tehsiggi wrote on 2026-03-07, 16:19:

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

ChrisK wrote on 2026-03-08, 08:53:

tehsiggi wrote on 2026-03-07, 16:19:

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

Well, the 9800XXL is not a "real" XT - i'm interested in the thermal monitoring approach they've done on the XT.
While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not.
They use the LM63 in their design (apart from ASUS) and by the few pictures I found, they just placed a transistor as temperature sensor right next to the GPU. Still LM63, but a very different approach.

I was already able to confirm that R300, R350 and R360 have a thermistor (NTC) right next to the die on the top, so I'm just looking for good pictures of the "reference" 9800XT PCB to verify my claim: The R360 does NOT have a thermal sensing diode as the RV360 has. (And the temp measurements of the RV360 are way more accurate than R360)

Just wrapped up this "Q4/2001" Socket A build.

EpoX 8kha+
Athlon XP Palomino 1500+@1640MHz
Akasa Silver Mountain heatsink
Kingston HyperX 512MB@2-2-2-5-1T 1:1
120GB SSD + Win 2000 Pro
Asus V8200 Deluxe GeForce3 64MB
SB Audigy
3Com 10/100 NIC

I think the case is Lian Li pc-7a Plus II, but I'm not 100% sure. There are so many slightly different versions of this cases starting from the PC-60.

Few weeks ago I polymodded the motherboard and decided to make a period correct build out of it, after all 8kha+ was an absolute KJT266A OC legend back in the day. Pretty much everything connected to MB is very period correct except Kingston memory (KHX3000/512). It is in fact somehwhere around mid 2003 and should have BH-5:s. This made the OC process relatively easy, because I could just keep the memory 1:1 and at the tightest settings possible. For CPU I opted Palomino with smallest multiplier, again making the OC easier. I don't know if this build is a keeper, I have similar class sA systems already, but if I do, I probably throw water cooling stuff in. At this point I just didn't bother, it is just an unnecessary hassle if I decide to make something else in this.

Even the case fan with light is old school, it is not a modren led fan, but a cold cathode one 🤣.

Overall I had a productive retro computing week. I recapped one vintage 5V heavy PSU, one Radeon 9600 XT and while the gear for this build was on bench, I did a mid-late 2001 socket A cooler round-up, putting few of my vintage coolers to perfomance and sound test.

CharlieFoxtrot wrote on 2026-03-08, 14:29:

Just wrapped up this "Q4/2001" Socket A build.

Very nice looking system!

CharlieFoxtrot wrote on 2026-03-08, 14:29:

Kingston HyperX 512MB@2-2-2-5-1T 1:1

There was a time where I wished I had such RAM in my P4 system. Back in 2004 it was rumored that it can enable additional optimizations on 865PE ASUS motherboards (the famous ASUS PAT).

Now I look back and "meh, my Elixir-branded 3-3-3-8 is good enough". 🤣

I am testing out a somewhat-new-in-box Saitek Pro flightstick and two old Logitech USB gamepads that I got at the thrift store the other day. I definitely overpaid ($25 vs old price of $20) for the flightstick, but not in comparison to current prices for such things. The gamepads I got for $7 each.

I overclocked a s478 Prescott 3.2 to 3.8ghz on stock voltage. I could have went further but seeing the system pull close to 200w under a cpu load had me worried about the motherboard vrm, so I stopped there.

Also my Sapphire AGP x1950pro died. Or more specifically, I think the VRM failed, because now the card simply would not POST and the GPU stays cold.

dr_st wrote on 2026-03-08, 15:29:

Very nice looking system! […]
Show full quote

CharlieFoxtrot wrote on 2026-03-08, 14:29:

Just wrapped up this "Q4/2001" Socket A build.

Very nice looking system!

CharlieFoxtrot wrote on 2026-03-08, 14:29:

Kingston HyperX 512MB@2-2-2-5-1T 1:1

There was a time where I wished I had such RAM in my P4 system. Back in 2004 it was rumored that it can enable additional optimizations on 865PE ASUS motherboards (the famous ASUS PAT).

Now I look back and "meh, my Elixir-branded 3-3-3-8 is good enough". 🤣

Thanks!

Yeah, I don't think fast OC memory really is mandatory, but nice to have and worth getting if you bump into these sort of sticks cheap. I have bunch of different DDR1 stuff, HyperX, Team Group, Corsair etc., so I naturally use them over some run of the mill basic RAM when feasible. It also depends on the platform how much you benefit from tight timings. Anyways, it is still more for the love of the sport, usually couple of more frames here and there don't drastically change the experience 😀

In the past few days I repaired and tested, a 462 Gigabyte GA-7VAXP motherboard, as a test CPU I used an AMD with Duron 1000 (Morgan), and initially as a VGA I had an 8MB ATI Rage XL PCI, which I replaced with a Radeon 9200 AGP.
If it may interest you I ran some benchmarks, including those of Phil's, testing all the possible frequencies (standard VCORE), the minimum goes from 500 MHz and the maximum was 1333-1340 MHz, with three different FSB (100 133 166), I did not go further because an overvolt would have been necessary, but I saw the PC start even at 1400 MHz, a sign that it is possible to go to higher frequencies, probably this frequency requires a VCORE of 1.75V, slightly higher than the one used in all tests (1.65V), there is also the possibility of raising the FSB from the BIOS, but I do not think it can give many advantages, since over 1340 MHz it starts to accuse startup problems in the Doom and Quake benches, so the only proof that can be done is to switch to strange FSBs like 183 MHz (366), always staying within the limits of 1333 - 1340 MHz.

PC@LIVE wrote on 2026-03-09, 10:05:

In the past few days I repaired and tested, a 462 Gigabyte GA-7VAXP motherboard, as a test CPU I used an AMD with Duron 1000 (Morgan), and initially as a VGA I had an 8MB ATI Rage XL PCI, which I replaced with a Radeon 9200 AGP.
If it may interest you I ran some benchmarks, including those of Phil's, testing all the possible frequencies (standard VCORE), the minimum goes from 500 MHz and the maximum was 1333-1340 MHz, with three different FSB (100 133 166), I did not go further because an overvolt would have been necessary, but I saw the PC start even at 1400 MHz, a sign that it is possible to go to higher frequencies, probably this frequency requires a VCORE of 1.75V, slightly higher than the one used in all tests (1.65V), there is also the possibility of raising the FSB from the BIOS, but I do not think it can give many advantages, since over 1340 MHz it starts to accuse startup problems in the Doom and Quake benches, so the only proof that can be done is to switch to strange FSBs like 183 MHz (366), always staying within the limits of 1333 - 1340 MHz.

Interesting.
Nice work!

tehsiggi wrote on 2026-03-08, 11:21:

ChrisK wrote on 2026-03-08, 08:53:

tehsiggi wrote on 2026-03-07, 16:19:

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

Well, the 9800XXL is not a "real" XT - i'm interested in the thermal monitoring approach they've done on the XT.

Yes, for sure.
For the records, on the XXL they've used the 2N3904-transistor-as-thermal-diode method as described in the LM63 datasheet.
The transistor is placed right next to the R360 package.

tehsiggi wrote on 2026-03-08, 11:21:

While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not. They use th […]
Show full quote

ChrisK wrote on 2026-03-08, 08:53:

tehsiggi wrote on 2026-03-07, 16:19:

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not.
They use the LM63 in their design (apart from ASUS) and by the few pictures I found, they just placed a transistor as temperature sensor right next to the GPU. Still LM63, but a very different approach.

I was already able to confirm that R300, R350 and R360 have a thermistor (NTC) right next to the die on the top, so I'm just looking for good pictures of the "reference" 9800XT PCB to verify my claim: The R360 does NOT have a thermal sensing diode as the RV360 has. (And the temp measurements of the RV360 are way more accurate than R360)

You mean the NTC sitting on the package like those little Cs and Rs?

ChrisK wrote on 2026-03-09, 11:39:

Yes, for sure. For the records, on the XXL they've used the 2N3904-transistor-as-thermal-diode method as described in the LM63 d […]
Show full quote

tehsiggi wrote on 2026-03-08, 11:21:

ChrisK wrote on 2026-03-08, 08:53:

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

Well, the 9800XXL is not a "real" XT - i'm interested in the thermal monitoring approach they've done on the XT.

Yes, for sure.
For the records, on the XXL they've used the 2N3904-transistor-as-thermal-diode method as described in the LM63 datasheet.
The transistor is placed right next to the R360 package.

tehsiggi wrote on 2026-03-08, 11:21:

While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not. They use th […]
Show full quote

ChrisK wrote on 2026-03-08, 08:53:

If a 9800XXL (MS-8956) will be OK too I could also provide photos.

While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not.
They use the LM63 in their design (apart from ASUS) and by the few pictures I found, they just placed a transistor as temperature sensor right next to the GPU. Still LM63, but a very different approach.

I was already able to confirm that R300, R350 and R360 have a thermistor (NTC) right next to the die on the top, so I'm just looking for good pictures of the "reference" 9800XT PCB to verify my claim: The R360 does NOT have a thermal sensing diode as the RV360 has. (And the temp measurements of the RV360 are way more accurate than R360)

You mean the NTC sitting on the package like those little Cs and Rs?

The attachment RadeonXXL_R360_ThermalMonitoring.jpg is no longer available

Yup, exactly like that.
Thanks for the pictures.. that underlines my assumption. Interesting stuff. Again, I'd love to sit down in a round and talk about thinks like those with engineers involved. But I guess those interesting stories will be kept secret and will get lost with time.

And yup, that's the NTC right there. Present on all R300/R350/R360. It appears to be used on some fireGL cards in conjunction with a MIC502 for fan control.

PC@LIVE wrote on 2026-03-09, 10:05:

In the past few days I repaired and tested, a 462 Gigabyte GA-7VAXP motherboard, as a test CPU I used an AMD with Duron 1000 (Morgan), and initially as a VGA I had an 8MB ATI Rage XL PCI, which I replaced with a Radeon 9200 AGP.
If it may interest you I ran some benchmarks, including those of Phil's, testing all the possible frequencies (standard VCORE), the minimum goes from 500 MHz and the maximum was 1333-1340 MHz, with three different FSB (100 133 166), I did not go further because an overvolt would have been necessary, but I saw the PC start even at 1400 MHz, a sign that it is possible to go to higher frequencies, probably this frequency requires a VCORE of 1.75V, slightly higher than the one used in all tests (1.65V), there is also the possibility of raising the FSB from the BIOS, but I do not think it can give many advantages, since over 1340 MHz it starts to accuse startup problems in the Doom and Quake benches, so the only proof that can be done is to switch to strange FSBs like 183 MHz (366), always staying within the limits of 1333 - 1340 MHz.

I don't remember anyone pushing a Morgan much over 1500 back in the day. They were very closely related to Palomino in core design, but implemented with Aluminum rather than Copper (Like coppermine) which I think put limits on their thermal envelope and current leakage ceiling. So 1300s with the 1000 is doing pretty good. They should be performing like top end Thunderbirds at that speed, they've got a few percent over them clock for clock. I believe they could be mobile modded like a palamino and thus might go down to 300 then. Decent budget CPU overshadowed by the sexier stuff. I think AMD nerfed the sales of the faster ones though by letting the XP 1500 get discounted to the same range.

Nexxen wrote on 2026-03-09, 11:22:

PC@LIVE wrote on 2026-03-09, 10:05:

In the past few days I repaired and tested, a 462 Gigabyte GA-7VAXP motherboard, as a test CPU I used an AMD with Duron 1000 (Morgan), and initially as a VGA I had an 8MB ATI Rage XL PCI, which I replaced with a Radeon 9200 AGP.
If it may interest you I ran some benchmarks, including those of Phil's, testing all the possible frequencies (standard VCORE), the minimum goes from 500 MHz and the maximum was 1333-1340 MHz, with three different FSB (100 133 166), I did not go further because an overvolt would have been necessary, but I saw the PC start even at 1400 MHz, a sign that it is possible to go to higher frequencies, probably this frequency requires a VCORE of 1.75V, slightly higher than the one used in all tests (1.65V), there is also the possibility of raising the FSB from the BIOS, but I do not think it can give many advantages, since over 1340 MHz it starts to accuse startup problems in the Doom and Quake benches, so the only proof that can be done is to switch to strange FSBs like 183 MHz (366), always staying within the limits of 1333 - 1340 MHz.

Interesting.
Nice work!

Thanks a lot

BitWrangler wrote on 2026-03-09, 14:36:

PC@LIVE wrote on 2026-03-09, 10:05:

In the past few days I repaired and tested, a 462 Gigabyte GA-7VAXP motherboard, as a test CPU I used an AMD with Duron 1000 (Morgan), and initially as a VGA I had an 8MB ATI Rage XL PCI, which I replaced with a Radeon 9200 AGP.
If it may interest you I ran some benchmarks, including those of Phil's, testing all the possible frequencies (standard VCORE), the minimum goes from 500 MHz and the maximum was 1333-1340 MHz, with three different FSB (100 133 166), I did not go further because an overvolt would have been necessary, but I saw the PC start even at 1400 MHz, a sign that it is possible to go to higher frequencies, probably this frequency requires a VCORE of 1.75V, slightly higher than the one used in all tests (1.65V), there is also the possibility of raising the FSB from the BIOS, but I do not think it can give many advantages, since over 1340 MHz it starts to accuse startup problems in the Doom and Quake benches, so the only proof that can be done is to switch to strange FSBs like 183 MHz (366), always staying within the limits of 1333 - 1340 MHz.

I don't remember anyone pushing a Morgan much over 1500 back in the day. They were very closely related to Palomino in core design, but implemented with Aluminum rather than Copper (Like coppermine) which I think put limits on their thermal envelope and current leakage ceiling. So 1300s with the 1000 is doing pretty good. They should be performing like top end Thunderbirds at that speed, they've got a few percent over them clock for clock. I believe they could be mobile modded like a palamino and thus might go down to 300 then. Decent budget CPU overshadowed by the sexier stuff. I think AMD nerfed the sales of the faster ones though by letting the XP 1500 get discounted to the same range.

Thanks a lot
I don't rule out trying a faster FSB in the future, like at 190 MHz (380), unfortunately I tried unsuccessfully the selection for 200 MHz (400), and it doesn't work unfortunately 😣

My journey of becoming friends with Linux (Fedora KDE is my choice of poison) now had another milestone - almost a year in, i found a tool called "rclone", that can do cloud storage mounting with limited cache sizes and on-demand delivery. It can do it even with SMB mounts (or even over FTP), with all the control over those same caching settings as well as doing it in "user-level" instead of "root-level". (no need to modify /etc/fstab)
It also solved all my issues with computer sometimes being unable to go to sleep - it was all fault of those OneDrive mounting programs. Going to sleep now always works and is near instantaneous.

Minor thing for those who have maybe always known it. But i struggled with multiple OneDrive tools, none working well, until today.

I now have all my "remote storage" (OneDrive, Google Drive and SMB shares from my NAS) working in my setup with such performance that it gives a perception of having all the data locally. Yay!

EDIT: Wait, that should have gone to the "Modern activity" thread, not to here... Ups.

tehsiggi wrote on 2026-03-09, 14:16:

Yup, exactly like that. Thanks for the pictures.. that underlines my assumption. Interesting stuff. Again, I'd love to sit down […]
Show full quote

ChrisK wrote on 2026-03-09, 11:39:

Yes, for sure. For the records, on the XXL they've used the 2N3904-transistor-as-thermal-diode method as described in the LM63 d […]
Show full quote

tehsiggi wrote on 2026-03-08, 11:21:

Well, the 9800XXL is not a "real" XT - i'm interested in the thermal monitoring approach they've done on the XT.

Yes, for sure.
For the records, on the XXL they've used the 2N3904-transistor-as-thermal-diode method as described in the LM63 datasheet.
The transistor is placed right next to the R360 package.

tehsiggi wrote on 2026-03-08, 11:21:

While the 9600XT (RV360) has a thermal sensing diode inside the die of the GPU, I assume the 9800XT (R360) does not.
They use the LM63 in their design (apart from ASUS) and by the few pictures I found, they just placed a transistor as temperature sensor right next to the GPU. Still LM63, but a very different approach.

I was already able to confirm that R300, R350 and R360 have a thermistor (NTC) right next to the die on the top, so I'm just looking for good pictures of the "reference" 9800XT PCB to verify my claim: The R360 does NOT have a thermal sensing diode as the RV360 has. (And the temp measurements of the RV360 are way more accurate than R360)

You mean the NTC sitting on the package like those little Cs and Rs?

The attachment RadeonXXL_R360_ThermalMonitoring.jpg is no longer available

Yup, exactly like that.
Thanks for the pictures.. that underlines my assumption. Interesting stuff. Again, I'd love to sit down in a round and talk about thinks like those with engineers involved. But I guess those interesting stories will be kept secret and will get lost with time.

And yup, that's the NTC right there. Present on all R300/R350/R360. It appears to be used on some fireGL cards in conjunction with a MIC502 for fan control.

I'll post the 9800XT photos tomorrow as well, just in case mine may have it. I didn't get to see closeups of the core on my card and it appears to be some OEM version as far as it looks to me. (it has a dual slot type of VGA bracket, but single slot cooler)

After testing out my Voodoo 2, and realizing that I don't really have a use for it thanks to my Voodoo 3 300, I've decided to swap out the build in my ATX case once more, this time with the intent being a complete overkill 1997 build, albeit with a few QoL differences. The idea I have is as follows:

• Using my single Socket 8 board as the base with one of my PPro at 200(233 is possible I think)
• 128MB of EDO, as I have that available and currently not in use
• Matrox Millennium 1 with my Voodoo 1 tied in to it
• My MPact! DVD card for, uh, it's 1997. Everyone wants a DVD card. totally.
• my AWE64 Gold with the 8MB expansion board that is otherwise sitting around.

My only point of contention is if I should try to find a 32 bit SCSI card or just use my slightly anachronistic 64bit PCI card to connect my working 40GB SCSI hard drive, or if I should just throw in a 66MB/s IDE paired with some random IDE drive and call it a day. I would be using a SD-IDE converter as a secondary drive for the sake of quick file transfers.

Beyond that, PCI or ISA for networking purposes is the only other thing I can think of.

This was a cheap score on ebay a few weeks ago, I finally got around to puttering with it. Tyan S1564D dual socket7 board, came with a pair of 233MMX processors.

Seller said it wouldn't POST....and initially, it didn't for me either....so time to start troubleshooting.

This was the first red flag....but all contacts were good (passing voltage).

First check was vcore, it was @ 3.3v! 🤦 ...but CPU's were ice cold. So immediate power down and tested the CPU's in a known good board, they were still good! 😌

Ok, that was my flub, as any board that has jumpers I normally always check and make sure they're right; this has been the saving grace for many-a-boards!!! This time I didn't. Well, good thing I checked, every farking jumper was wrong for CPU settings; voltages, FSB speeds, and number of processors. So I set them according to the manual and shazam!!

VCORE dead nuts on!

It was on my service bench, so when I fired it it began beeping 'no video' at me and I knew it was alive at that point.

So now over to the 'oven'. You'll have to excuse the crude heatsinks....they're just resting on the processors for testing.

Of course the RTC is toast....so of course it won't hold settings. I'll order one for it.... Sorry, I just won't do the hackjob RTC fixes.....I find it repulsive resto work; like fixing a rust hole in a car by filling it full of bondo. RTC chips are plentiful and cheap.

No build in mind for it.....but I'll come up with one! I love stupid people!!

tehsiggi wrote on 2026-03-07, 16:19:

If everything works out, this will work with either R300, R350 or R360. Next week will give some progress on this. […]
Show full quote

ChrisK wrote on 2026-03-07, 15:57:

Following the other thread I'd guess this has something to do with measureing the temperature sensor on those cards.

If everything works out, this will work with either R300, R350 or R360. Next week will give some progress on this.

Shponglefan wrote on 2026-03-07, 16:07:

Diagnosed a faulty Diamond Viper V550 TNT AGP graphics card. […]
Show full quote

Diagnosed a faulty Diamond Viper V550 TNT AGP graphics card.

The attachment Diamond Viper V550 AGP no video.jpg is no longer available

The card is successfully detected by the PC, but fails to display any video output.

After some physical inspection I discovered it's missing an inductor (L16) and several capacitors (C255, C256, C257). There was also some abrasion on nearby resistors, so I'm assuming this card physically hit something to knock these components off.

The attachment Diamond Viper V550 AGP missing inductor.jpg is no longer available

The attachment Diamond Viper V550 AGP missing capacitors.jpg is no longer available

Not sure if this is the reason it won't output video, but it's a starting point. I plan to remove these components from a working card so I can measure them and then order some replacements.

For testing, you can bridge L16 with a short piece of wire or a 0 Ohms resistor. It is a filter inductor for incoming power, without it in place, the GPU is likely missing a voltage. That can definitely explain why the card isn't showing an image.

---

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

Thats unfortunate. I scrapped a Dell 9800XT (bad artifacting @ bios) a few months ago when I got rid of the massive box of scrap electronics because the box was falling apart.

TheAbandonwareGuy wrote on Yesterday, 00:11:

tehsiggi wrote on 2026-03-07, 16:19:

If everything works out, this will work with either R300, R350 or R360. Next week will give some progress on this. […]
Show full quote

ChrisK wrote on 2026-03-07, 15:57:

Following the other thread I'd guess this has something to do with measureing the temperature sensor on those cards.

If everything works out, this will work with either R300, R350 or R360. Next week will give some progress on this.

Shponglefan wrote on 2026-03-07, 16:07:

Diagnosed a faulty Diamond Viper V550 TNT AGP graphics card. […]
Show full quote

Diagnosed a faulty Diamond Viper V550 TNT AGP graphics card.

The attachment Diamond Viper V550 AGP no video.jpg is no longer available

The card is successfully detected by the PC, but fails to display any video output.

After some physical inspection I discovered it's missing an inductor (L16) and several capacitors (C255, C256, C257). There was also some abrasion on nearby resistors, so I'm assuming this card physically hit something to knock these components off.

The attachment Diamond Viper V550 AGP missing inductor.jpg is no longer available

The attachment Diamond Viper V550 AGP missing capacitors.jpg is no longer available

Not sure if this is the reason it won't output video, but it's a starting point. I plan to remove these components from a working card so I can measure them and then order some replacements.

For testing, you can bridge L16 with a short piece of wire or a 0 Ohms resistor. It is a filter inductor for incoming power, without it in place, the GPU is likely missing a voltage. That can definitely explain why the card isn't showing an image.

---

BTW: If anybody has a legit DEAD 9800XT - give me a message.. by dead I mean really dead. I need the PCB for science. They're just hard to get.. OR a high res photo of the GPU area without cooler. Much appreciated.

Thats unfortunate. I scrapped a Dell 9800XT (bad artifacting @ bios) a few months ago when I got rid of the massive box of scrap electronics because the box was falling apart.

Huh, that explains the long bracket the one I'm getting has. It's a DELL OEM card, according to what Google can tell me.

Well, after my P3 Voodoo3 build started to throw RAM errors AGAIN, even after already replacing the DIMMs, it was probably time to retire it. That QDI Advance 10T served me well for a full decade, but failing DIMMs so soon after the previous RAM failure is VERY suspicious... I only had these in that board for 2 weeks...
But what could replace it anyway? The reason I had that system was to have the best possible compatibility with 90's games, while also having good performance. Not exactly period-correct, since I never really go for that, although its original specs were very much period-correct, with a Soyo SY-6VBA133, P2 400, SBLive! and Voodoo3, it quickly evolved to a PGA370 P3 (mostly because the original board died, and I ended up upgrading to the QDI), only keeping the Voodoo and SBLive! constantly. It even was a P4 build briefly, before settling on that QDI board (I really dislike Netburst, and for some reason that particular config caused problems with some games anyway.) That QDI in of itself went from P3 650MHz to Tualeron 1.2GHz, to P3-S 1266MHz OCed @ 1.3GHz. Surprisingly, the same Win98SE install took all those changes (even the motherboards) like a champ over the years without a hitch. I wouldn't really mind reinstalling the OS, it just didn't happen to be necessary.
So, what is the next step? Well, I have an MSI K7T266 Pro2 V2.0 (with the extra Promise FastTrak100 and NEC USB2 builtin), a few socket 462 CPUs, plenty of DDR DIMMs (although just one 512MB was more than enough while also being perfectly stable without patches), and more than enough time to rebuild everything. I settled on an Athlon XP 1800+ Palomino (~1.56GHz), 512MB DDR400 (allows for very tight timings on 266 speeds), and basically kept everything else the same. Build went nice, almost too nice...
Just for shits and giggles, I hooked up that same Win98SE HDD to see if that will work. And well... IT ACTUALLY DID! 🤯 No BSOD, no conflicts, no falling back to DOS Compatibility Mode, just automatically reinstalling the proper drivers for the new configuration, and after that, happily playing the startup sound, and loading my desktop. One reboot just for good measure, and everything is still fine. Now,I'm no stranger to getting Windows XP and later installs working on different motherboards even when SATA controllers were different (injecting drivers on existing installations is especially easy on Vista and later, and yes, I also did it just for fun), but that was from G41 chipset to Z77, and it still needed some finagling with drivers and reboots, even though when I was done, it was just fine. This? I expected it to be considerably worse, since everyone here knows how temperamental Win9x can be, but it was even smoother instead. It probably helped that they were similar enough VIA chipsets, and used the same drivers.
I also have another HDD on that PC with WinXP, mostly to see if I can get working Glide support in that, but the drivers have failed me there again and again. (Any recommendations for XP Voodoo3 drivers??). I also tried that, but ironically, that failed spectacularly with an UNMOUNTABLE_BOOT_VOLUME. Not good, and that's not due to the motherboard being different, it might be an actual HDD problem.