Hi TeAMmates,
I expect to gain some "traction" in SIMAP and Rosetta in February. Both projects have resisted the attentions of my hordes to this point.
The local horde is reeling from a spate of recent failures! November, December, and January have seen a record number of apparent failures in the local horde.
[*]Guardian (Tyan S2460 Dual Processor Motherboard) - This computer ran pretty much flawlessly for about 42 months. It began with a random crash. On reboot, a processor failed to be recognized. Testing has shown that the Antec TruPower 480 PSU has a failed power ready output which shouldn't be that big of a deal since all other outputs are within normal tolerances. Further testing showed one of the MP2000 processors is dead. The 512MB PC2100 ECC REG RAM refuses to be recognized by any motherboard that I have so I presume it is dead too! I replaced the suspect PSU and I pulled the Shuttle AN35N 400 Ultra Motherboard and Athlon Mobile XP3000+ from my library computer and now Guardian lives again. I upgraded the RAM to 768MB from 512MB. The Household File server lives again!
[*]Tyan S2462 Dual Processor Motherboard - a replacement for the S2460 in Guardian that was DOA.
[*]Bikini (Shuttle AN35N 400 Ultra) just crashed and failed to post ever again. Bad Caps in a cheap PSU! A lesson learned. Bikini may yet live again but that is a down the road thing as this computer is a pure cruncher.
[*]TheBrain (Shuttle MN31N) I built three identical micro ATX based computers in 2002. Trinity, Pinky, and TheBrain, to replace 3 aging dual pentium 200 computers. All had Shuttle MN31N motherboards, Athlon XP2400+, and 2 x 256MB Golden Dragon Value PC3200. TheBrain has experienced random Lock-Ups (usually about every other month) for about the last year. I finally decided it was time to upgrade it to Windows 2003 Server. I saved everything off ... installed the new OS and TheBrain turned up lame. Memory tests confirm that something is amiss with the motherboard or processor or onboard video memory. There are very apparent anomalies in the video during the memory testing. I'll pick the motherboard as the failed component. The PSU tests OK but I have never been pleased with the voltages from it so it will be history too. At this point, a replacement PSU, motherboard, and processor are on the way. This will be the first of the 3 to be upgraded to an Athlon 64 3200+ Venus. I'll test the XP2400+ and, if OK, will replace a XP2000+ in another computer ... a fairly nice upgrade for both systems. Alas, an Opteron or Dual Core was not in the budget this time
[*]Smoke (Shuttle MN31N) Yet another micro ATX based computer built in 2002. This has a transparent smoke gray plexiglass case. Apparently, the video RAM has failed in the onboard video. I added an AGP card and all seems well for the moment.
[*]Ghost (ancient ASUS Duron system) Another motherboard has apparently failed. Again, this computer may live again but not in the short term as it was only a cruncher.
[*]At the strong urging of the wife, I have divested several other sub 1GHz computers. The local horde is now less than a dozen processors/cores.
I am beginning to formulate a few theories from these failures:
[*]Processors and/or memory may tolerate long term overclocking but, perhaps, motherboard chipsets may not be as durable or tolerant. I don't believe DC projects bear on this one way or another.
[*]Cheap and/or generic power supplies are a waste. I will bite the bullet and purchase only Active PFC power supplies in the future. If for no other reason, the high efficiency of these PSU's should have the effect of reducing electricity consumption. Because of all the advanced features of these PSU's and the high cost, these power supplies seem to be built with better components to go the distance. MTBF/MTF for these PSU's is almost double that of the low cost generics. A bonus is that most of these advanced PSU's are genuinely QUIET.
[*]I keep my systems running cool. I routinely clean out the dust that seems to accumulate. I absolutely believe that if I did not keep my systems cool and clean, the failure rate would be unacceptable.
[*]Bad Caps in PSU's remain a problem even for some presumed high end PSU manufacturers like Antec. I don't find any visible evidence of Bad Caps on these failed mothboards but I have repeatedly seen cheap and generic PSU's with Bad Caps.
So much for the ramblings of a deranged cruncher.
mondo
I expect to gain some "traction" in SIMAP and Rosetta in February. Both projects have resisted the attentions of my hordes to this point.
The local horde is reeling from a spate of recent failures! November, December, and January have seen a record number of apparent failures in the local horde.
[*]Guardian (Tyan S2460 Dual Processor Motherboard) - This computer ran pretty much flawlessly for about 42 months. It began with a random crash. On reboot, a processor failed to be recognized. Testing has shown that the Antec TruPower 480 PSU has a failed power ready output which shouldn't be that big of a deal since all other outputs are within normal tolerances. Further testing showed one of the MP2000 processors is dead. The 512MB PC2100 ECC REG RAM refuses to be recognized by any motherboard that I have so I presume it is dead too! I replaced the suspect PSU and I pulled the Shuttle AN35N 400 Ultra Motherboard and Athlon Mobile XP3000+ from my library computer and now Guardian lives again. I upgraded the RAM to 768MB from 512MB. The Household File server lives again!
[*]Tyan S2462 Dual Processor Motherboard - a replacement for the S2460 in Guardian that was DOA.
[*]Bikini (Shuttle AN35N 400 Ultra) just crashed and failed to post ever again. Bad Caps in a cheap PSU! A lesson learned. Bikini may yet live again but that is a down the road thing as this computer is a pure cruncher.
[*]TheBrain (Shuttle MN31N) I built three identical micro ATX based computers in 2002. Trinity, Pinky, and TheBrain, to replace 3 aging dual pentium 200 computers. All had Shuttle MN31N motherboards, Athlon XP2400+, and 2 x 256MB Golden Dragon Value PC3200. TheBrain has experienced random Lock-Ups (usually about every other month) for about the last year. I finally decided it was time to upgrade it to Windows 2003 Server. I saved everything off ... installed the new OS and TheBrain turned up lame. Memory tests confirm that something is amiss with the motherboard or processor or onboard video memory. There are very apparent anomalies in the video during the memory testing. I'll pick the motherboard as the failed component. The PSU tests OK but I have never been pleased with the voltages from it so it will be history too. At this point, a replacement PSU, motherboard, and processor are on the way. This will be the first of the 3 to be upgraded to an Athlon 64 3200+ Venus. I'll test the XP2400+ and, if OK, will replace a XP2000+ in another computer ... a fairly nice upgrade for both systems. Alas, an Opteron or Dual Core was not in the budget this time
[*]Smoke (Shuttle MN31N) Yet another micro ATX based computer built in 2002. This has a transparent smoke gray plexiglass case. Apparently, the video RAM has failed in the onboard video. I added an AGP card and all seems well for the moment.
[*]Ghost (ancient ASUS Duron system) Another motherboard has apparently failed. Again, this computer may live again but not in the short term as it was only a cruncher.
[*]At the strong urging of the wife, I have divested several other sub 1GHz computers. The local horde is now less than a dozen processors/cores.
I am beginning to formulate a few theories from these failures:
[*]Processors and/or memory may tolerate long term overclocking but, perhaps, motherboard chipsets may not be as durable or tolerant. I don't believe DC projects bear on this one way or another.
[*]Cheap and/or generic power supplies are a waste. I will bite the bullet and purchase only Active PFC power supplies in the future. If for no other reason, the high efficiency of these PSU's should have the effect of reducing electricity consumption. Because of all the advanced features of these PSU's and the high cost, these power supplies seem to be built with better components to go the distance. MTBF/MTF for these PSU's is almost double that of the low cost generics. A bonus is that most of these advanced PSU's are genuinely QUIET.
[*]I keep my systems running cool. I routinely clean out the dust that seems to accumulate. I absolutely believe that if I did not keep my systems cool and clean, the failure rate would be unacceptable.
[*]Bad Caps in PSU's remain a problem even for some presumed high end PSU manufacturers like Antec. I don't find any visible evidence of Bad Caps on these failed mothboards but I have repeatedly seen cheap and generic PSU's with Bad Caps.
So much for the ramblings of a deranged cruncher.
mondo
Facebook
Twitter