Did you protect / cover the southbridge chip from solder being blown underneath? The southbridge is right beside PCI slot 2!
Otherwise there might be shorts between the BGA solder balls now.

Fail! I didn't cover the southbridge. I didn't realize there was a noticeable gap. But the ball pitch doesn't seem catastrophically close. I tried sticking some stranded wire underneath and scouring. Didn't help.
Also, it seems the CLK LED on the POST card means whether there's a bus clock, not CPU. Then why would there be no PCI clock if only the south bridge is broken. PCI is from the northbridge and the southbridge is only a PCI device.

One thing I noticed was there was no solder splatter on the 2 nearby PCI slot pins. There definitely was on the plastic part. So that suggests the spatter is unlikely to go per underneath something.

I'm confused why you used the iron from above the board.

The logical way it to flip the board and do everything from the bottom.

But there's a tell tale melt mark on the slot.

I'd pull the slot back out, clean and inspect, and test without.

I didn't use a soldering iron on the front side of the board. How's that possible. I did blow the solder out from both the front and back side.

Those are rotary tool marks, not melt marks.

Actually, I did test the board after removing the PCI slot and cleaning, but it didn't POST. I thought maybe some power lines to other important components would be disconnected if the PCI slot wasn't soldered, so I went ahead and soldered it. Maybe that was a mistake and should've not proceeded until it was working.

Rotary tool?
What were you doing with that?

There's visible spray from something. Plastic or solder is hard to tell.

https://vinafix.com/threads/schematics-ms-61xx.20379/

You might be able to find it elsewhere,but basically, check voltages.
If you didn't preheat the board, it's possible that it warped and disconnected a ball under the chipset.

Personally, id run it through the dishwasher without soap to clean it properly before testing again.

In the future, I'd recommend low melt solder from below the board. The slot should basically fall out, and then you can solder wick the holes clean easily. Then it's a simple matter of new slot, and solder the pins from below.

Rotary tool?

I said in my 1st post, I had to cut the PCI slot into pieces, so it was easier to pull out. Otherwise, I don't think the heat gun can heat such a large area, without a special duct. I was very careful to not cut into the PCB, obviously. Too bad I didn't take a picture after removing the slot. Also, the chipset is pretty far from the PCI slot, so I don't expect the board in that area to warp.

You might be able to find it elsewhere,but basically, check voltages.

Good idea, I actually have another 1 of these boards, except version 1.0. I've seen repair shop videos before where the 1st thing they probe are the power lines. There doesn't seem to be many places to test around the southbridge, except for those exposed parts of the traces (without solder mask).

Personally, id run it through the dishwasher without soap to clean it properly before testing again.

Uh, water and chips seemed like a bad idea from all the moisture indicators I've seen in DigiKey packages. But I didn't know until now that it's only a problem when you solder them. Actually, I could try a pressure washer, maybe at 800 PSI.

I also did try solder wick before using the air compressor, but it was only pulling the solder off the surface, not the thru hole. Maybe I needed to slide the wick around.

Jesus christ.
You're a man of extremes aren't you.
I suggest a dishwasher, and you want to immediately tear every component off with 800psi?!

They're not stupidly fragile, but they are a bit fragile.
DO NOT USE A HIGH PRESSURE WASH.

The dishwasher is fine (wihtout soap). If you don't have one, toothbrush and handsoap is fine.
Drying takes about 4 hours in full sun after you've shaken it as much as possible.
Or, about 15mins in a WARM (~90*C) oven.
Or about 3 days inside, vertical.

Please don't fire up the BBQ and sit it on the coals to dry.

The solder didn't wick well because it's 100yo. Low melt solves that problem by mixing with the existing.

Haha, I'll hold off on the pressure washing and try both the dishwasher (I hope I can still eat from stuff that goes in there...) and a toothbrush for suspicious spots. Also, I was talking about the hand held, battery powered pressure washer, which aren't as strong as the gas powered ones, which are 2000 PSI.

But before that, try probing some voltages.

Personally I would have not done a thing with the board, since it worked and even the burnt PCI slot worked.
Or if the burn spots "really" bothered you: gently (using a good magnifier/microscope) shave/sand the colored plastic a bit using some micro tools.
I think between the hot air gun and the solder splatter you ruined it. Sorry.
You did teach me one thing: never blow the solder from the back side toward component side, even though I have not done it yet..
added: the other thing "don't fix it if not broken" really does apply to vintage stuff....

spacesaver wrote on 2025-03-02, 03:24:

(I hope I can still eat from stuff that goes in there...) a

A bit of lead never hurt no one. Just ask anyone from Detroit.

myne wrote on 2025-03-02, 03:05:

The dishwasher is fine (wihtout soap).

It's really not because dishwashers use salt in the water treatment system and hot salt water corrode metal parts. Only use a dishwasher if the salt compartment is empty.

where the hell do you live?

Depending on what attachment you used, the vibrations from the dremel can also cause bad contacts... anywhere really.
After all, that's one of the reasons (besides wider operating temperature and pressure ranges) why aviation and military spec components and PCBs have different standards than consumer grades stuff.

As for using the dishwasher to wash electronics, I also don't generally recommend that, especially if you don't turn OFF the drying cycle. Hot moist air + PCB = moisture ingress between the layers... and a hell of a lot more likely to make the PCB popcorn if you do any rework. So no, I prefer to hand-wash my stuff with warm water and don't really keep complex PCBs wet in the sink for longer than 10-15 minutes tops while I am washing.

OK, some progress. I put it in the dishwasher for almost 1 hour, without drying, then dried in the oven @170F. Still nothing. Then I realized, it was a comedy of errors:

1. The BIOS chip was wrong. It was a 128 instead of 256 KiB! When the board was working, I must've used a 256 KiB one taken from another board while waiting for the replacement to arrive (the original that came with the board had some OEM version from NEC and had stuck bits). That means the flasher, Xgpro doesn't even warn you that the file you're loading exceeds the ROM size!
I later confirmed that with the other 2 items rectified, the wrong BIOS causes the BUS clock LED on the POST card to not even light up. That means the BIOS must 1st initialize the northbridge to get a PCI clock.
2. The power supply (EVGA SuperNova 1300G2) was unreliable or incompatible. I forgot the symptom, maybe some abnormal beeps. Switched to a HDPLEX 250.
3. Only the 1st PCI slot is working, but I plugged the POST card into either the 2nd or 3rd.

After rectifying all 3 in that order, then it finally POSTed and booted to Windows 98. But the USB doesn't work (the host does show up in device manager without error) and PCI slots 2 & 3 don't work (after fixing all 3 issues, the POST card does show codes when in slot 1, but not in 2 & 3).
So that means the board might've been fine even before I started the thread 🙁

Lpt: hasten slowly.

spacesaver wrote on 2025-03-02, 14:10:

So that means the board might've been fine even before I started the thread 🙁

When someone finally invents a time machine, perhaps only then we will know if you do revisit it. 😁

Hey, still better to hear the board is working (albeit in somewhat crippled manner) than none at all.

It's possible the burned contacts in the original PCI slot might have been on a trace carrying power to these slots, causing the trace to open somewhere. If so, that could be why this and the other PCI slots below it are not working anymore.
Or maybe some logic in the Southbridge that handles PCI comm lines is permanently fuxzored??

OR OR OR... maybe you just need to let the board dry for a few more days (naturally, without any hot air or oven or other extreme measures.) After all, 170F = around 77C... which might not have dried all of the moisture out.
That doesn't mean re-bake it again - that's not so great for the electrolytic caps.
Gotta be more patient with these things, man! 😉

Ah, the address 13 line was broken. Most likely what happened, is it got delaminated due to excessive heating when removing the slot, and got ripped off.

After adding some bodge wires, all 3 PCI slots are working again! ISA and AGP also work, but still no USB (there is power coming out).

Personally I would have not done a thing with the board, since it worked and even the burnt PCI slot worked.

But the melted plastic means there's nothing for the pin to push against, so could lose contact. In the picture, you can barely see the shorted pins because they've been pushed inside. I might've had to push them out some for them to work again.

You did teach me one thing: never blow the solder from the back side toward component side

Actually, that's only creates a small puddle around the thru-hole on the component side. The extreme spatter on the component side is only when you blow from that side.

Wait, the USB was working. The reason I claimed it wasn't was because

1. A USB keyboard is working during boot, using the BIOS driver. Loading any USB host driver in Win98 (both NUSB 3.6 and the earlier Intel drivers from 2000) makes the keyboard stop working.
2. Somehow this motherboard's IO backplate blocks 1 of the USB ports. I had removed it to try the other USB port, but apparently it doesn't work. My other version 1.0 of the motherboard definitely allows both.

Anyone know what could be causing devices to not work with the Windows driver? I'm pretty sure the USB port worked 1 month ago, when I used it to copy ATI drivers. Is there any chance there's still a hardware fault, like maybe only the transmit line is connected, but not receive? Or is it because BIOS uses low speed (1.5 MBit/s) instead of full speed (12 MBit/s)?

let this debacle be a cautionary tale to some of you: BUY A VACUUM DESOLDERING PUMP