my zombi-puter has finally died.

[goontron]

my main desktop has been slowly dieing, first the video went out and i got a card for it, then the network quit and i got a card for that, then the audio and again a card. now both drive controllers, IDE and SATA died. shes dead jim.

PSU is known good. jap caps, good ripple, good fan, etc. it a delta PSU. so would this be a southbridge issue?

[shovenose]

Re: my zombi-puter has finally died.

Probably just a bad motherboard? IDK, time for a new one don't reuse any parts.

[goontron]

Re: my zombi-puter has finally died.

Probably just a bad motherboard? IDK, time for a new one don't reuse any parts.

HDD's and the audio/graphics are new so all thats coming over to the new build. the 100mb network card, meh, the new build will have Gb onboard network.

[Topcat]

Re: my zombi-puter has finally died.

what motherboard is it?

[goontron]

Re: my zombi-puter has finally died.

what motherboard is it?

an intel d945gcnl with a fucked up Powerspec OEM BIOS and a Pentium (left in the) D(ust).

[kaboom]

Re: my zombi-puter has finally died.

an intel d945gcnl with a fucked up Powerspec OEM BIOS and a Pentium (left in the) D(ust).

No big loss. I still think the 875s were better. As for the rotten netburst press-hot, who cares?


You going Intel or AMD?

Do you think the chipset had poor heatsinks/lousy cooling, to where the BGAs got stressed over many heat/cool cycles?

[goontron]

Re: my zombi-puter has finally died.

No big loss. I still think the 875s were better. As for the rotten netburst press-hot, who cares?


You going Intel or AMD?

Do you think the chipset had poor heatsinks/lousy cooling, to where the BGAs got stressed over many heat/cool cycles?

not sure yet.
the chipsets were well cooled. ive seen smaller heatsinks on AMD geodes.

[Topcat]

Re: my zombi-puter has finally died.

an intel d945gcnl with a fucked up Powerspec OEM BIOS and a Pentium (left in the) D(ust).

I've had a bunch of Intel desktop boards poop the bed on me for no apparent reason. Not as bad as Asus, but they're close. Intel server boards were a bit tougher though.

Unless you're running a nifty dual CPU dual core Paxville Xeon setup (which you aren't), yea, it probably is time to say goodbye to the netbust wonders. If you don't have a backup plan, let me know....I've found some other goodies in the attic that may be suitable replacements.

[goontron]

Re: my zombi-puter has finally died.

well, im going for used custom machine. i volunteer at a non profit here in denver that refurbishes computers and gives them to people who cant afford them. im going to buy a high power kit from the owner of the place for like $100. ive seen hardware with a 4000+ passmark, so thats what im going for.

i actually remember seeing server hardware with a SBS 2008 key on the side with a 4000+ passmark, i think its still sitting there. any IDE raid card s in the attic? these quantum bigfoots still haven't died.

[Topcat]

Re: my zombi-puter has finally died.

these quantum bigfoots still haven't died.

Why in the name of molasses in January would you want to use those?!

[goontron]

Re: my zombi-puter has finally died.

Why in the name of molasses in January would you want to use those?!

LOL.
because Linux still fits on one
maybe i like the sound of them running.

[stj]

Re: my zombi-puter has finally died.

they are reliable, unlike modern seagates.
well, unlike any seagates in my experience!

my first quantum was a 40meg scsi drive in my amiga2000.
it never did fail, it just got binned a couple of years ago.

[Wester547]

Re: my zombi-puter has finally died.

I've had a bunch of Intel desktop boards poop the bed on me for no apparent reason. Not as bad as Asus, but they're close. Intel server boards were a bit tougher though.

Sounds like it might have been either a Southbridge, Northbridge, or Super I/O chip (SMSC) failure (the latter of which is very commonplace), if not that possibly a fake batch of capacitors that would take more than a visual inspection to figure out (I believe some possibly fake Panasonic FJs have been spotted on Pegatron and ASUS boards here on this forum).

In any case, Intel's OEM, until a few years ago I believe, was Foxconn (and was ASUS until Intel decided to quite the motherboard business?). Foxconn did not make the best decisions with many of their motherboard designs, IE, placing hot running linear regulators all over the board in proximity of 85*C capacitors, allowing Prescott CPUs to run with two-phase CPU VRMs (and one-phase VRMs for Pentium III and older CPUs), etc. I would say Foxconn motherboards are more finnicky than anything else. I haven't had too many fail on me, though I don't doubt the possibility. Consequently and as with any other product, when millions of these boards are being produced, you are bound to have some lemons no matter the overall quality.

Goontron, you don't spot any KZGs on your board by any chance, do you? I'm wondering if even the presently good capacitors on the board, that came stock with it, might possibly be fake as well. Seeing as how the southbridge doesn't have a heatsink on that board I wonder if it's failing and if the Super I/O chip is another possible culprit.

[goontron]

Re: my zombi-puter has finally died.

Sounds like it might have been either a Southbridge, Northbridge, or Super I/O chip (SMSC) failure (the latter of which is very commonplace), if not that possibly a fake batch of capacitors that would take more than a visual inspection to figure out (I believe some possibly fake Panasonic FJs have been spotted on Pegatron and ASUS boards here on this forum).

In any case, Intel's OEM, until a few years ago I believe, was Foxconn (and was ASUS until Intel decided to quite the motherboard business?). Foxconn did not make the best decisions with many of their motherboard designs, IE, placing hot running linear regulators all over the board in proximity of 85*C capacitors, allowing Prescott CPUs to run with two-phase CPU VRMs (and one-phase VRMs for Pentium III and older CPUs), etc. I would say Foxconn motherboards are more finnicky than anything else. I haven't had too many fail on me, though I don't doubt the possibility. Consequently and as with any other product, when millions of these boards are being produced, you are bound to have some lemons no matter the overall quality.

Goontron, you don't spot any KZGs on your board by any chance, do you? I'm wondering if even the presently good capacitors on the board, that came stock with it, might possibly be fake as well. Seeing as how the southbridge doesn't have a heatsink on that board I wonder if it's failing and if the Super I/O chip is another possible culprit.

All caps in the VRM are pollys, i recaped the thing when I got it, so all panny, or rubby caps.

[RJARRRPCGP]

Re: my zombi-puter has finally died.

It probably has KZGs. Just a heads up.

[BigTroll]

Re: my zombi-puter has finally died.

I have a 945gcnl also, mine i put a e6600 with 8800gts and its run off a X-CLIO GOODPOWER! all caps holding up. mine also has KZG on the VRM but since I recapped the X-CLIO I know its getting GOODPOWER! no pun intended.

[c_hegge]

Re: my zombi-puter has finally died.

The D945GCNL board is actually made by ASUS, which explains it all. From the pictures I've seen of it, it has the tell-tale backwards polarity markings on the caps. I've had a few ASUS made Intel boards die out of the blue (just like ASUS branded boards), but the Foxconn made ones seemed to be bit better.

[Heihachi_73]

Re: my zombi-puter has finally died.

This sounds exactly like my POS Intel desktop board, a D865PERL. All Rubycon caps, none failed, random BSODs every single time it's turned on and randomly fails memtest86+ regardless of which RAM is used. Ironically it's also being run with an ancient Quantum relic from the 1990s (a 6GB 3.5" unit; I do have a 4.3GB 5.25" Bigfoot in the garage though). Of course, it didn't help that its 3GHz Prescott CPU hit 93°C once, someone (looks in mirror) managed to sit the heatsink/fan on top of one of the clips after doing a CPU swap! Also, swapping the CPU with a 2.8 Northwood did nothing whatsoever, just slowed it down ever so slightly.

The only thing I didn't look at was cracked solder joints around the RAM; removing the RAM slots just to look for that is a waste of time anyway - P4 boards simply aren't worth the trouble these days and they are everywhere, if one goes bad, swap boards and wait for XP to readjust to the handful of different parts in the new system (or reinstall/repair if it does happen to shit itself). I might just reclaim those Rubies and part it out, I really have no use for it.

[Wester547]

Re: my zombi-puter has finally died.

^ Sound similar to a very bewildering issue I encountered with an Intel 850 series Foxconn motherboard the other day, and two of them at that (two Dell Dimension 8200s). I don't know if a BIOS update would fix them as one of the BIOS updates is said to "modify the SMI handler to prevent spurious shutdowns", but evidently, those earlier boards at least don't like Newton/Delta power supplies at all and only seem to run stably with Hipros. When I replaced a 250W Hipro with a 250W Newton with all Japanese capacitors (Rubycon YXG and Nichicon PW) I started getting all sorts of RAM-related errors and this happened across two different motherboards with two different sets of RAM (128MB PC800 45ns Elpida RDRAM x2 and 128MB PC800 45ns Samsung RAM x2 to make 512MB total, one board has the ECC versions of that Samsung RAM and the other non-ECC but those Dell boards don't support ECC RAM anyway so they all run in non-ECC mode). These boards have good capacitors, populated with Chemicon LXZ and Nichicon PW and Nichicon NA polymers in the VRM output, and the 250W Newton with Japanese caps runs fine on any other motherboard with a 20-pin connector.

The BIOS would complain that the system memory has changed and that it's reducing the available memory to prevent address errors, I'd get blue screens, random restarts, the computers would shut down instead of restart when I saved BIOS changes (though the changes saved anyway), etc... put a Hipro back in there and all those problems magically disappate, and they run completely stable. What I found ironic is that it was actually stable with another 250W Newton with LTEC capacitors, but it would occasionally give me RAM related errors (memory leaks) and would still shut down upon saving BIOS changes unlike the Hipros (the 250W Newton w/LTEC had a lesser transformer whine than the 250W Newton w/Japanese capacitors, and the Hipro had none at all, but I don't think that's important). I tested both boards with 300W and 250W Hipros and they ran fine. I don't know if the 'cleaner' output of the linear regulated +3.3V rails in the Hipros (as opposed to the magamp approach of the Newtons) has anything to do with it but I have never experienced anything like that with any other motherboard....

[kaboom]

Re: my zombi-puter has finally died.

^ Sound similar to a very bewildering issue I encountered with an Intel 850 series Foxconn motherboard the other day, and two of them at that (two Dell Dimension 8200s). I don't know if a BIOS update would fix them as one of the BIOS updates is said to "modify the SMI handler to prevent spurious shutdowns", but evidently, those earlier boards at least don't like Newton/Delta power supplies at all and only seem to run stably with Hipros. When I replaced a 250W Hipro with a 250W Newton with all Japanese capacitors (Rubycon YXG and Nichicon PW) I started getting all sorts of RAM-related errors and this happened across two different motherboards with two different sets of RAM (128MB PC800 45ns Elpida RDRAM x2 and 128MB PC800 45ns Samsung RAM x2 to make 512MB total, one board has the ECC versions of that Samsung RAM and the other non-ECC but those Dell boards don't support ECC RAM anyway so they all run in non-ECC mode). These boards have good capacitors, populated with Chemicon LXZ and Nichicon PW and Nichicon NA polymers in the VRM output, and the 250W Newton with Japanese caps runs fine on any other motherboard with a 20-pin connector.

The BIOS would complain that the system memory has changed and that it's reducing the available memory to prevent address errors, I'd get blue screens, random restarts, the computers would shut down instead of restart when I saved BIOS changes (though the changes saved anyway), etc... put a Hipro back in there and all those problems magically disappate, and they run completely stable. What I found ironic is that it was actually stable with another 250W Newton with LTEC capacitors, but it would occasionally give me RAM related errors (memory leaks) and would still shut down upon saving BIOS changes unlike the Hipros (the 250W Newton w/LTEC had a lesser transformer whine than the 250W Newton w/Japanese capacitors, and the Hipro had none at all, but I don't think that's important). I tested both boards with 300W and 250W Hipros and they ran fine. I don't know if the 'cleaner' output of the linear regulated +3.3V rails in the Hipros (as opposed to the magamp approach of the Newtons) has anything to do with it but I have never experienced anything like that with any other motherboard....

Did you put the scope on it? Maybe ripple, or really just cyclical current pulses from the board somehow upset the loop in those supplies. Like a few years ago when certain video cards, with external power connections, despite being below the limits of certain supplies, would cause them to shut of or become unstable during gaming or other GPU-intensive tasks.