So I have changed some ram settings to match the specs of my ram. Everything was already on "Auto" or "Set from SPD" also voltage to 2.5v from 2.6v

AGP/PCI clock settings are set to 66.6/33.33

I have a performance mode option which is set to slow, dynamic overclocking in disabled.

One conflict was found in device manager and I have adjusted it manually.

That PCI universal serial bus device is still being problematic in device manager. Any attempt to change anything to do with it from device manager results in a locked system.

And this points me towards overlapping IRQs, ACPI tables incompatible with Windows 98, PCI steering and so on. And that means that you need memorize IRQ values during POST, compare it to the values in device manager, search for resource conflicts. After that go to BIOS and try to manually reallocate IRQs. If it doesn't work play ACPI modes, APM, PNP OS settings. But google it first and read it carefully. You can't damage anything, but with some settings combination you will need to reinstall OS from scratch before any further experiments.

How will I know if a reinstall is required?

I asked for something else 😀 Did you installed chipset drivers, reboot and install everything else OR install all you may think of chipset+graphics+sound+etc and reboot after.

I've been restarting each time it is requested.

Are you sure nusb36 isn't the issue ? Where did you get the motherboard ? Have you tried any other ram ? After reading thru this I cannot find any ref that you tried diff RAM, diff Video card or tried not using nusb32.

How will I know if a reinstall is required?

OS don't boot.

I've been restarting each time it is requested.

nice, again points towards ACPI

Horun wrote:

Are you sure nusb36 isn't the issue ? Where did you get the motherboard ? Have you tried any other ram ? After reading thru this I cannot find any ref that you tried diff RAM, diff Video card or tried not using nusb32.

Tried different RAM, also just tried nusb33 now and had the same lock up

Are these the IRQ conflicts that are giving me all the pain?

I can confirm in device manager that IRQ11 is very popular but Windows will not let me change anything. I don't see any options in BIOS to change them either. I can change them from "PCI/PnP" to "ISA/EISA".

I can also enable "APIC ACPI SCI IRQ". This causes the infamous os won't boot. But I'll do a fresh install and see how it likes it then.

Why don't they space out a bit more? Everything is on Auto and I've disabled all the onboard ports that I don't need.

APIC ACPI SCI IRQ enabled caused bluescreens, other errors.

Are these the IRQ conflicts that are giving me all the pain?

The way how windows 98 handle it. That is all about evolution in ACPI and PCI IRQ steering.

Google it and read something, then try to do some adjustments.

Installed again with APIC disabled and with the /p i switch which forces Windows NOT to use ACPI/PnP.

Things moved along well. Until I installed the nusb drivers. Maybe it is as Horun suggested. I just found an old nusb32 version so I'll try those. But why would I be the only person who can't have USB support bwith those drivers?

I don't know if I'm closer or further from a solution.

The device that is giving me all the problems is the "Intel(R) 82801EB (ICH5 Family) USB2 Enhanced Host Controller - 24DD". It doesn't matter what drivers I use, it's always the same lockup. It is insisting on being assigned to IRQ0 and also in a memory range that was already occupied by "System board extension for ACPI BIOS". Changing these values in either normal or safe mode do not persist after reboot.

How do I measure battery voltage while system is running?

Put the positive probe of multimeter on top of the battery, negative probe against the PSU case. There's lots of ways to get the ground but the PSU case is the easiest and safest way when the power is on.

There's a middle ground between "good battery" and "dead battery", in which the battery's ability to do it's job falls into a gray area and it's effect on the system becomes uncertain.
Some motherboards draw from the battery when the system is running, which is why they sometimes don't boot when it's dead.
When the power is off, it's possible a marginal battery could corrupt the CMOS settings, although those really ought to be powered by the PSU's standby supply.
I'd measure both with system on and when it's off.

5V and 12V can be measured from the red and yellow pins of an unused 4pin power connector that goes to a hard/CD drive.
12V for the CPU (4pin) can be measured by pushing the probe into the back side of one of the yellow wires on that connector while it's plugged in and running. Negative probe goes to the PSU casing.
3.3V can be measured by doing the same at an orange wire at the ATX power connector.
5Vsb is the purple wire. It's still relevant when the system is running.
CPU Vcore can be measured on the large tab at the backside of half of the MOSFETs near the CPU. The other half will have the upstream input voltage there (should be +12V). But if this board has heatsinks in the way, then another place to find Vcore is at either of the solder points attached to an inductor (copper coil). Checking this is optional of course, don't do it if you're not comfortable.
If you ever find yourself poking the probe somewhere that you're worried about shorting something, then wrap the probe with tape so only a minimal tip is exposed. But for the stuff I mentioned it shouldn't be necessary.
It's possible to find Vmem and others, but it gets more difficult.
If you check these things, try to observe not only if they look low, but also if they appear to be sagging under a load.

I don't mean to turn this voltage thing into too much of a distraction though. IMO checking voltages is easy to do and best practice on any PC, even moreso if it's exhibiting problems. Hardware stability is the foundation the system is built on. 😀
However, your symptoms sound like a resource allocation issue with Win9x, and it looks like you've narrowed it down to that Intel NIC. Is it stable in another OS?
I had a P3 Thinkpad that started randomly locking up, turned out the Intel NIC in the laptop was failing. In that case it was hardware - linux logged an error message that revealed the NIC was failing, and replacing the NIC fixed it.

Following this guide https://www.techspot.com/community/topics/fix … q-conflict.886/ which basically says uninstall all the conflicting hardware besides the one that I want to be there. Then, on reboot windows will put it on a more appropriate IRQ and memory range. This seems to solve the problem and also lets me use a usb drive. Until I reboot. Then the USB2EHC needs to be disabled in safe mode again.

How many usb connections you wish have active at all time? According to your BIOS screenshot USB controller can use IRQ 3 and 5, maybe you can lower number of active usb ports in BIOS (idk if motherboard allow it). Or you may disable USB 2.0 (Enchanced/Hi-Speed). Any HID (keyboard/mouse/gamepad) will work just fine without any issues or slowdowns, but it will be painful to use external drives.

_UV_ wrote:

How many usb connections you wish have active at all time? According to your BIOS screenshot USB controller can use IRQ 3 and 5, maybe you can lower number of active usb ports in BIOS (idk if motherboard allow it). Or you may disable USB 2.0 (Enchanced/Hi-Speed). Any HID (keyboard/mouse/gamepad) will work just fine without any issues or slowdowns, but it will be painful to use external drives.

I'd probably need

Mouse
Keyboard
Maybe a steering wheel
And definitely some external drives - Not all the time, but I'd like to use it as the main way to transfer files to this PC

I can disable the main culprit and things seem to work ok (at least for the usb keyboard). But it looks like anything to do with a removable drive will crash it.

However, your symptoms sound like a resource allocation issue with Win9x, and it looks like you've narrowed it down to that Intel NIC. Is it stable in another OS?

I just booted from a live linux mint cd and copied about 4gb to the hdd with no problems. Same goes for Windows XP even after installing those chipset drivers. Both used the sam usb port and same drive.

That is definitely something wrong with how MSI compiled BIOS. I had Intel boards on 865PE and 875 running Win98, they done many fixes towards this OS, and it was pretty stable. My own journey into Pentium 4 was started with MSI 845PE something with W98 too, i never had such troubles. We had many PCs at work built around different Asus and Gigabyte 848 & 865 boards and they worked fine (till all of them died) under 98 since it was mandatory for our software. I had similar troubles with Supermicro 865/875 boards.

Do you mean that the BIOS has an incorrect setting somewhere or is it a deeper problem than just a setting?

I'm considering that it might be time to start a new thread on this since it started out as an unknown lockup thread and has now become much more specific.

Just look at this quick explanations

https://en.wikipedia.org/wiki/Advanced_Config … Power_Interface
https://wiki.osdev.org/ACPI
https://docs.microsoft.com/en-us/windows-hard … cription-tables

The way how developer tune all this settings will affect how OS will perform and configure devices.
I think you don't need another thread, try to search for other BIOS versions, maybe even downgrade it. You need to find one magic combination of all available settings about ACPI, PNP OS, IRQs, APM, APIC that will be stable and not causing this lockups with USB. Or don't use USB storage under W98, install WinXP and dual boot in case you need it.

This might be the end. There is a BIOS setting I've changed.

From the manual...
"USB Device Legacy Support
Set to All Device if your need to use any USB 1.1/2.0 device in the operating
system that does not support or have any USB 1.1/2.0 driver installed, such as
DOS and SCO Unix. Set to No Mice only if your wan t to use any USB device
other than the USB mouse. Setting options: Disabled, Keyboard+Mouse, All
Device."

I have changed it to Disabled.

Also from the manual...
"If you want to boot from any of the USB-interfaced devices,
please set USB Legacy Support to All Device."

If it solve the issue that is great. I always have it set to legacy just in case.

This problem is similar to the one WinRaid forum guys are having with XP and newer Win10-only Skylake boards with USB3/Newer HDAudio peripherals. HAL, PnP issues and memory corruption are everywhere in XP with these boards.

ACPI is an always evolving technology. If you want integrated USB2 and Win98, you really want an USB2 motherboard with Win98/WinME compatibily specified in box. That ensures they compiled the ACPI tables and ACPI/PCI routing information in a format which can be understand by the old crusty ACPI interpreter built in Win98 ACPI.sys driver. Going to legacy setting will not solve it as the PCI routing tables could have also be compiled in an NT/XP compatible format, and onboard USB2 is very reliant in these as it uses a lot of IRQs. You will have to disable USB to get it to boot somewhat.