The monitor will not take an IRQ.

Changing the CPU should not have changed anything.

Does the motherboard have the latest BIOS available or at least a new enough version that supports the 933?

Have you tried cleaning the edge connector contacts on the CPU? Rubber pencil eraser is good to get the oxidation off. You can also use electronic contact cleaner.

Have you tried resetting the BIOS to default settings via the BIOS clear jumper on the motherboard?

What is the brand and model of the power supply and how old is it?

It might be also your video driver, try to do a fresh new install of your video driver or try a different version.

This kind of happend to me but with a different hardware, I installed the wrong sound driver and my Windows 98 frozed and after that I couldnt were able to boot it back to normal, It only let me do safe mode, so I removed my sound driver in safe mode, and system boot up normal again, I re install the correct driver and everything went back to normal.

You can give it a try and see if that works.

Hello,

• I updated to the latest bios available
• Connectors look clean
• PSU is a 350W "POWMAX"
• BIOS Settings were reset
• Couldn't install any driver since the system freezes at boot

Is this a Slot 1 Pentium 3, or are you using a socket 370 P3 with a slotket?

bloodem wrote on 2025-03-25, 06:09:

Is this a Slot 1 Pentium 3, or are you using a socket 370 P3 with a slotket?

Sorry, forgot to mention. It's a Slot 1 CPU on a Slot 1 system.

Interesting. The only thing I can think of is a power delivery issue (i.e. underpowered MOSFETS that can handle the 600 MHz SKU, but not the 933 MHz) . I had the same situation with an Amptron PII-3100B board, which was freezing with higher clocked CPUs, but worked fine with slower/less power hungry CPUs. I fixed the issue by replacing the two MOSFETs, which were part of the synchronous buck converter.

However, in your case, I would imagine that the newer Soyo board should already have enough power to handle any CPU. Can you show a close-up picture with the CPU voltage regulation circuitry?
Also, did you confirm the P3 933 MHz CPU works on any other board?

I took pictures of the area around the CPU, one should be the voltage regulator, unless is under the Soyo heatsink.
I noticed a slightly bulging capacitor (mb1, TC17). Could that be the issue?

Vio voltage may be unstable due to TC17. The nearby AIC1084CM is a 3.3V linear regulator. Looking at the board image, there is nothing else that seems to be generating the Vio voltage.

lazalius wrote on 2025-03-25, 07:13:

I took pictures of the area around the CPU, one should be the voltage regulator, unless is under the Soyo heatsink.
I noticed a slightly bulging capacitor (mb1, TC17). Could that be the issue?

Yeah, those 56 Amps 76129S MOSFETs are more than adequate, so something else is going on.
Damaged capacitors don't usually cause this type of behavior, but it's definitely worth a shot. I would definitely replace all capacitors of the same type.
What about the CPU, are you sure it's in good shape?

I'm wondering the same as bloodem, is the CPU itself in good shape.

If it were me, and all else seems to be failing, I'd pop the CPU apart and inspect it. Give it some new thermal compound while I was in there.

I also agree on the caps. Going from the 600 to 933 is about a 9W power jump? But like bloodem, I don't think that's the issue either, though anything is possible.

All the documentation on your board says it should work...

I bought the CPU on ebay, listed as working. Came without heatsink so I put thermal paste and mounted an heatsink. What I find baffling is that it works in a lot of contexts, including dos and safe mode, so I'm being cautious before saying it doesn't work.

lazalius wrote on 2025-03-25, 08:05:

Came without heatsink so I put thermal paste and mounted an heatsink.

That is already a red flag. Are you sure the heatsink is making a good contact with the die? The behavior you are describing can definitely be a result of overheating, I've seen it many times before.
DOS is a lot less intensive on the CPU. For example, if you have an unstable overclock, it tends to work in DOS just fine, but not in Windows.

bloodem wrote on 2025-03-25, 08:17:

lazalius wrote on 2025-03-25, 08:05:

Came without heatsink so I put thermal paste and mounted an heatsink.

That is already a red flag. Are you sure the heatsink is making a good contact with the die? The behavior you are describing can definitely be a result of overheating, I've seen it many times before.
DOS is a lot less intensive on the CPU. For example, if you have an unstable overclock, it tends to work in DOS just fine, but not in Windows.

It's a definite possibility. As an example...

I recently built an Athlon XP based Win98 system. During that process, the CPU was overheating and causing stability issues because the heatsink I was trying to use was inadequate, it was the only one I had on hand at the time, but it overheating the processor, causing lockups. Replaced it was a proper one rated for my CPU, the problem was solved.

It would freeze during install, at or near the same spot almost every time. Once I got Win98 installed, it froze on startup more often than not. Safe mode was indeed fine, as was boot to DOS prompt, as I was troubleshooting. I realized the issue when I physically felt the cooler and how warm it was. Right after a lockup I reset the system and hopped in the BIOS to see the temp, and it was much higher than it should have been.

What happens if you change the FSB setting to 100MHz or even 66MHz? Is it stable then? (I don't know if this motherboard lets you set that)
Alternatively, what happens if you put the 600EB back in? Is the system still working properly with that CPU?

If a cap is bulging at all then it's extremely out of spec and not doing much of anything anymore. If 1 cap is bulged there's probably others that just aren't showing physical signs yet. Unfortunately this could be causing the board to be in a borderline status, where it still runs but isn't entirely stable anymore. The faster CPU increases the demands on the motherboard.

A long time ago I had a "400W" PowMax that caused me stability problems on a 440BX Slot-1 machine. A reputable 250W PSU (which weighed twice as much) proved to be much better. I don't know if this is your problem but I wouldn't trust that PSU a bit. Do you have a multimeter? If so, check voltages (BIOS readings are unreliable).

When did you install Windows? Was it with this hardware configuration, or did you make changes since then? Win98 doesn't handle hardware changes very well. But swapping the CPU is one thing that shouldn't bother it at all.

Does this video card work in other systems? I've had a Geforce4 Ti4200 that locks up the system as soon as an AGP driver is loaded, but works fine if AGP driver is prevented from loading (performance is just compromised). I confirmed that on 2 working systems. It's one of the weirdest failures I've run into.

If you have any other cards installed, remove them to simplify the system until you can narrow down the issue.

Hello,
As far as the heatsink goes, it looks like it's making contact from what I can see, it is tightly locked with the metal clips.

I couldn't lower the FSB of the cpu (from bios the lowest i can set is 126MHz, probably requires a motherboard jumper set to go to 100 MHz), but can confirm the 600EB is working fine.

I reinstalled Windows with this configuration, and the video card works and plays games fine with the 600EB.

That cap is definetly bulging, I already ordered a replacement and will test again as soon as it's replaced. Do you have any advice on how to test the remaining capacitors? I'm reading I need an ESR multimeter to avoid desoldering every one of them.

Checking the PSU voltage is definetly worth a shot, I will keep you updated as soon as I can test.

Just trying to set your expectations: I don't think it's the capacitors, in fact I'm pretty sure they're not to blame.
Fun fact: I have a KT7A board on my desk right now, with at least 10 x bulging & leaking caps, and it has been running an Athlon XP-M CPU (overclocked to 2.2 GHz!) for the past two days with absolutely no sign of instability (if I find the time, I'll do a full recap today, though). In its current state, the board might sometimes need a few tries to POST, but once it does POST, it just works.

No, I think there is an actual problem with your new CPU. It would be good if you could get your hands on another known working board and see if the behavior changes.

lazalius wrote on 2025-03-26, 06:59:

That cap is definetly bulging, I already ordered a replacement and will test again as soon as it's replaced. Do you have any advice on how to test the remaining capacitors? I'm reading I need an ESR multimeter to avoid desoldering every one of them.

Most motherboard capacitors are in parallel with other caps, so an ESR meter will still show a good result if just one of the caps is still good. This really limits the usefulness IMO.
Unfortunately you can't really measure caps properly without removing them, and once you've done that you might as well replace them. But - if you remove them and you do have a meter handy, then it's still nice to get a measurement to satisfy curiosity.
I think the only way you might be able to rule out caps without desoldering is to use an oscilloscope and look at the ripple on various power circuits while the board is running. Then you can see if the power looks clean while it's under real working stress. But that's getting pretty spendy and involved unless you do this often.

Typically the most failure-prone caps are the ones on the output of CPU Vcore. But the cap you have visibly bulging looks like it would probably be on the 3.3V supply to RAM (and maybe the chipset). The other caps don't look like any of the quality brands so I wouldn't trust them, but that doesn't mean they've actually failed. A cap that's bulged is definitely bad but one that hasn't is just unknown.

Bad caps can cause stress that sometimes damages other components, such as MOSFETs and even the VRM controller IC. Catching them early is better for the board in the long run.
Bad cap(s?) may or may not be a primary cause of the immediate problem you're having, but if you keep the board then they will still need to be addressed if possible.
Bad CPUs are rare, but of course it can certainly happen. There's no telling what the CPU's history was.

I managed to test another PSU and the result is the same, and both PSU look good to a PSU tester. Thermal paste was replaced and heatsink is correctly installed with the metal clips; the heatsink itself comes from an 800MHz CPU and is almost twice the size of the 600MHz heatsink (it has a working fan as well). Since we're talking a vanilla windows 98 installation with no drivers installed, it is not a driver issue.
Everything seems to indicate a faulty CPU, but I don't have another working slot 1 board to confirm.