Comments on: Smoke from the PSU http://etbe.coker.com.au/2008/05/14/smoke-from-psu/ Linux, politics, and other interesting things Wed, 08 Feb 2012 17:45:05 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: etbe http://etbe.coker.com.au/2008/05/14/smoke-from-psu/comment-page-1/#comment-14033 etbe Sun, 18 May 2008 00:12:55 +0000 http://etbe.coker.com.au/?p=582#comment-14033 martin: True, spinning up the hard disk when it's cold can take a significant amount of energy (more than spinning up a disk that has been parked for a while due to viscosity in the lubricating oil). This problem did occur on a cold morning after the machine had been turned off overnight. Also when it initially starts it has to charge all the capacitors that surround the CPU which may impact the current draw. It seems quite unlikely that a machine with a default configuration (no extra PCI cards and only one hard disk) when starting will use as much power as a fully configured machine running under load (think about having two hard drives and a CD-ROM drive doing seeks, three PCI cards doing stuff, and 100% CPU use). So if a machine dies under load when starting with a default configuration it seems most likely that it would also have died in normal operation if it had the maximum set of extras added to it. martin: True, spinning up the hard disk when it’s cold can take a significant amount of energy (more than spinning up a disk that has been parked for a while due to viscosity in the lubricating oil). This problem did occur on a cold morning after the machine had been turned off overnight. Also when it initially starts it has to charge all the capacitors that surround the CPU which may impact the current draw.

It seems quite unlikely that a machine with a default configuration (no extra PCI cards and only one hard disk) when starting will use as much power as a fully configured machine running under load (think about having two hard drives and a CD-ROM drive doing seeks, three PCI cards doing stuff, and 100% CPU use). So if a machine dies under load when starting with a default configuration it seems most likely that it would also have died in normal operation if it had the maximum set of extras added to it.

]]> By: martin http://etbe.coker.com.au/2008/05/14/smoke-from-psu/comment-page-1/#comment-14030 martin Sat, 17 May 2008 11:28:43 +0000 http://etbe.coker.com.au/?p=582#comment-14030 I'd been told that PSUs get most stress when started. So maybe it just broke because of that and not because of anything else beeing faulty. I’d been told that PSUs get most stress when started. So maybe it just broke because of that and not because of anything else beeing faulty.

]]> By: Paul Johnson http://etbe.coker.com.au/2008/05/14/smoke-from-psu/comment-page-1/#comment-13952 Paul Johnson Wed, 14 May 2008 03:26:56 +0000 http://etbe.coker.com.au/?p=582#comment-13952 Stupid me, I read the headline as "smoke from <a href="http://en.wikipedia.org/wiki/Portland_State_University" rel="nofollow">PSU</a>" and clicked on it in aKregator wondering "OMGWTF is going on in downtown <a href="http://en.wikipedia.org/wiki/Portland,_OR" rel="nofollow">Portland</a>? <a href="http://blog.oregonlive.com/breakingnews/2008/05/homeless_protest_impasse_conti.html" rel="nofollow">Did the homeless revolt or something</a>? Why haven't we fed them to the hungry yet? Stupid me, I read the headline as “smoke from PSU” and clicked on it in aKregator wondering “OMGWTF is going on in downtown Portland? Did the homeless revolt or something? Why haven’t we fed them to the hungry yet?

]]>