VOGONS


First post, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie

Hello VOGONS,

Today, I have a truly tragic tale to share, and need some guidance afterwards.

Recently, I decided to try to do some FSB OCing with my 1.4Ghz PIII-S Tualatin on my newly-recapped Asus TUSL2-C to try to squeeze out some extra performance. I did a mild OC of 5MHz on the FSB, which also set the SDRAM frequency to 138Mhz and the PCI freq. to 34Mhz. I had finished stress-testing with an older version of Prime95 for Windows 9x, and went to play some NFS3 when this happened:
FhMgQYhl.jpg

That's right. It seems as though that seemingly light OC somehow damaged my V5. I quickly shut off my PC, reverted my OC to stock, tried again, same result. Let's just say I wasn't too happy at the proposition of damaging my collector's item.

About a week has passed, and I tried again on my Asus P3V4X test bench with Windows ME, and lo and behold, nothing changed:
HcNkAnrl.jpg
I've even tried reflashing the VBIOS on the card, thinking perhaps the VBIOS was corrupted somehow. No die.

To summarize the issue, in 2D-mode, the card works normally, but whenever I try to throw any 3D load at it, it freaks out, and I have to shut down the machine it's running on.

If someone could tell me what exactly I have done to my poor V5, please let me know. I doubt it's a cap issue - I checked the caps earlier, and they aren't that far off from their rated capacitance. I'm willing to recap the card, though. RN I'm thinking it's either a damaged GPU die or memory chip issue, but I have no clue. If that's the case... I'm kinda SOL unless someone with way better soldering skills and equipment would like to have a crack at my card for free.

Any help would be appreciated.

P.S. : I'll never, ever do FSB OCing again with a 3dfx card in the system unless I can lock the FSB freq. in the BIOS. Still shocked, though, that such a minor FSB increase actually did harm to my GPU in such a "permanent" way. Every other device in my Tually rig works perfectly fine still.

Last edited by ultra_code on 2020-11-06, 08:01. Edited 1 time in total.

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 2 of 16, by ODwilly

User metadata
Rank l33t
Rank
l33t

There are some threads around here with people replacing memory chips that were damaged, which that looks like it could be

Main pc: Asus ROG 17. R9 5900HX, RTX 3070m, 16gb ddr4 3200, 1tb NVME.
Retro PC: Soyo P4S Dragon, 3gb ddr 266, 120gb Maxtor, Geforce Fx 5950 Ultra, SB Live! 5.1

Reply 3 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
ODwilly wrote on 2020-11-06, 06:26:

There are some threads around here with people replacing memory chips that were damaged, which that looks like it could be

Any idea how exactly it could happen? Also, have any links perhaps to those threads?

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 5 of 16, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

Why do you think the AGP slot was running at 138Mhz? AGP normally runs at 66Mhz.
https://en.wikipedia.org/wiki/Accelerated_Graphics_Port

At most, overclocking the fsb by 5Mhz would have overclocked the AGP slot to 68.5Mhz which is no big deal at all and would not be the cause of the issue with the V5.

My guess is that it was just a sad coincidence that the V5 messed up at the same time.

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK

Reply 6 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
cyclone3d wrote on 2020-11-06, 07:26:
Why do you think the AGP slot was running at 138Mhz? AGP normally runs at 66Mhz. https://en.wikipedia.org/wiki/Accelerated_Graph […]
Show full quote

Why do you think the AGP slot was running at 138Mhz? AGP normally runs at 66Mhz.
https://en.wikipedia.org/wiki/Accelerated_Graphics_Port

At most, overclocking the fsb by 5Mhz would have overclocked the AGP slot to 68.5Mhz which is no big deal at all and would not be the cause of the issue with the V5.

My guess is that it was just a sad coincidence that the V5 messed up at the same time.

My b. I was thinking of SDRAM clock, not AGP clock.

I agree that such a tiny bump in FSB affected it when it shouldn't have it is unfortunate. Regardless, I want to try to fix this card, or give it to someone who can.

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 7 of 16, by ODwilly

User metadata
Rank l33t
Rank
l33t

Voodoo 5 5500 Repair (project thread) heres the 1st one I came across. I think there are a few more, I juzt googled Vogons and Voodoo 5 repair 🤣.

Main pc: Asus ROG 17. R9 5900HX, RTX 3070m, 16gb ddr4 3200, 1tb NVME.
Retro PC: Soyo P4S Dragon, 3gb ddr 266, 120gb Maxtor, Geforce Fx 5950 Ultra, SB Live! 5.1

Reply 8 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
ODwilly wrote on 2020-11-06, 08:17:

Voodoo 5 5500 Repair (project thread) heres the 1st one I came across. I think there are a few more, I juzt googled Vogons and Voodoo 5 repair 🤣.

Thanks.

Hmm... I don't think it has anything to do with my GPU dies getting toasted. In fact, recently, I de-heatsinked my V5 and replaced them with brand-new push-pin heatsinks and Arctic MX-4 thermal paste.
jmA7sEHl.jpg

On top of that, in that guy's case, even in 2D mode things were going awry.

Also, if anyone's interested, here were the capacitance values I measured (granted, for the smaller caps, I only measured one; measuring all of them would probs be in my best interest):

  • Small 10uf caps - 8uf
  • Larger 100uf cap - 98uf
  • Largest 470uf cap - 437uf

They are aging, that's for sure. I'll recap the card shortly, but am doubtful it'll solve the issue.

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 10 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
chrismeyer6 wrote on 2020-11-07, 00:01:

Random thought have you checked into make sure it's getting proper voltage on the auxiliary power connector?

Not exactly sure how I would test that, nor what the proper voltage it would need to be receiving.

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 12 of 16, by RockstarRunner

User metadata
Rank Member
Rank
Member

I just had something very similar happen to my V3 3000, but nothing was overclocked, it just started behaving this way one time. I made a thread about it, might be a bad vram chip.
Does yours still look so bad with non-glide software? Such as 3dmark? Glide hard locks my machine, but other 3d modes work, but with artifacts.

Reply 13 of 16, by Hanamichi

User metadata
Rank Member
Rank
Member

Did you up the voltages on other components for the overclock?

I wonder if a power supply rail that was wobbly went out of spec during the overclock.

Hopefully you get it back up and running though.

Reply 14 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
RockstarRunner wrote on 2020-11-08, 09:01:

I just had something very similar happen to my V3 3000, but nothing was overclocked, it just started behaving this way one time. I made a thread about it, might be a bad vram chip.
Does yours still look so bad with non-glide software? Such as 3dmark? Glide hard locks my machine, but other 3d modes work, but with artifacts.

NFS3, 3DMark, if it's 3D, it Black-Screen-of-Deaths.

Hanamichi wrote on 2020-11-08, 16:20:

Did you up the voltages on other components for the overclock?

I wonder if a power supply rail that was wobbly went out of spec during the overclock.

Hopefully you get it back up and running though.

Nope. Literally had my CPU at the lowest voltages available in the BIOS, and the V5 wasn't OCed.

Builds
ttgwnt-6.png
kcxlg9-6.png

Reply 15 of 16, by Hanamichi

User metadata
Rank Member
Rank
Member

Weird coincidence then maybe!
Might be worth checking the PSU just in case it's a component killer.

The only other thing I can think of is that you mentioned taking off the heatsinks. Did you have to use a lot of force? It could have damaged a solder joint.

Reply 16 of 16, by ultra_code

User metadata
Rank Oldbie
Rank
Oldbie
Hanamichi wrote on 2020-11-09, 02:14:

Weird coincidence then maybe!
Might be worth checking the PSU just in case it's a component killer.

The only other thing I can think of is that you mentioned taking off the heatsinks. Did you have to use a lot of force? It could have damaged a solder joint.

The PSU is like a 2-3 yo Corsair CX450M PSU. Shouldn't be a problem.

As for the heatsinks, I used the freeze-spray/flat-head screwdriver method. Required a bit more force than my V4 and Banshee needed IIRC, but the heatsinks did just pop off. Again, very doubtful that was the cause.

Builds
ttgwnt-6.png
kcxlg9-6.png