VOGONS


First post, by pixelatedscraps

User metadata
Rank Member
Rank
Member

Or just the PSU?

PSU was set to 110v and I’m in Hong Kong with 220v. Completely forgot that old PSUs used to come with a voltage switch, plugged it in and it fizzed.

Have I killed other components aside from the PSU?

I ask because I can’t get any signal from the computer despite trying a variety of RAM, VGA cards and HDDs and was wondering if I had fried something else too.

Motherboard is an Asus P3B-F and was listed with screenshots as working before the sale and subsequent amateur power hour.

Main: PII 2x 333Mhz | P2L97-DS | 128 | Millenium II 8MB | Voodoo2 SLI 12MB | AWE64 Gold CT4390 | Cheetah 9.1 | InWin A-500 | Win98 / NT 4.0

Test bench: Asus P3B-F | 1.3Ghz Tualeron w/ Powerleap | Geforce 2 Ti500 | SB Live! 5.1 CT4760

Reply 1 of 12, by mkarcher

User metadata
Rank Oldbie
Rank
Oldbie

Usually, the secondary side protection (using zener diodes) on a quality PC power supply is strong enough to prevent dangerous overvoltage to the mainboard and disk drives until the primary side breaks down in your scenario. So I would have expected no damage except the fried primary side of your PSU.

OTOH, if the components on the secondary side are dead, they are dead. But I suggest to thoroughly test what's broken before throwing it all to the bin.

Reply 2 of 12, by retardware

User metadata
Rank Oldbie
Rank
Oldbie

The zeners mentioned by @mkarcher are not designed to cut a surge like this.
Power was quadruple.
Not unlikely that you made a veritable total kill.

Had such things already in my repair practice.
Maybe some chips survived, but they should be considered as compromised due to the invisible damage.
Latent failures are to be expected.

Reply 3 of 12, by mkarcher

User metadata
Rank Oldbie
Rank
Oldbie
retardware wrote on 2021-09-19, 12:34:

The zeners mentioned by @mkarcher are not designed to cut a surge like this.
Power was quadruple.

"Power was quadruple" would apply on resistive loads if the voltage is doubled. Zeners are even worse: They take virtually no power if the supply is OK, but take a lot of power (way more than four times virtually nothing) on 20% overvoltage. They get hot very fast, but they tend to fail safe: They become a short circuit if the core temperature exceeds 200°C.

Furthermore, a PC switch-mode supply is regulated. It will not output twice the voltage just because the input voltage is twice as high. The primary caps tend to fail within seconds (usually short, blowing the fuse) on 220V into a 110V unit, so the amount of energy available until catastrophic failure of the supply is limited. If the secondary side overvolts at all, it is possible that the zeners short out, kind of like a single-use crowbar circuit, but I doubt there is enough energy to melt the bond wires, which would allow overvoltage to leave the supply and damage the PC.

That's why I still have high hopes that no serious damage outside the supply was caused by the incident. Yet, of course, I wouldn't use the affected components in mission critical systems anymore.

Reply 5 of 12, by retardware

User metadata
Rank Oldbie
Rank
Oldbie
mkarcher wrote on 2021-09-19, 13:05:

Furthermore, a PC switch-mode supply is regulated. It will not output twice the voltage just because the input voltage is twice as high. The primary caps tend to fail within seconds (usually short, blowing the fuse) on 220V into a 110V unit, so the amount of energy available until catastrophic failure of the supply is limited. If the secondary side overvolts at all, it is possible that the zeners short out, kind of like a single-use crowbar circuit, but I doubt there is enough energy to melt the bond wires, which would allow overvoltage to leave the supply and damage the PC.

There are some more things that can happen:
Rectifier and/or transistors short out and/or open.
Fuse is slo-blo, and arc might persist a while.
Zeners are slow, not fast diodes designed to cut down flyback voltages much higher than in normal operation.
There is enough possibility to have the mobo etc zapped.

I wish it were not this way, but the OP's examinations post-mortem look like that there is a good portion of damage.

Reply 7 of 12, by pixelatedscraps

User metadata
Rank Member
Rank
Member

Wow, you guys are good. I’m assuming some past / present lives lived as engineers.

The new PSU I’m using appears to be working, but I’ll setup a proper test desk once I get time this week and test everything methodically. I have two other PII motherboards I can test the other components with as well.

Main: PII 2x 333Mhz | P2L97-DS | 128 | Millenium II 8MB | Voodoo2 SLI 12MB | AWE64 Gold CT4390 | Cheetah 9.1 | InWin A-500 | Win98 / NT 4.0

Test bench: Asus P3B-F | 1.3Ghz Tualeron w/ Powerleap | Geforce 2 Ti500 | SB Live! 5.1 CT4760

Reply 8 of 12, by pixelatedscraps

User metadata
Rank Member
Rank
Member

Update here, I moved everything over to an open bench table, took everything apart and cleaned each component.

I get:

• single beep on power on
• green light on motherboard
• CPU fan spinning
• HDD spinning up
• RAM seems to work (remove both sticks and I get the long beep RAM error)
• tried both the original VGA card (Cirrus Logic CL-GD5434 QC-F) and a Matrox Millenium II with VGA out
• no signal on monitor

I am connected to my rather modern Dell S3221QS 32’ 4K monitor via VGA to HDMI adapter with 5v plugged into an empty USB port. Could it be this? I hope so. My Dell 2001fp arrives in a few days so I can further test this with a standard VGA cable without the need for an adapter.

Main: PII 2x 333Mhz | P2L97-DS | 128 | Millenium II 8MB | Voodoo2 SLI 12MB | AWE64 Gold CT4390 | Cheetah 9.1 | InWin A-500 | Win98 / NT 4.0

Test bench: Asus P3B-F | 1.3Ghz Tualeron w/ Powerleap | Geforce 2 Ti500 | SB Live! 5.1 CT4760

Reply 10 of 12, by pixelatedscraps

User metadata
Rank Member
Rank
Member

It’s alive!

Well, only after I switched to yet another PSU - this time a Seasonic after getting zero display connection from the initial replacement Fortron unit.

Attachments

Main: PII 2x 333Mhz | P2L97-DS | 128 | Millenium II 8MB | Voodoo2 SLI 12MB | AWE64 Gold CT4390 | Cheetah 9.1 | InWin A-500 | Win98 / NT 4.0

Test bench: Asus P3B-F | 1.3Ghz Tualeron w/ Powerleap | Geforce 2 Ti500 | SB Live! 5.1 CT4760

Reply 11 of 12, by retardware

User metadata
Rank Oldbie
Rank
Oldbie

Glad it worked out well!
Nice test bench btw 😀 And a nice background green view, too 😀

And it confirms again why it is not a good idea to use a Fortron / FSP, at least not as test bench PSU 😀

Reply 12 of 12, by pixelatedscraps

User metadata
Rank Member
Rank
Member
retardware wrote on 2021-09-25, 00:24:

Glad it worked out well!
Nice test bench btw 😀 And a nice background green view, too 😀

And it confirms again why it is not a good idea to use a Fortron / FSP, at least not as test bench PSU 😀

Thanks! My spare room overlooks Hong Kong’s oldest mosque and these trees form the boundary between the apartment and the grounds below the mosque. The living quarters behind the mosque must be at least 90-100 years old and are in quite a state of decay but the mosque is still going strong with a fresh lick of paint every few years.

I’m tempted to send the Fortron to e-waste recycling and pick up another of these Seasonic units to keep as spare.

Main: PII 2x 333Mhz | P2L97-DS | 128 | Millenium II 8MB | Voodoo2 SLI 12MB | AWE64 Gold CT4390 | Cheetah 9.1 | InWin A-500 | Win98 / NT 4.0

Test bench: Asus P3B-F | 1.3Ghz Tualeron w/ Powerleap | Geforce 2 Ti500 | SB Live! 5.1 CT4760