Yes, 430TX chipset.
IDE devices detected as UDMA4 will not get DMA tick in win98 because board can only do UDMA2 but reports 4. Win98se then calls it a day and goes pio.
Chkcpu explained this in great detail in his first post in this thread.

ux-3 wrote on 2024-07-18, 13:48:

Yes, 430TX chipset.
IDE devices detected as UDMA4 will not get DMA tick in win98 because board can only do UDMA2 but reports 4. Win98se then calls it a day and goes pio.
Chkcpu explained this in great detail in his first post in this thread.

Sorry about the confusion. I had that wrong. Thanks for taking the time to set me right.

ux-3 wrote on 2024-07-18, 11:46:

Unfortunately, my BCN IN530 passed away. I replaced it with a Gigabyte GA-586ATX3 (=rev3), patched it to latest 4.0 bios. Is a […]
Show full quote

Chkcpu wrote on 2024-06-24, 19:40:

Fixing the UDMA bug in an Award BIOS is not that hard, I only have to find the time to do it. 😀

Cheers, Jan

Unfortunately, my BCN IN530 passed away.
I replaced it with a Gigabyte GA-586ATX3 (=rev3),
patched it to latest 4.0 bios.
Is a UDMA2 bios for this board already in existance and where would I have to look for it?
Thanks in advance.

Hi ux-3,

I’m sorry to hear about the IN530 board, but that GA-586ATX3 is a nice replacement.
However the latest 4.0 BIOS from 07/1998 has the usual limitation of that era:
- 32GB IDE HDD limit bug
- Win98 UDMA bug
- No K6-2+/K6-III+ support.
But this BIOS does support the K6-2 up to 400MHz, the WinChip 2, and the Tillamook!

I’m unaware of a patched BIOS with the UDMA fix for this board, but next week I have more time and can do the 32GB and UDMA bugfixes on both the 586STX2 and 586ATX3 BIOSes.

I will keep you posted,
Jan

Chkcpu wrote on 2024-07-18, 20:24:

I’m sorry to hear about the IN530 board, but that GA-586ATX3 is a nice replacement.

Well, I am still in mourning though... 🙁 It may not have been king of the hill in its time, but it was great for running older DOS stuff...and very compact.

However the latest 4.0 BIOS from 07/1998 has the usual limitation of that era: - 32GB IDE HDD limit bug - Win98 UDMA bug ... But […]
Show full quote

However the latest 4.0 BIOS from 07/1998 has the usual limitation of that era:
- 32GB IDE HDD limit bug
- Win98 UDMA bug
...
But this BIOS does support the K6-2 up to 400MHz

I expected and confirmed it. Same manufacturer, very similar boards.

next week I have more time and can do the 32GB and UDMA bugfixes on both the 586STX2 and 586ATX3 BIOSes.

Are you using the procedure that Sappa3dfx outlined? I could try that myself. I guess I should learn doing this, at the rate I run through these boards. Should I do a complete fix or only those two items? Are there any critical steps to be aware of, before one takes them?

I will keep you posted,

Thanks!

There is a BIOS for that board listed here that has been patched to support 128 gigabyte hard drives. You would be better off waiting for Jan's BIOS which will have that and much more fixed, but just in case you want to try it.

https://web.archive.org/web/20051105011452/ht … ex.php?count=-1

Hi ux-3,

I did found an up-to-date GA-586ATX3 Award BIOS afterall. 😀
It is a 04/2000 BIOS upgrade from Unicore Software, specifically for the GA-586ATX3.
It contains all fixes, like the UDMA bugfix and 32GB limit fix for full 128GiB IDE HDD support. It also supports the K6-2+/K6-III+ CPUs.

Please let us know how this BIOS works on your GA-586ATX3.

Repo Man11 wrote on 2024-07-18, 21:45:

There is a BIOS for that board listed here that has been patched to support 128 gigabyte hard drives. You would be better off waiting for Jan's BIOS which will have that and much more fixed, but just in case you want to try it.

https://web.archive.org/web/20051105011452/ht … ex.php?count=-1

The link from Repo Man11 is also a great resource of BIOSes that have the 32GB HDD limit fixed for proper 128GiB support. I contributed to this list myself, but these patched BIOSes don’t contain fixes for other bugs.

ux-3 wrote on 2024-07-18, 20:51:

Are you using the procedure that Sappa3dfx outlined? I could try that myself. I guess I should learn doing this, at the rate I run through these boards. Should I do a complete fix or only those two items? Are there any critical steps to be aware of, before one takes them?

No, I use a different BIOS patching method.

The BIOS Patcher by developer “apple_rom” is a nice tool though.
Member soggi wrote about this tool in Re: Modding support for AMD K6-2 into a Socket 5 Award BIOS? and member Scorp did a video about it in his Necroware YouTube channel: https://www.youtube.com/watch?v=VwZT40sRMzM

The latest stable version 4.23 works only on compressed Award BIOSes v4.5x and v6.x and usually works well, as Sappa3dfx showed it this thread. But there have been several cases where it bricked a system, so I would only recommend it when you’re board has the BIOS flashchip in a socket and you have an EEPROM programmer to re-program the BIOS chip if needed. Luckily the developer(s) have put in a fall-back feature that allows you to boot with the original BIOS. Just hold down the “–“ (minus) key on the numeric keypad during boot.

My BIOS patching method of directly editing the BIOS firmware is a different (and more time consuming) approach but works on both compressed and uncompressed BIOSes. It also allows me to fix rare bugs that apple_rom’s tool doesn’t know about.
I have my patched BIOSes tested on real hardware before publishing them on my k6plus webpage. These “ready to go” BIOSes then work like downloading and flashing a BIOS update from the board manufacturer.

Cheers, Jan

Many thanks but I am not at home at the moment, will look into it asap.

Chkcpu wrote on 2024-07-19, 09:54:

I did found an up-to-date GA-586ATX3 Award BIOS afterall. :) It is a 04/2000 BIOS upgrade from Unicore Software, specifically fo […]
Show full quote

I did found an up-to-date GA-586ATX3 Award BIOS afterall. 😀
It is a 04/2000 BIOS upgrade from Unicore Software, specifically for the GA-586ATX3.
It contains all fixes, like the UDMA bugfix and 32GB limit fix for full 128GiB IDE HDD support. It also supports the K6-2+/K6-III+ CPUs.

The attachment 2A59IG0E.ZIP is no longer available

Please let us know how this BIOS works on your GA-586ATX3.

I downloaded the bios. The readme says:
"Super I/O: W83877F (Check this first!!)"
I did check, but my chip is W83877TF.
Is that ok?

ux-3 wrote on 2024-07-25, 18:22:

I downloaded the bios. The readme says: "Super I/O: W83877F (Check this first!!)" I did check, but my chip is W83877TF. Is that […]
Show full quote

Chkcpu wrote on 2024-07-19, 09:54:

I did found an up-to-date GA-586ATX3 Award BIOS afterall. :) It is a 04/2000 BIOS upgrade from Unicore Software, specifically fo […]
Show full quote

I did found an up-to-date GA-586ATX3 Award BIOS afterall. 😀
It is a 04/2000 BIOS upgrade from Unicore Software, specifically for the GA-586ATX3.
It contains all fixes, like the UDMA bugfix and 32GB limit fix for full 128GiB IDE HDD support. It also supports the K6-2+/K6-III+ CPUs.

The attachment 2A59IG0E.ZIP is no longer available

Please let us know how this BIOS works on your GA-586ATX3.

I downloaded the bios. The readme says:
"Super I/O: W83877F (Check this first!!)"
I did check, but my chip is W83877TF.
Is that ok?

Yes, this is oke.
The W83877TF adds only little changes over the W83877F. The 877TF adds ACPI support for the devices connected to it, but if the BIOS doesn’t use that functionality it treats the 877TF just like an 877F with regular Device Power Management support.

I’m sure the original GA-586ATX3 BIOS didn’t have ACPI support, but maybe the Unicore BIOS does have it and can use ACPI on your 877TF Super I/O. 😀
Just check the BIOS Setup menus for any ACPI settings.

Jan

Chkcpu wrote on 2024-07-25, 20:42:

...

Meanwhile, I flashed the 586STX2 (SIS chipset) with a bios I made with bios patcher 4.23. The patch and flash worked.
However:
If I leave UDMA in bios on OFF, nothing changes.
If I turn UDMA in bios ON, the new patched bios will display UDMA2 during boot. (so this part is working as intended)
With the old bios, this will then untick DMA in win98 during next boot.
With the new bios, the startup of win will hang indefinately.

Looks as if I traded one problem with another. (Shouldn't be a problem to revert back though)

This was using a UDMA4 CF card. I will try a few other storage media.

SKIP TO NEXT POST!
OK, I experimented a bit further on the 586STX2 (SIS chipset):
I tried CF media from another manufacturer, same behavior.

In the past, using SSDs did lead to disk corruption once windows did the first post install boot.
Now, with the new bios and UDMA OFF, it will reboot windows BUT keep DMA unticked. While this is a small improvement, it will not make SSDs more attractive than CFs.

The odd thing is: All this media worked fine (UDMA & DMA tick) with the SIS530, which uses the SIS 5595. Totally different behavior.
Unless there is something essential that I missed, there is little use in upgrading this bios, as I can't use SSDs nor do I get UDMA with CF. The board remains usable with CF dma at pio4.

As I have several PCI controller cards for IDE and mixed IDE/SATA, I will just try out those now.

I haven't yet dared to flash the intel board with a modified bios and see what happens there.

GOT IT!

I first tried a Silicon Image IDE-133 raid controller, which would basically display full speed: CF max read is at 30MB/s and SDD reads at ~40 MB/s. That worked ok.

I then decided to try the onboard IDE again and swapped to the generic Standard-Dual PCI IDE controller driver. (dates 22-10-2005, so perhaps from an update)
Now I can run both CF and SATA at UDMA2 (bios patch) and get the DMA check mark in win98. CF reads at 20MB/s max and SSD at 29 MB/s. Both is good enough for early win98.

I just confirmed: booting with the unpatched bios will not give the DMA mode check.

I later found out that the SIS driver works with udma as well. Just the DVD drive refuses to use UDMA, as soon as that DVD is forced on pio4, the rest works in UDMA. Master/slave makes no difference

Edit: Further experimenting shows that one CF card will do UDMA if connected in a plug-in-ide-port adapter, but not if connected to an adapter via IDE cable. Another card doesn't care and works either way. Seems that the IDE signal is at the fringe here. This would explain several confusing results before.

Chkcpu wrote on 2024-07-25, 20:42:

I’m sure the original GA-586ATX3 BIOS didn’t have ACPI support, but maybe the Unicore BIOS does have it and can use ACPI on your 877TF Super I/O. 😀
Just check the BIOS Setup menus for any ACPI settings.

I have patched and flashed my bios successfully. DMA check works fine now.
Not sure if I want to take a risk with an unknown bios now that everything works so far.

ux-3 wrote on 2024-07-30, 18:48:

Chkcpu wrote on 2024-07-25, 20:42:

I’m sure the original GA-586ATX3 BIOS didn’t have ACPI support, but maybe the Unicore BIOS does have it and can use ACPI on your 877TF Super I/O. 😀
Just check the BIOS Setup menus for any ACPI settings.

I have patched and flashed my bios successfully. DMA check works fine now.
Not sure if I want to take a risk with an unknown bios now that everything works so far.

Hi ux-3,

Thanks for the detailed follow-up. Nice that the BIOS Patcher tool worked for the 586STX2 BIOS and you got UDMA working. 😀

How about the GA-586ATX3 BIOS? Are you going to use the BIOS Patcher tool on that as well, or are you going to use the Unicore BIOS?
I’ve tested the 586ATX3 Unicore BIOS here in the 86Box emulator and it ran perfectly fine. 😉

Cheers, Jan

Chkcpu wrote on 2024-07-31, 08:10:

How about the GA-586ATX3 BIOS? Are you going to use the BIOS Patcher tool on that as well, or are you going to use the Unicore BIOS?

Hi.
I quoted your last post where you said 586ATX3. So yes, after I got the STX2 patched sucessfully, I also patched the ATX3. I wanted to you to know, so you don't spend your time on it. As I patched the latest bios 4.0, I am not sure if I should risk going back to an earlier bios, albeit fixed.

I noted some odd remaining errors. The STX2 mistakes 16 GB CF cards from Transcend as 8GB, while the ATX3 has the same problem with SanDisk 16 GB.
The ATX3 also does some odd power LED blinking when shut down but PSU is still switched on. I have not seen that before. It keeps blinking for a few minutes when you unplug or switch off the PSU. But it doesn't yet power down with button in win98. If you push the button, it will just go off. So perhaps ACPI would be nice.

Meanwhile, I could revive the IN530 board, which uses an AMI Bios. I would like to "unhide" the "disable cache" feature in a later version of the bios. But having 32 GB in the early bios isn't a severe limitation.
The board also had a somewhat similiar UDMA issue, it would not DMA check when UDMA4 and UDMA2 were on the same IDE port. Using a 40 wire cable forced UDMA2 on both devices, and made it work. The actual speed of the drive needs no UDMA 4 anyway. Only good for moments when bandwidth matters.

ux-3 wrote on 2024-07-31, 09:34:

Hi. I quoted your last post where you said 586ATX3. So yes, after I got the STX2 patched sucessfully, I also patched the ATX3. […]
Show full quote

Chkcpu wrote on 2024-07-31, 08:10:

How about the GA-586ATX3 BIOS? Are you going to use the BIOS Patcher tool on that as well, or are you going to use the Unicore BIOS?

Hi.
I quoted your last post where you said 586ATX3. So yes, after I got the STX2 patched sucessfully, I also patched the ATX3. I wanted to you to know, so you don't spend your time on it. As I patched the latest bios 4.0, I am not sure if I should risk going back to an earlier bios, albeit fixed.

I noted some odd remaining errors. The STX2 mistakes 16 GB CF cards from Transcend as 8GB, while the ATX3 has the same problem with SanDisk 16 GB.
The ATX3 also does some odd power LED blinking when shut down but PSU is still switched on. I have not seen that before. It keeps blinking for a few minutes when you unplug or switch off the PSU. But it doesn't yet power down with button in win98. If you push the button, it will just go off. So perhaps ACPI would be nice.

Meanwhile, I could revive the IN530 board, which uses an AMI Bios. I would like to "unhide" the "disable cache" feature in a later version of the bios. But having 32 GB in the early bios isn't a severe limitation.
The board also had a somewhat similiar UDMA issue, it would not DMA check when UDMA4 and UDMA2 were on the same IDE port. Using a 40 wire cable forced UDMA2 on both devices, and made it work. The actual speed of the drive needs no UDMA 4 anyway. Only good for moments when bandwidth matters.

Hi ux-3,

Right, I misunderstood.
So you successfully patched the GA-586ATX3 BIOS as well, nice!

Note that the latest 586ATX3 v4.0 BIOS is from July/1998 and therefore uses a mid-1998 core. Although the BIOS Patcher tool removed several bugs and improved CPU and HDD support, your patched version v4.0 is still an 1998 BIOS at heart.
The 586ATX3 Unicore is almost 2 years younger, being from April 2000, so it has an updated core.

I’ve checked the Unicore BIOS for ACPI, and it has indeed both APM and ACPI support. The POWER MANAGEMENT menu allows to control both features.
So this Unicore BIOS may be worth a try when troubleshooting the Win98 shutdown issues on your ATX3 rig.
Maybe even the SanDisk 16GB issue will go away. 😉

Great that you could revive the IN530 board! I haven’t looked for the disable cache option in the later AMI BIOS versions yet, but I can give it a try. It may take some time though…

Cheers, Jan

Chkcpu wrote on 2024-08-02, 18:22:

Great that you could revive the IN530 board! I haven’t looked for the disable cache option in the later AMI BIOS versions yet, but I can give it a try. It may take some time though…

Hi ux-3,

I had some time this morning and I took a quick look at the latest v1.29 AMI BIOS for the FR520 as available on TRW, and I already found something that may help you.

I looked at all the BIOS options, both active and hidden, and their default values for Optimal defaults and Failsafe defaults. There is surprising little difference between these settings and this is the short list:

1Option name			Optimal	Failsafe
2
32nd Boot Device		CDROM	IDE-0
4*External Cache		Enabled	Disabled
5*System BIOS Cacheable	Enabled	Disabled
6Power Management/APM 	Enabled	Disabled
7Video Power Down Mode	Suspend	Disabled
8OnBoard Parallel Port	378h/IRQ7	Auto
9

All other BIOS options get the same value , whether you select Set Optimal Default or Set Failsafe Default in the BIOS Setup.
The two options with the asterisk are hidden, but can be influenced by selecting Failsafe Default. As one of these options is the External Cache, you should be able to Disable it by simply selecting the Failsafe Default, and Enable it again via Set Optimal Default! 😀

I don’t know yet if I can activate these hidden options without bricking the BIOS, but that work is for another time. In the meantime, I hope that the above workaround helps controlling the onboard cache.

Cheers, Jan

OK, many thanks.
But can I load "failsafe" and then change visible parameters (like Sound, LPT, Boot drives)?

I guess I will test it when I find the time.

ux-3 wrote on 2024-08-03, 18:48:

OK, many thanks.
But can I load "failsafe" and then change visible parameters (like Sound, LPT, Boot drives)?

Yes, after selecting Failsafe defaults, you can change any visible option to your liking. The hidden External Cache option will stay Disabled, until you select Optimal defaults.

Jan

Chkcpu wrote on 2024-08-02, 18:22:

Note that the latest 586ATX3 v4.0 BIOS is from July/1998 and therefore uses a mid-1998 core. Although the BIOS Patcher tool remo […]
Show full quote

Note that the latest 586ATX3 v4.0 BIOS is from July/1998 and therefore uses a mid-1998 core. Although the BIOS Patcher tool removed several bugs and improved CPU and HDD support, your patched version v4.0 is still an 1998 BIOS at heart.
The 586ATX3 Unicore is almost 2 years younger, being from April 2000, so it has an updated core.

I’ve checked the Unicore BIOS for ACPI, and it has indeed both APM and ACPI support. The POWER MANAGEMENT menu allows to control both features.
So this Unicore BIOS may be worth a try when troubleshooting the Win98 shutdown issues on your ATX3 rig.
Maybe even the SanDisk 16GB issue will go away. 😉

I tried the unicore bios and it turned out to be a disaster. The board failed to recognize any IDE devices. After flashing back the 4.0 Bios, the board no longer posts a picture at all nor does it boot from even floppy.

I now need a way to flash the bios chip with the former 4.0 bios. What hardware do I need? Or in case it isn't economical to buy a programmer, is there a service for this?

I just contacted a service in my country. If it is affordable, I'll just go that route.