I've got a flakey Windows box. Pretty sure this is screwed up at the hardware level: I get intermittent but frequent BSOD's after waking from S3 suspend, and now I've seen an instance where it came up and couldn't find the boot media the first time, but then found it after a soft reboot.
So with that second piece of information, it seems definitely hardware. But the most likely culprit? Motherboard or hard drive?
If that doesn't cure it, run a good memory checker.
In fact, run the mem-check first.
I'm a bit worried that reflashing a bios on already flakey hardware could turn a semi-functioning machine into a dead pile of circuitry if the process is interrupted.
Good point about the memory checker, though.
(This thing ran fine for years with the BIOS that's in there.)
2. set your BIOS to boot from your CD drive.
3. insert the Linux Live CD.
4. tell WinBlowz to shutdown/restart
Your box will boot up using the Linux Live CD, which will NOT do ANYTHING to your existing hard disk(s).
Let the machine run. Use the hell out of Linux - there's a web browser and a few other fun toys. See if the box frotzes on you. If it runs for more than a day, the problem is in your WinBlowz installation (probably a bad driver or 3... oh, sorry, Mickey$lop calls them "dll files" - same diff).
If the machine frotzes under Linux, and assuming you ran the mem-check and the RAM tested good, then it's either the BIOS corrupted or the CPU or the motherboard itself. Same fix in any of those cases - replace the motherboard and CPU. About $150 for something decent (but not high-end).
Either way, at least you'll have some fun with Linux!!
I do not believe in memcheck. In the last 10 years, no box ever failed because of failing memories. I saw defective memory chips simply being ignored by the system, but no harm done. Defective memory being harmful to a machine is from the pre-DDR1 time, in my opinion.
I do sincerely and ultimately believe in defective capacitors. If your board is older than 4 years, even significantly older maybe, have a look at the capacitors. If any of them has a swollen top, it is likely that the machine will continue to fail. http://en.wikipedia.org/wiki/Capacitor_plague some sample images included. Most often they are between the cpu socket and the back panel. They die because they are either badly manufactured and/or because heat dries out the electrolytes.
- Waking from suspending to ram/hibernation fails
- USB ports do not give enough power to attached devices
- Other effects in scenarios where more power than usually is required
The simple reason is, that the capacitors play a major role in delivering the required voltages and amperes for operation. My previous desktop board failed to resume from sleep, the attached 5port switch which was powered over ps2 port failed to operate. This board now serves as my home server without any problem, I do not need to power usb/ps2 devices and it is always on. It might one day stop booting at all, but currently it is fine. Some of the capacitors showed swollen tops, btw.
I've got a flakey Windows box. Pretty sure this is screwed up at the
This is, of course, by definition.
Right. But I depend on Windows to run Lightroom, these days. (No, there is no open source product that even comes remotely close to replacing Lightroom.)
And the occasional game.
Anyway, this is weird:
(1) Windows generates a Minidump on BSOD, by default.
(2) I see a minidump in C:\Windows\Minidump or wherever, from early September-ish, before this problem began happening in earnest.
(3) But now the system is configured for a full crash dump (C:
\Windows\Memory.dmp I believe). The weird part is that I didn't change the configuration since September, for sure. And probably not ever.
(4) The full memory dumps were failing to appear, so I switched it back to generate Minidumps. And finally got one. It fingered the Intel Matrix Storage driver, so I updated that. We'll see if that helps.
I ran Microsoft's memory checker in extended mode overnight, and it says "no problemo, boss."
(I use Windows at home, too... although I've been more fortunate, it seems).
I still think this is probably a hardware issue, otherwise this would be in Microsoft Bashing). But I'll run with the updated(?) driver for a while just in case.
I do sincerely and ultimately believe in defective capacitors. If
your board is older than 4 years, even significantly older maybe,
Probably. This board is a P55 so I guess it's from 2009.
Yep, that fix didn't last long.
Yeah, my mom has a Dell from that timeframe -- 2001-2 or so -- that shipped with Windows XP. I finally had to advise her to retire it because XP was no longer supported and Vista-or-newer would not be a good tenant on that hardware.
That thing was a tank - all my other machines since then have failed or needed upgrading.
I just ordered a stack of Haswell stuff -- almost an entirely new machine, I expect to re-use the video card, power supply, Blu-Ray, and Windows license; all the rest might as well be upgraded.
Melvin does not wish to upgrade.
This was true when I last upgraded his machine. He did not wish to part with that one, either. Nor the one before that.
He has only ever owned three machines. His current machine has a single core running at 1 Ghz, back when such things were how CPUs were measured.
I am *not* counting the absolute ton of machines, routers, hubs, switches (smart and otherwise), RAS boxes, etc. etc. etc. that were a part of the NetK2NE ISP operation. Some of that stuff survives to this day in my storage areas up here on the mountain. Most has since been sold, given to friends, etc.
I am also not counting the TRS-80 Model I (1977) or the Apple II (1978) or the Commodore 64/128 (1983/1985). They couldn't run anything that we run today (except for Citadel - I ported that to the Commodore 64 and then to the 128 before "falling into" an IBM PC AT clone in 1987, never looking back and beginning the Citadel:K2NE project).
Enough from me!
I thought the C= 128 could run the original Citadel...
I have seen lots of old boxes surviving more than 10 years. Most of the time, the harddisk or power supply become faulty.
But of the failing ones, most died of capacitor issues, between 5 and 10, do not recall exact amount. Two died of dc-to-dc converter overheating (distinctly visible darker area around one of them), one because the PSU connecter burned out on one pin (that one looked like "phew, this could have been worse, like starting a real fire"). Only one died of uncertain cause.
Oh, and I killed some accidentally, because I slipped with a screw driver when I tried to mount an AM2 heatsink for the CPU. Luckily, it was replaced it. Another one had a memory slot which I broke in half while trying to insert a memory module with one hand, while not looking into the case. Actually, it "lived", but I buried it.
2014-10-09 14:28 from fleeb
I thought the C= 128 could run the original Citadel...
Nope - it ran that same stupid Commodore BASIC that the C64 used, with a few extra commands.
Citadel-128 was nothing more than a port of my Citadel-64 software, with added support for external hard drives (I ran my BBS w/ a 20 meg hard drive - there were only 2 C128 BBSs using hard-drives. Me and the other one.).