Had a floptical drive for a Silicon Graphics Iris Indigo that wouldn't read floppies. To test if it was the special Teac-chipset SCSI to floppy bridge or not I haphazardly plugged it into another machine with SCSI but didn't observe pin 1. Fried it. I've never seen a SCSI device fry like that before but that was the last time I used unkeyed SCSI cables.

Managed to plug in the DC cable for a regular 3.5" floppy drive upside down. You don't need to guess what that did.

Desoldered and removed a VRM heatsink on an 865PE Neo2-Platinum that had cap issues. When reinstalling it I didn't clamp the heatsink in place so it floated over the regulator. It got about 1/2 into booting XP before it bluescreened and dropped stone dead with a shorted regulator. Took three of the four sticks of 1gb PC3200 in the process.

That ceramic pile of deceased Duron and Athlon CPUs I have is looking at me and it's creepy

LewisRaz wrote on 2021-05-07, 07:35:

I slipped while fitting the socket 370 heatsink and destroyed a tiny IC next to the socket.

I had a similar problem with the ECS K7S5A with the Titan D5TB cooler. After installation, its wide buckle touched the resistor field next to the socket and shorted it out when the PC was turned on. Luckily, the board continued to work.

appiah4 wrote on 2021-05-06, 11:57:

[...]

I have a board with this exact issue; I wonder if I can just replace the socket and it would work? Or does this mean traces/components on that pin's rail already got wrecked?

Can't comment on your board, but see the rest of my post you quoted - that board had bashed-in pins on the chipset and multiple burnt out chips. There was absolutely no point in trying anything more there so I didn't.

I was doing some reflowing on the GUS Max 2.1 card since it would lock up at the detection using the DOS setup program or when I enable it in Windows 95C's device manager. I reflowed the following to see if the issue would get fixed:

Gravis GF1 ICS1614 (brains for the operation)
Crystal CS4231A-KL (sound processor)
Advanced Gravis GF1D1 (controller of some kind)
Gravis/NEC chip (not seeing the model #)
NEC 42S4260-70 512KB RAM

I then reinserted the card after I cleaned the flux off and let the 99.9% IPA dry and I thought the issue would get fixed, right? Dead wrong.

Apparently, the ICS1614 GF1 chip is done for and the card isn't detected any longer (throws an error saying the HW address is in use, the CD drive isn't detected, and so on).

I'm planning on removing the GF1 chip and the CS4231A-KL chip to see what happened and remove excess solder in case too much got stuck underneath.

@bjwil1991 - You are whinging, should be more proud of your killings.

I assembled a 486 on the test bench today and the BIOS decided not to boot and now requires a password (!?!?). I don't know what happened. Static discharge maybe? Luckily, I have ROM file for it. Would anyone be willing to flash a new BIOS for me?

Not one, but TWO Gigabyte GA-K8VM800M died of BIOS failure while playing some retro games. 🙁 They rebooted and asked for BIOS on floppy; resetting the machine bricked the first MB, while floppy reading error bricked the second. It was very frustrating when you thought you've prepared a floppy with BIOS just in case.

Those Award BIOS chips were in PLCC32 package. However, instead of socketed, they were soldered on the MB. Flashing them would require desoldering, followed by a chip programmer via an adaptor (modern systems usually use UEFI chips in DIP8 package, so a clip can be used to connect the chip without desoldering). Way too troublesome; bought an Asus K8V-MX with AMI BIOS instead (no more vintage Gigabyte MB with Award BIOS).

That's the one thing I hate: soldered-on BIOS/ROM chips. Even cheap Chinese boards (clones) are like that as well. I'm planning on installing a socket on my Mega Memory Card for my GameBoy saves and check to see if they are still there in case my games lost their saves after many years.

And the main reason why I'd like to add sockets to my GUS Max 2.1 for easier chip swapping without having to break out my hot air station to replace the chips in case they go bad periodically.

Does this card contains any SMD fuses? I personally found several times that SMD 0805 or 1206 sized fuses got several kilo ohms after heated for longer period with hot air. Please check power supply lines for mentioned ICs.

bjwil1991 wrote on 2021-08-01, 22:40:

I'm planning on removing the GF1 chip and the CS4231A-KL chip to see what happened and remove excess solder in case too much got stuck underneath.

Try to inspect surface of package of this IC. If you find some bubbles or cracks on it this can be some signal that infamous popcorn issue happened to this IC. Hopefully this is not the case and something else around ICs happened.

1nanoprobe wrote on 2021-08-01, 23:31:

I assembled a 486 on the test bench today and the BIOS decided not to boot and now requires a password (!?!?). I don't know what happened. Static discharge maybe? Luckily, I have ROM file for it. Would anyone be willing to flash a new BIOS for me?

Standard 486 BIOSes do not store passwords in flash, but in the CMOS RAM. If the board has a clear CMOS process, just use it. If not, and it has a dedicated battery, pull out the battery for a day. Worst case: No clear CMOS process, and the CMOS chip has an integrated battery and no external clear pin (DS12C887 without "A"). In that case, try to clear the chip using a different board.

Also: Google for master passwords. In case it's an award BIOS, try "lkwpeter". This password has the same hash code as the intended password "award_sw". The advantage of "lkwpeter" is that it contains no characters that depend on the keyboard type. In case it's an AMI BIOS, try "ami".

dormcat wrote on 2021-08-02, 00:04:

Not one, but TWO Gigabyte GA-K8VM800M died of BIOS failure while playing some retro games. 🙁 They rebooted and asked for BIOS on floppy; resetting the machine bricked the first MB, while floppy reading error bricked the second. It was very frustrating when you thought you've prepared a floppy with BIOS just in case.

Those Award BIOS chips were in PLCC32 package. However, instead of socketed, they were soldered on the MB. Flashing them would require desoldering, followed by a chip programmer via an adaptor (modern systems usually use UEFI chips in DIP8 package, so a clip can be used to connect the chip without desoldering). Way too troublesome; bought an Asus K8V-MX with AMI BIOS instead (no more vintage Gigabyte MB with Award BIOS).

What is you general opinion regarding the reliability of Gigabyte boards? In my experience it more of a hit and miss, with for example 1.0 boards being somewhat overfitted wrt/ capacitors and such, and subsequent revisions on the cheaper side which is evident with visual inspection.

I THOUGHT I'd killed my entire PC a month or two ago after putting in an Aureal Vortex 2. I went to turn it on after install and the power blipped but ultimately didn't turn on, and then the sweet, horrifying scent of electrical burning wafted from the case.

I tore open the PC and pulled the Aureal Vortex 2 out and gave it a quick smell and, yeah, the nice electrical burning smell was coming from that card. However, when I removed the CD-in cable between the card and the CD-ROM drive everything worked fine, including the drive and the vortex 2. No issues since, though I've not put a CD-in cable back in the PC at all (I don't think I run anything that requires it accept the original Half-Life). I'm not really sure what the issue is, both the card and drive denote the CD-in/out pinouts as R-G-G-L, so even if the connectors were reversed on the two ends there shouldn't have been a short.

I've since put the Vortex 2 in the bin and am running a Labway YM719 card and can't find the pinout for the CD-in connector on that card, so I'm not risking it right now, 🤣.

cde wrote on 2021-08-02, 12:15:

What is you general opinion regarding the reliability of Gigabyte boards? In my experience it more of a hit and miss, with for example 1.0 boards being somewhat overfitted wrt/ capacitors and such, and subsequent revisions on the cheaper side which is evident with visual inspection.

You've just spoken out the general and most accepted opinion on Gigabyte in its home country. 😉 Unfortunately those two GA-K8VM800M were Rev. 2.0 boards.

On the other hand, its driver database for legacy products is much better than that of Asus and MSI: the latter two must manually pick an OS first instead of simply listing them all (and Win9x compatible drivers are often labeled with just one OS; you have to switch through OS in order to make sure all drivers have been downloaded); sometimes a not-so-old legacy product gets delisted. ASRock is roughly on par with Gigabyte in this regard.

For the record, Gigabyte has the same photo for Rev. 1.0 and 2.0 on its website, but Rev. 2.0 actually looks exactly the same as GA-K8VM800M-RH; the latter simply acquired RoHS compliance.

This is not how to repair something thread, create some other one for it..

I killed 2 or 3 floppy drives before i figured out that I had switched 12v and 5v in the molex connector for a fan I just had installed. Not my best day...

Just remembered another incident: In mid-2000's (should be 2005 or 2006) I turned the power on to test an used Acer Socket 370 MB without realizing it sitting on the top of a metallic case of another computer. Saw a small spark and the MB bricked. One PCI slot has some solder points yellowed.

mkarcher wrote on 2021-08-02, 07:50:

Standard 486 BIOSes do not store passwords in flash, but in the CMOS RAM. If the board has a clear CMOS process, just use it. If not, and it has a dedicated battery, pull out the battery for a day. Worst case: No clear CMOS process, and the CMOS chip has an integrated battery and no external clear pin (DS12C887 without "A"). In that case, try to clear the chip using a different board.

Also: Google for master passwords. In case it's an award BIOS, try "lkwpeter". This password has the same hash code as the intended password "award_sw". The advantage of "lkwpeter" is that it contains no characters that depend on the keyboard type. In case it's an AMI BIOS, try "ami".

Thanks for the help. There's actually no battery connected. Also, tried "ami" password. There's no clear CMOS jumper, so I will have to try in another board. Only have non-working ones so hopefully they are good enough to clear the CMOS.

Is it normal for these BIOSes to randomly crap out and require passwords when none were previously set?

1nanoprobe wrote on 2021-08-03, 21:15:

Is it normal for these BIOSes to randomly crap out and require passwords when none were previously set?

I had this happen only when I changed BIOSes on a board, so that the CMOS layout was incompatible, but the checksum scheme was identical. The new BIOS recognized that the CMOS checksum is valid, but interpreted it in a completely wrong way and thus assumed a password was set. On AMI BIOSes, you might have success holding "INS" during power-up for a software-triggered CMOS clear/override. If INS is accepted as hotkey, you should get a beep response. I don't know whether your board/BIOS supports this feature, though.

I reinstalled a 486 DX2-50 cpu back into the socket incorrectly, didn’t line up the notches. This was in a Leading Edge 486 WinPro system, my first PC. When I powered it on I got a pretty good spark and a corner of the chip had cracked completely through. Thankfully this wasn’t my main machine at the time.