with the NUSB drivers, no problems.

NETWORK:
Motherboard chip Realtek RTL8111H has 98 drivers (unless I'm mistaken? I'll use USB Ethernet otherwise).

So, what do you think? Should I build this?

@Rasteri @Nachtrave

I decided I'll go ahead with this build!
With these parts I reckon I can get everything except for the onboard USB 3.0 up and running in 98SE.

I'll let you know how it goes.

I emailed DMP about the IDE Controller problem, I'd asked for the source code for the rdcatapi.sys driver but of course it's not available.

This is the message they sent me:

Thanks for using Vortex86DX. […]
Show full quote

Thanks for using Vortex86DX.

The source code is not available. But I’ve checked procedure to install Window98, you don’t need special driver, it should work directly with standard driver.

As you can see in attached photos ( it’s in Chinese, but some English words are readable)

1, Search “Standard Dual PCI IDE Controller)

2, Display all driver list, and pick the suitable one

3, Select the “Standard Dual IDE Controller [5-5-1999]

4, Install that driver

5, Restart

6, After restart, In System -> Setting, you will see the Standard Dual PCI IDE Controller. For Dual IDE channel setting, please select “Only enable primary IDE Channel”

The BIOS setting is also attached for your reference.

Best regards,

Gary Chen / DMP Electronics INC

www.vortex86.com

The last part is interesting. - please select “Only enable primary IDE Channel”

I guess try that?

@rasteri
Hey guys..

Looking to maybe make a weeCee of my own; have a couple questions.

1 - Where do I buy a SOM for cheap? Here in the US direct from icop-shop it's 200 - ish bucks, and that doesn't include shipping. Shipping is more than 80 bucks, and you can pay 375 for first class! I'm not spending 350 bucks on just the SOM.

2 - Does the SOM come with that metal thing, and is that metal thing required? icop-shop shows it without it. Worried that it won't come with it and that it won't fit.

3 - How do I get the connectors for the PCB from icop-shop? Do they come with it? I can't find any place to buy them because I don't know the name.

4 - How hard is it to solder the components? I've only done some through-hole soldering and a practice board with some solder pads for surface mount. But I'm a little worried that it is going to be hard.

5 - Anything else I ned to know?

Thanks so much, and this project is so cool!

I FINALLY got a mouse working. Third PS/2 splitter I tried. Do not get them from Amazon 😒
EBAY FTW. The one that is black with purple and green connectors from eBay worked for me finally.

@Duffman I tried this fix... it was interesting.
After selecting the "Standard Dual PCI IDE Controller" and rebooting, I got the same blue screen that indicates the 16 vs 32bit issue.
Then I changed the BIOS "Onboard PCI IDE controller" setting to "Primary" from "both" as shown in your screenshot.
Then it booted fine into windows. Partial success! But there was still an exclamation point on the driver.
Then I followed the instructions to change it from "both IDE channels" to "Only enable primary IDE Channel" and rebooted
same blue screen...

@fluxrez

Thanks for trying.
If it comes down to it I'll backport XP's atapi.sys if I have to, but I'm still hopeful we won't have to do that.

@rasteri

Would you be willing to risk trying to flash an ITE BIOS onto the IDE controller on the SOM? Like with a hardware EEPROM writer?

Re: Requesting BIOS files for ITE IT8212F ATAPI/RAID Controller

Duffman wrote on 2022-10-28, 15:11:

@rasteri

Would you be willing to risk trying to flash an ITE BIOS onto the IDE controller on the SOM? Like with a hardware EEPROM writer?

Re: Requesting BIOS files for ITE IT8212F ATAPI/RAID Controller

The BIOS on the Vortex86 is internal to the chip, so no chance of recovery if it goes wrong...

@rasteri

Oh wow, really? bummer...

Welp, Do you want me to try and do the backport of XP's atapi.sys driver then?

Duffman wrote on 2022-10-28, 15:54:

Welp, Do you want me to try and do the backport of XP's atapi.sys driver then?

Honestly, I really don't care about the disk performance in Windows 98 😀 It works fine in DOS and that's all that matters to me.

I'm skeptical that it's even a driver issue at all, really... it needs more diagnosis (for example, nobody's even tried using a regular hard drive to see if it's an SDcard-only issue or one that affects the entire controller). And I'm not interested enough in the problem that I'm willing to put the time in to research it.

@rasteri

Can you do a test install of Windows ME to see if it shows the same symptoms? I'd imagine it would, but I'd like to get proof if you can.

Duffman wrote on 2022-10-28, 17:45:

@rasteri

Can you do a test install of Windows ME to see if it shows the same symptoms? I'd imagine it would, but I'd like to get proof if you can.

Yeah that is a good idea.

@rasteri

Same result on Windows ME? or it works?

rasteri wrote on 2022-10-28, 15:17:

Duffman wrote on 2022-10-28, 15:11:

@rasteri

Would you be willing to risk trying to flash an ITE BIOS onto the IDE controller on the SOM? Like with a hardware EEPROM writer?

Re: Requesting BIOS files for ITE IT8212F ATAPI/RAID Controller

The BIOS on the Vortex86 is internal to the chip, so no chance of recovery if it goes wrong...

If the SOM304RD-VI is anything like the SOM-128-EX, it probably has some sort of small eeprom chip soldered on the carrying board. If you've got a programmer with a sop/soic clip attachment, it's possible to do a recovery programming of the original BIOS (save it first, of course) without lifting the chip. I've done this on the SOM-128-EX lots of times when messing up my BIOS.

Duffman wrote on 2022-10-29, 05:27:

Same result on Windows ME? or it works?

Just tried it.

Windows ME has the exact same problems loading the driver - but with default Microsoft drivers the speed is MUCH faster than on 95/98(about the same as XP).

I think I've just proved that the speed problems aren't caused by inability to load drivers.

If I had to guess, I'd say that it could be something to do with how 98 accesses hard disks being suboptimal for SD cards. Maybe messing with FAT32 sector sizes would improve things.

I think the next test has to be to run Windows 98 off a standard IDE hard disk rather than from an SD card. I have a vortex86dx board I can test this on, but it has no video card so I'll have to do some hacking