I am not disappointed in the overall result.. Got a cheap proper mobo for my testing.

I ordered the right polyfuse and will put it back in, even though I'll never use the firewire ports anyway.

The PCI slot may stay or be replaced.. I'll just order some for good measure and then see how i feel.

Well I have some news on my dead Voodoo3 PCI card, seems its problem is a burnt out finger on the PCI connector, not 100% sure what this finger is for but I assume its part of the power delivery for the card. The card did get warm when I initially went to test it and its totally undetected by the system.

Doesn't bode well for the card itself and I'm not sure how to repair this kind of damage in a way that's more permanent or what caused the finger to be burnt off. Have done a little scratching and the finger is totally gone, there isnt anything to solder to rebuild a finger.

If you have a spare finger to stick on there, you can scrape the via leading to it and solder the new finger to that via.

Although you'll have to secure the finger down and that... idk how you'd go around doing it.

Trashbytes wrote on Yesterday, 20:38:

Well I have some news on my dead Voodoo3 PCI card, seems its problem is a burnt out finger on the PCI connector, not 100% sure what this finger is for but I assume its part of the power delivery for the card. The card did get warm when I initially went to test it and its totally undetected by the system.

The attachment Burnt Finger.jpg is no longer available

Doesn't bode well for the card itself and I'm not sure how to repair this kind of damage in a way that's more permanent or what caused the finger to be burnt off. Have done a little scratching and the finger is totally gone, there isnt anything to solder to rebuild a finger.

I've done a repair of a missing pad before, took a PCI modem and cut off the pci pad with a sharp blade. Then superglued the pad into place and soldered a wire to it: Re: What retro activity did you get up to today?

Thermalwrong wrote on 2023-12-12, 23:45:

I think it's finally time for me to put away the hot air station - I've now broken the Elsa Victory II enough that one of the memory data lines is just broken. I mean I was fixing corrosion but with hindsight I can say that drilling through the PCB to run new traces through was a terrible idea. I didn't notice until after I'd drilled those holes that the PCB is 6-layer and some of the data traces route internally, or they did until I started on it. Maybe I'll fix that in the future but it's going away for a while now - repairing a Banshee card is tricky business since there isn't the plethora of information like there is for the Voodoo 1 and Voodoo 3 cards.

The attachment IMG_2750 (Custom).JPG is no longer available

I got one card to actually post once after reflowing it so thought I'd have a go at reflowing or removing the GPU of this PCI Voodoo Banshee - I had thought its GPU was fried since there were missing pads and traces on the PCI slot which I repaired but it would never run, just gave the no video card detected beeps.

That card actually works now since its main fault was a BGA connection issue.

I wonder why it seems like only 3dfx voodoo PCI cards can have burned out pci pins? I suppose it's a survivorship bias since they're the only broken old PCI cards worth enough to not scrap, but I suspect maybe they got fitted into an incorrect port type at some point for this to happen.

Good news for you though, that is just a ground pin (A18 by my count) and there are lots of those on the PCI connector, all you really need to do is clean up any burnt PCB to ensure it's not conductive to the adjacent pads.

Thermalwrong wrote on Today, 01:15:

I've done a repair of a missing pad before, took a PCI modem and cut off the pci pad with a sharp blade. Then superglued the pad […]
Show full quote

Trashbytes wrote on Yesterday, 20:38:

Well I have some news on my dead Voodoo3 PCI card, seems its problem is a burnt out finger on the PCI connector, not 100% sure what this finger is for but I assume its part of the power delivery for the card. The card did get warm when I initially went to test it and its totally undetected by the system.

The attachment Burnt Finger.jpg is no longer available

Doesn't bode well for the card itself and I'm not sure how to repair this kind of damage in a way that's more permanent or what caused the finger to be burnt off. Have done a little scratching and the finger is totally gone, there isnt anything to solder to rebuild a finger.

I've done a repair of a missing pad before, took a PCI modem and cut off the pci pad with a sharp blade. Then superglued the pad into place and soldered a wire to it: Re: What retro activity did you get up to today?

Thermalwrong wrote on 2023-12-12, 23:45:

I think it's finally time for me to put away the hot air station - I've now broken the Elsa Victory II enough that one of the memory data lines is just broken. I mean I was fixing corrosion but with hindsight I can say that drilling through the PCB to run new traces through was a terrible idea. I didn't notice until after I'd drilled those holes that the PCB is 6-layer and some of the data traces route internally, or they did until I started on it. Maybe I'll fix that in the future but it's going away for a while now - repairing a Banshee card is tricky business since there isn't the plethora of information like there is for the Voodoo 1 and Voodoo 3 cards.

The attachment IMG_2750 (Custom).JPG is no longer available

I got one card to actually post once after reflowing it so thought I'd have a go at reflowing or removing the GPU of this PCI Voodoo Banshee - I had thought its GPU was fried since there were missing pads and traces on the PCI slot which I repaired but it would never run, just gave the no video card detected beeps.

That card actually works now since its main fault was a BGA connection issue.

I wonder why it seems like only 3dfx voodoo PCI cards can have burned out pci pins? I suppose it's a survivorship bias since they're the only broken old PCI cards worth enough to not scrap, but I suspect maybe they got fitted into an incorrect port type at some point for this to happen.

Good news for you though, that is just a ground pin (A18 by my count) and there are lots of those on the PCI connector, all you really need to do is clean up any burnt PCB to ensure it's not conductive to the adjacent pads.

Yeah I've been pondering a good way to repair the damage that will last, your idea will work ok if I put it in a system and just leave it there. More concerning is what caused the initial damage, that pin doesn't seem to connect to either of the voltage regulators on the card but I might be missing something and it does actually find its way to the regulator. Since I don't have a ton of experience at fixing 3dfx cards this repair will have some ...winging it involved but Ill be checking the mosfets and caps before it gets put back in a board.

Edit - did more investigation and this burnt pin connects to the via above it which connects directly to the third pin on the opposite side of the edge connector via that via...it goes nowhere else, so perhaps these two pins are shorted together on purpose and do nothing.

I need to find out what the fingers do on the PCI slot. (Just checked and both are unused by the PCI bus so that explains why they are shorted together)

tehsiggi wrote on Yesterday, 18:30:

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken. […]
Show full quote

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken.

I was on the lookout for a good 939 AGP board and nForce3 was really nice to have.

So today I took care of it and got it working. Turns out we have an exploded polyfuse in the Firewire section - that's what I call irony!

I removed the burned part and cleaned everything. The PCI slot will need replacement if it is ought to look like new again at some point. But so far - the board works just fine and will be my primary testing platform from now on.

Here are some pictures of FireWire with burnout!

Those are sweet MB's!
Mine is now rocking an x2 4400+ under a Wraith Prism cooler, keeps it barely over room temp =)

Trashbytes wrote on Today, 01:40:

Edit - did more investigation and this burnt pin connects to the via above it which connects directly to the third pin on the opposite side of the edge connector via that via...it goes nowhere else, so perhaps these two pins are shorted together on purpose and do nothing.

I need to find out what the fingers do on the PCI slot. (Just checked and both are unused by the PCI bus so that explains why they are shorted together)

I'm wondering. I the orientation is correct, the missing pin is A18 and therefore GND. If it is going to the third pin from the edge on the opposite side of the connector, it would short with the clock signal. Not sure if that's supposed to be that way 😉

bofh.fromhell wrote on Today, 03:09:

tehsiggi wrote on Yesterday, 18:30:

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken. […]
Show full quote

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken.

I was on the lookout for a good 939 AGP board and nForce3 was really nice to have.

So today I took care of it and got it working. Turns out we have an exploded polyfuse in the Firewire section - that's what I call irony!

I removed the burned part and cleaned everything. The PCI slot will need replacement if it is ought to look like new again at some point. But so far - the board works just fine and will be my primary testing platform from now on.

Here are some pictures of FireWire with burnout!

Those are sweet MB's!
Mine is now rocking an x2 4400+ under a Wraith Prism cooler, keeps it barely over room temp =)

I moved over my X2 3800 and 2GB RAM onto that board now. Together with a Samsung EVO 840 I'm able to leverage that sweet onboard SATA. My old platform (MS-7075) had a K8T800 with the SATA bug, meaning it only detects SATA I drives and nothing else. This now works so much better now. Back in the day I would have OC'ed the snot out of it with such a nice board. But for today I'll leave it humming at default and be happy with it.

tehsiggi wrote on Today, 05:02:

I'm wondering. I the orientation is correct, the missing pin is A18 and therefore GND. If it is going to the third pin from the […]
Show full quote

Trashbytes wrote on Today, 01:40:

Edit - did more investigation and this burnt pin connects to the via above it which connects directly to the third pin on the opposite side of the edge connector via that via...it goes nowhere else, so perhaps these two pins are shorted together on purpose and do nothing.

I need to find out what the fingers do on the PCI slot. (Just checked and both are unused by the PCI bus so that explains why they are shorted together)

I'm wondering. I the orientation is correct, the missing pin is A18 and therefore GND. If it is going to the third pin from the edge on the opposite side of the connector, it would short with the clock signal. Not sure if that's supposed to be that way 😉

bofh.fromhell wrote on Today, 03:09:

tehsiggi wrote on Yesterday, 18:30:

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken. […]
Show full quote

I recently snacked a EPOX 9NDA3+ from ebay. It was listed as broken.

I was on the lookout for a good 939 AGP board and nForce3 was really nice to have.

So today I took care of it and got it working. Turns out we have an exploded polyfuse in the Firewire section - that's what I call irony!

I removed the burned part and cleaned everything. The PCI slot will need replacement if it is ought to look like new again at some point. But so far - the board works just fine and will be my primary testing platform from now on.

Here are some pictures of FireWire with burnout!

Those are sweet MB's!
Mine is now rocking an x2 4400+ under a Wraith Prism cooler, keeps it barely over room temp =)

I moved over my X2 3800 and 2GB RAM onto that board now. Together with a Samsung EVO 840 I'm able to leverage that sweet onboard SATA. My old platform (MS-7075) had a K8T800 with the SATA bug, meaning it only detects SATA I drives and nothing else. This now works so much better now. Back in the day I would have OC'ed the snot out of it with such a nice board. But for today I'll leave it humming at default and be happy with it.

Yes my bad its pin B17 on the front and A18 on the back counting from the face plate, both are listed as Unconnected/ground in the docs I have here.

A18 is the missing one and its connected to B17 via the via.

https://wiki.attie.co.uk/wiki/PCI_pinout

is what I used to determine pin out.

I think it's the other way around. Backside pins are named A, front side are named B (don't get me started 😁).

But yeah, I assume they'll both be ground, so probably not your main issue.

tehsiggi wrote on Today, 05:13:

I think it's the other way around. Backside pins are named A, front side are named B (don't get me started 😁).

But yeah, I assume they'll both be ground, so probably not your main issue.

Yup .. fixed that ...

But yes its not the issue ..the card isn't detected on the PCI bus, no tool I have can find it so there is a power issue there as the card isn't powering up, even with a dead BIOS the card should still be detected on the PCi bus.

So my next step will be to poke at the mosfet and regulator, my card is identical to the image below.

The regulator on the heat sink will be the first to get the poke followed by the fet down near the edge connector, IIRC its normally the regulator that dies on these cards. (I think its a fet on the heatsink and the regulator is the one near the edge connector, shrug they both get the poke)

Not quite sure atm but I think both are linear regulators. The one with the heatsink should make 3.3V for I/O, memory, etc. and the one near the slot is for V_core of the Voodoo chip, so something around 2.xx volts iirc. Look for datasheets/pinouts of these to better troubleshoot them.

ChrisK wrote on Today, 06:37:

Not quite sure atm but I think both are linear regulators. The one with the heatsink should make 3.3V for I/O, memory, etc. and the one near the slot is for V_core of the Voodoo chip, so something around 2.xx volts iirc. Look for datasheets/pinouts of these to better troubleshoot them.

Fingers crossed the one near the core has not failed and sent too much voltage to the core, the card did get fairly warm when I was testing it.