Check the +12V line while under a heavy +5V load. Could be cross-loading problems with that PSU; low +5V and high +12V issue. Use a good calibrated DVM to catch the spikes if any.

Thanks guys for your responses and help. Well, as I was typing my last post I heard a loud click. Looked over at the 98 machine and it had turned off, but this time no smoke. This is because I had the CPU fan plugged directly into the PSU via a 3pin to 4pin D-connector adapter. The SF450 has all sorts of safety trips when something goes wrong and sure enough the PSU shut down as soon as a problem arose. So I unplugged the CPU fan and turned on the system and everything turned on immediately. I turned it off quickly since the CPU copper heatsink only gives me 5-10 seconds before the CPU gets uncomfortable. I plugged the Startech CPU fan back into the PSU directly and the system turns on for just a second and turns off immediately. If this isn't a clear sign of the CPU fan having a short or some sort of electrical problem I don't know what is. I think I found my problem.

Once again, though, it took a couple of hours of the system just sitting there at the desktop idling before the problem came up. Doesn't seem to matter if the system is under load or not and usually happens when it is just idling. Looking forward to testing with the new CPU heatsink/fan.

Holy !@#$! Found the problem. I am really pissed at myself (for not checking) and at Startech. So I inspect the CPU fan a little bit and sure as shit the red wire is completely exposed and bare wire touching the copper heatsink. No freakin wonder. Basically whoever screwed down the fan during manufacturing (robot/person) clamped down on the wire breaching its sleeve. Wow. Something so simple can be so destructive. Doubt I will get anywhere with Startech since I bought this through a reseller new in box via Ebay, but we'll see.

Let me ask an electrical question. If the red wire is exposed and touching the copper heatsink why did I not receive a shock when I touched the heatsink while the system was on?

Here is a photo of the lovely offender. Bet anyone doing a build now is going to double check their CPU fan wires. The sickest part of it is this is sold as a 1U server Socket 370 cooling solution. Can you imagine if this had happened to an expensive server board?

LOVE, just LOVE the QC sticker in this shot. It was basically stuck on right next to the exposed wire. 🤣

liqmat wrote:

Let me ask an electrical question. If the red wire is exposed and touching the copper heatsink why did I not receive a shock when I touched the heatsink while the system was on?

Because you have there 12v and very little current (probably 0.1A or so). It appears that my second guess was correct and the two incidents were unrelated.

liqmat wrote:

The power button breaking was the plastic that broke the camels back. I basically laughed when that happened.

Don't give up. The harder the project the bigger the satisfaction once it's completed 😁

I believe both incidents are related. The Slotkets checked out fine and the BX motherboards actually are able to handle the P-III 850 FSB100 Coppermine in a slotket configuration according to all the specs I checked. This was the MSI slotket v2 which is highly compatible. I would bet I could setup a BX board with the MSI slotket now with a new CPU fan/heatsink without issue. I am actually going to repair the least damaged BX board and give it a whirl and see if I'm right.

The seller discounted the power buttons for me so I have a few on the way. Easy fix.

liqmat wrote:

Here is a photo of the lovely offender. Bet anyone doing a build now is going to double check their CPU fan wires. The sickest part of it is this is sold as a 1U server Socket 370 cooling solution. Can you imagine if this had happened to an expensive server board?

LOVE, just LOVE the QC sticker in this shot. It was basically stuck on right next to the exposed wire. 🤣

fanwire.jpg

Wow..this cost you several motherboards 😵
It's a good idea to always do a visual inspection, even if something is brand new (as something may also by chance get damaged during transit).

Thanks for documenting this 😀

A little good news is that the Ebay reseller has refunded the cost of the fan/heatsink after I explained what happened even though it's not their fault. We'll see how Startech responds if they do at all. Luckily the Intel D815EEA works just fine minus the blown CPU fan plug so I will continue to use that motherboard while I repair the least damaged Intel BX board. As mentioned in another Vogons thread BX boards have gone up in price lately and so has the D815EEA and I am glad I am not forced to pay those prices.

Adding a 1/4 amp slow blow fuse inline with the other fan headers on the motherboard would be a good idea.

TOBOR wrote:

Adding a 1/4 amp slow blow fuse inline with the other fan headers on the motherboard would be a good idea.

Thank you and good idea. Going to take a break from this for a few days. I feel relieved, at least, finding the source of the problem. Thanks for everyone's input.

Well, some progress, Startech has taken an interest in this via email and I will be linking them to this thread to show what transpired. I'll let you guys know if they help me out or not. Who knows, there might be a happy ending to all of this.

liqmat wrote:

Here is a photo of the lovely offender. Bet anyone doing a build now is going to double check their CPU fan wires. The sickest part of it is this is sold as a 1U server Socket 370 cooling solution. Can you imagine if this had happened to an expensive server board?

LOVE, just LOVE the QC sticker in this shot. It was basically stuck on right next to the exposed wire. 🤣

This gave me the motive to open up my P1 build to check the CPU fan cable and it looks good so thankfully I'm not affected by this. I chose a Startech heatsink/fan (the larger black one, which isn't the one you have) for that build back when I built it and it's still working great 4 months later. My fan is manufactured by Evercool if that matters.

liqmat wrote:

Well, some progress, Startech has taken an interest in this via email and I will be linking them to this thread to show what transpired. I'll let you guys know if they help me out or not. Who knows, there might be a happy ending to all of this.

I feel bad that you ended up with a DOA CPU fan. On the bright side at least your PSU and motherboards are fine (aside from the damaged fan header). I guess the manufacturers at Startech will start double-checking the heatsink assembly lines to make sure this doesn't happen again from now on once they see that picture. 😀

Good to hear that your Gateway build is back on track now.

Actually both my BX motherboards fried and I might try repairing one of them. The D815EEA board lost the fan header. So that fan actually cost me three motherboards technically since a damaged motherboard is still unacceptable due to a faulty fan.

liqmat wrote:

Actually both my BX motherboards fried and I might try repairing one of them. The D815EEA board lost the fan header. So that fan actually cost me three motherboards technically since a damaged motherboard is still unacceptable due to a faulty fan.

If you wanted to keep using the D815EEA motherboard you can get away with using a Molex cpu fan, or use a 3pin to molex adapter if you want to keep using a 3pin fan. It should halt the boot in the bios, but it should prompt you to continue past it by pressing F2.

ODwilly wrote:

liqmat wrote:

Actually both my BX motherboards fried and I might try repairing one of them. The D815EEA board lost the fan header. So that fan actually cost me three motherboards technically since a damaged motherboard is still unacceptable due to a faulty fan.

If you wanted to keep using the D815EEA motherboard you can get away with using a Molex cpu fan, or use a 3pin to molex adapter if you want to keep using a 3pin fan. It should halt the boot in the bios, but it should prompt you to continue past it by pressing F2.

If you read a few posts back that's what I'm doing. Just plugging the CPU fan (a replacement unit for now of course) direct to the PSU bypassing the burned out 3pin. The motherboard would never detect that anyway and the BIOS would not halt on that. Either way, all three motherboards were damaged due to this faulty fan. Startech, so far, has agreed to replace the faulty fan, but I also asked them to reimburse me for two of the three motherboards. The first motherboard I purchased as part of the entire system for close to nothing. The other two I bought on Ebay. We're not talking a lot of money here. $50 total for both boards including shipping. They referred me to a limited liability policy of theirs, but said they would discuss in a meeting whether they would help considering the unusual circumstances involved. I have provided them with all receipts and a link to this thread. We'll see what happens.

My goal is to get reimbursed and buy one more mobo so I don't have a damaged board in this system. It's unfortunate because the Startech copper heatsink/fan 1U design is ideal for a Slotket configuration and did a damn good job paired with IC Diamond TC. I probably will end up using the replacement Startech unit, after physical inspection of course, when I start rebuilding the system. Right now my replacement cooler arrived which is a NIB Antec aluminum body/copper core unit while I wait for their decision.

Just a quick update:

Startech.com had a meeting and made an exception in my case and provided me with a $50 credit to their store. Very cool of them and the replacement copper heatsink/fan unit works great as well. A happy ending to a harrowing ordeal. Thanks Startech!