You are not logged in.
I've been experiencing the wierdest set of hardware (I think) problems recently. I've tried hard to find a solution, but I'm just lost. It's a bit long, but there's a lot of info.
First, it takes a few tries for my computer to start. The VGA won't post, it'll try to start, but it'll shut down and try again. Sometimes it'll continue without VGA, but not often. It does this with any amount of hardware attached to it -- even with just the CPU and motherboard with the onboard graphics chip, and with two PSUs. After it gets a successful VGA post, the computer continues to work fine.
Until recently. A kernel update from 3.2 to 3.3 brought crashes into the picture. Could be completely unrelated to the kernel update, but it consistantly started happening after rebooting to 3.3. One of three things would happen: X would crash and restart, everything would freeze and require a cold restart, or it would take a few seconds to recover. I noticed that the more video I was watching (ex higher res, full screen vs low res in a window) would trigger crashes. A 1080p flash video would bring the system down in 5 minutes, as would any HD video in mplayer or VLC. VirtualBox also triggered eed crashes pretty often while working with Autodesk Inventor inside XP.
Now, here's where things get real bad. Last night I ran a -Syu and nervously updated my nvidia drivers. After rebooting, my system is unusable. My KDE desktop takes about 40 seconds to load, and then I can't even close any windows, open anything. Black boxes show up around / in windows, graphics are fragmented, things don't load. Totally unusable. Switching to TTY's gives a blank screen.
Upon rebooting to windows (which isn't a good assessment of the situation because it's installed on a really slow old IDE HDD) it worked for a bit at a much slower speed than normal, eventually bluescreening and rebooting. I'm working on burning a liveCD with an X interface now to see if I can trigger crashes from there.
I rebooted to an arch liveCD last night and ran a long test ith smartmonctl on my main HDD that Arch resides on. No problems.
Another interesting thing: upon taking out my Nvidia 8800GTX to see if the onboard graphics chipset (Nvidia 8200) would fare any better, I couldn't get any picture at all. The top 1/3rd or so had a picture, but the rest simply flickered with both DVI and VGA outputs on the mobo.
I'll check the logs to see if I can find anything interesting when I get that liveCD up.
Offline
IMO the easiest hardware failure to diagnose is RAM. Either try with one stick of RAM at a time, or just run memtest. If I were you I'd try the former first, then run memtest on any sticks that seem to be good. Then add or swap in one of the questionable sticks and run memtest again until all the sticks have been covered. If your mobo doesn't have RAM slot population constraints, you should try different slots for any RAM that seems to be questionable too.
But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner
Offline
I'm running memtest right now.
Offline
I suspect:
overheating - check the cooling fan on the CPU or graphics card, clean the dust with a vacuum cleaner, put some machine oil in it if possible, or replace it
faulty (swollen) electrolytic capacitors on graphics card or motherboard - replace them with the same farrad values, if you know electronics
bad/corrupt BIOS (CMOS) settings - reset them by turning off PC, removing battery, wait one minute, put it back on, start PC. enter BIOS, set all values as you need, save and reboot, see if it works as it should
faulty RAM - remove one of the sticks, boot PC, check if it works, swap the memory sticks, repeat until you've tried them all; test memory with memtest afterwards
faulty VGA/DVI/HDMI cable - reconnect them if loose; if an VGA cable, keep it away from power cables, speakers or any other device that creates magnetic fields (cellphones too); replace the cable with a new one, if possible
faulty monitor - try with another one that works
buggy drivers/kernel - update the system or search the forums for anything similar to your problem
Did you find the culprit?
Last edited by bjornoslav (2012-04-16 19:23:51)
asus ux303la, core i5@1.6ghz, 8 gb ram, 500gb hdd, hd4400 gpu, crux x64 with openbox
Offline
Well, I can't take apart my PSU and look for bad capacitors (still under warranty) but I can try my other one. I'm going to play around with it tonight and see if I can get somewhere with it since the problem's recently gotten worse. I can't see any bad capacitors on the mobo, and I looked very carefully.
It's not overheating. Everything is of proper temperature.
I've used two monitors with both DVI and VGA cables.
The CMOS battery was replaced a few months ago (after the VGA-post problem started) and I tried reseting the BIOs to 'optimal settings' with no changes.
Memtest has been running for a while. 8 passes with no errors.
I'll play around with the PSU and the memory tonight.
Offline
System is fully updated. Still doesn't function well under heavy graphics load. What's interesting is that KDE doesn't work at all, but XFCE can handle pretty much anything, at some speed, for some time, before crashing. The system functions like normal if I work simply from the terminal, but as soon as heavier graphics get involved, it slows down.
Running mprime, at 100% cpu through four tests made it seem like my CPU was fine. No slowdown while running mprime. Memtest86+ ran four about 8 hours and reported 10 passed tests, so it doesn't sound like it's the ram (sometimes it crashes before I would even get to slim, which I have since disabled.)
I changed my PSUs again. Still having the exact same problems, no differences. This PSU worked just a few months ago in another computer of a little more power with no issues, and has been sitting in a box since. The other one is only a month or so old.
So what do you guys think? Graphics card? Motherboard?
Offline