I can usually solve my own conundrums, but I need some assistance.

I recently changed my motherboard and graphics card, and also reinstalled Windows with SP2. Graphic drivers are up-to-date.

The problem is, every so often (at least once during computer usage of 1 hour or more), my computer beeps and locks up. I've also noticed it seems to do this most often when starting or ending a program (I could be starting LFS, or it could be as simple as closing my browser window). The exact way in wich it locks is always the same:

1. Everything on the screen becomes unresponsive, but the mouse still moves.
2. After clicking a few times the computer beeps, the mouse freezes, and the whole computer seems to be locked (task manager cannot be summoned)
3. As a note, it has done this with music playing before, and somehow the music kept playing even though everything seemed to be hard-locked.

Lastly, I have run adware checkers and virus detectors which have shown nothing. So there's your mystery. Any clues? I appreciate any help.

one of several reasons:

1. faulty ram. if you have more than 1 ram stick, try using only one at a time, with each of the sticks. if you have a faulty one you'll find it.

2. CPU temperature/GFX card temperature. for GFX you can usually use the driver to get the info (right-click-desktop ->settings ->advanced->temp monitor). for CPU monitor there are many tools, but when your pc crash, reboot and you can usually enter the bios and see a cpu temp monitor. above 60 (celsius) more or less should have your attention. if that's the case, remount the cpu fan.

there are many more reasons, but i think these are the most common ones.

good luck.

Could be your PSU doesn't have the power to run the new card.

What card is it, and what watt is your PSU and who by?

Does the comp itself beep, or does Windows beep?

If it's the comp it's probably hardware related. On the other hand, Windows should NEVER freeze because of a software problem that doesn't shut down hardware (that's why we have those blue screens).

I can only assume that there's something in your BIOS settings that turns off something it shouldn't turn off because of some threshold being reached. Usually this is temperature related. Even if it's not hardware related in the end, this is the best place to start experimenting, since there's only so much hardware in the comp.

Try the following:
Take out everything except the proc, and some memory (only one strip, not all if you have more)... Run the comp, and wait and see what happens. Then one by one start reinstalling hardware, until the beeping occurs again. (of course an old PCI videocard would be useful if you don't want the comp to seriously stress out because it wants to draw out the pictures in thin air.)

It's gonna take time, but at least you're ruling out hardware as the problem.

Also, check the proc for any bent pins (there's about 187 of them, so that should keep you busy)... Since you changed the MB, I'm assuming you inserted the proc. Could be you messed that part up just a tad. Or applying the cooling block might have slightly cracked the core (depending on what proc you got). <-- Yes, I've spent days looking for this very problem of a tiny crack in the core with my PC stopping dead in its tracks after about 20 minutes of uptime. Since this is a pretty dangerous operation when the cooling is already in place (you could crack the core by removing it again), I suggest you try this last.

If none of these solve your problem, get back to us, and we'll start examining your system.

(BTW, Windows not responding and the mouse still moving is not so strange... The way the mouse works on a computer is in many ways the stupidest design flaw ever... They made sure that the mouse's I/O will not interfere with anything else, so unless the processor itself is in a deadlock, and windows hasn't killed all its threads (which is basically when you get a bluescreen), the mousepointer WILL move.)

Quote :

What card is it, and what watt is your PSU and who by?

My GPU is an EVGA GeForce 7800 GT PCI-E 256MB DDR3. Also, my power supply is an offbrand and is to spec with the power usage for the card (which I believe is 400W). I would put this as last on the list of possibilites since I had been using the same power supply for my previous card, which also required the external power.

Quote :

Does the comp itself beep, or does Windows beep?

The internal speaker beeps, not windows.



Anyway, thanks so far. Now I have some stuff I could try to solve the problem.

The first thing I will test is the memory, as I've had memory problems in the past. I'll update when I have a chance to really spend some time on the issue.

Also, it might be worthy to note that it has yet to lock while playing a game. I can play LFS for hours without anything happening, but it can easily freeze a couple minutes or seconds after I exit the program.

400W isnt enough for that card.

A while ago the cooler for my motherboard died out, and whenever i started up my computer a loud siren sound would come on, and my computer would shut down. Do you happen to have anything that will tell you the heat of your computer? Id say its a cooler problem, but their are so many possibilites. I reccomend having someone "trained" in the business come over and take a look at it.

Well I think if he can play LFS for hours then this pretty much rules out a heat problem. You really gotta check for faulty hardware parts, removing one (non critical) thing at a time, to see what causes the problem.

Quote from TaiFong :

My GPU is an EVGA GeForce 7800 GT PCI-E 256MB DDR3. Also, my power supply is an offbrand and is to spec with the power usage for the card (which I believe is 400W). I would put this as last on the list of possibilites since I had been using the same power supply for my previous card, which also required the external power.

Hate to break it to you. But it's not. If it's unbranded PSU then it uses an impossable rating system where it's only 400w at 25c. You'll never get it to run that cold so it'll not be 400w. I'd say that's your problem to be honest. Just because it ran an older card, doesn't mean it'll run a 7800.

"Single card requires min. 400W PSU"

And that'd be from a good brand PSU. My tip to you. Save some pennies and buy a better PSU, or if you want to test it, talk to someone who has a good PSU and ask if you can loan it for a few minutes.

GeForce 7800 GT PCI-E consumes 97 Watts. You should be fine unless you have several hdd's, fans and optical drives.

PSU Calculator: http://extreme.outervision.com/index.jsp

I used that handy PSU calculator and came up with 320w minimum, and I was generous with a few things. I think I'm OK in that department unless the PSU is defective or the calculator is inaccurate.

That will be aimed at a high end PSU. No PSU runs at 25c, which is why pretty much all PSU's are rated at. Your PSU wont be pumping out 400watt. That program said I'd need a 327watt PSU, but I know if I tried one it wouldn't power my system. Plus if you use a PSU close to the minimum you need it'll be running at full load all the time and use more power.

See if you know someone who has a good PSU, not unbranded (around 500watt) which you can loan to see if it can power your rig. If it does then you need a new PSU, if not then you need to find something else. But if you went to any PC Tech forum and said you had a unbranded PSU @ 400watt and your PC shutdown with the specs you have. 8/10 people will say get a new PSU.

Based on what has been mentioned up to this point, I'd think the most likely cause is memory or an motherbaord issue... (possibly an incompatability issue, check your supplier for any known issues)

I still wouldn't discount the PSU as being the issue also (like many are indicating) often with cheap PSU's they will have a very generous max rating, which many can not even come close to achieving reliability at that rating - I've seen 450W PSU for example that have the same rated components as a good quality 300W PSU so the 450W is absolute rubbish in that cercumstance The point is it would still be worth swaping the PSU with a known good brand to TEST it, even if for only a day or few hours...

With the memory (or any card with electrical connectors) it can quite often be as simple as tarnished connectors from a finger print, or even a human hair falling accross the M/B connector Get your PC under bright light, so you will spot hairs or any other hard to see foriegn materials. Use compressed air (can get from local electrical store in a can) to blow out the motherboard connectors before you install the memory.
With the memory take an eraser (rubber) and while holding the memory on the edge of a table so that the connector portion of the memory stick is supported by the edge of the table use the eraser to clean the connector surfaces (hey it works :P been doing it for 15+ years) that will get rid of any tarnish or contamination on the metal connector surfaces. Another option which I use these days instead of an eraser is Electrical Contact Cleaner and a lint free cloth, which can also be obtained from your local electronics supply store.

Hope that gives you a few more ideas and best of luck It can be frustrating trying to identify these types of problems sometimes, especially without all the extra components and equipement that a service centre has In a service centre they probably could identify the problem quite fast, if you can get friendly with some tech guys at your local PC shop it would make your life alot easier

[Edit] I'd have to agree with P5ychom4n there, I've worked as a PC tech before and if you brought your PC in discribing those symptoms the first thing I would have done is swap the PSU with a known good unit and test it, next would be swap the memory for known good memory, then if still getting problems start to look at heat issues, probably with CPU or the mortherboard multifunction controller chip, which gets quite hot these days (the controller chip will most likely have a heat sink attached and if the layout of components and cables, etc. in the case is not ideal it want be getting sufficient cooling, and if that heat sink is located near the memory it could even cause the memory to malfunction from heat issues) In alot of the newest PC's (especially P4 systems) heat is a major issue and good cooling design (air flow within the system) is very important. If still a problem at that point then I'd check for incompatibility issues, I've bought new M/B and video card from the same supplier in the past that had compatibility issues, so does happen... [/edit]

[Edit2] Something else just occured to me, being a new M/B it is likely that in the BIOS it has event logging... So if your familiar with BIOS setup you could check there for any clues as to whats happening (you probably have to turn it on then run the system to get it to log something as I think its set to off by default most times IIRC) [/edit2]

I knew this was the right place to ask this question. So many friendly, knowledgeable people.

Anyhow, I do appreciate the advice, and I definitely will be trying all options (even the PSU ). I used to be involved with buying and selling wholesale computer parts and had TONS on PSUs lying around, but they are all gone now! So it might be a little harder to get my hands on one, but I'll try.

Also, I prefer to do my own computer work, even if I have to learn a thing or two in the process. I'd sooner slit my wrists than spend hard earned money and take it to some "computer expert" who probably knows less than me.

im 99% sure its the psu.. Generic psu's are never near their rating in watts or amps.

EDIT: also 97Watts is wrong for a 7800GT.. maybe idle.. Load they draw around 250watts.

Quote from TaiFong :

Also, I prefer to do my own computer work, even if I have to learn a thing or two in the process. I'd sooner slit my wrists than spend hard earned money and take it to some "computer expert" who probably knows less than me.

I wasn't talking about paying for it if you get friendly with some guys (gift of the gab and maybe an icentive like carton of beer ) you may be suprised how much you can get for free and probably can learn in the process anyway

Quote from Roadie :

EDIT: also 97Watts is wrong for a 7800GT.. maybe idle.. Load they draw around 250watts.

See the graph at the bottom of this page. Note that the numbers are total system power consumption.

http://www.anandtech.com/video/showdoc.aspx?i=2636&p=5

Quote from Forbin :

See the graph at the bottom of this page. Note that the numbers are total system power consumption.

"This page"? Link?

Quote from Roadie :

also 97Watts is wrong for a 7800GT.. maybe idle.. Load they draw around 250watts.

Ok. So an SLI setup should have around 800W PSU then I guess. How much for my 850XT then?

Quote from AndroidXP :

"This page"? Link?

Silly me... :doh:

http://www.anandtech.com/video/showdoc.aspx?i=2636&p=5

The system specs are on the next page. For those too lazy to look, the system includes an AMD FX-55 2.6GHz.

You could be smart and put the old video card in to see if you have the same problem

Well I found the problem but I can't fix it. It seems anytime a Java application is used and then is closed, it freezes the computer. I accidentally had 2 different version installed at the same time. I uninstalled the newer version, but it will not let me uninstall the older version. The error says:

"You already have this version of JRE installed. Please uninstall the product using your add/remove programs utility before reinstalling."

I get this message even though I AM trying to uninstall using add/remove programs. So, then I tried deleting all registry files associated with Java and any folders that pertained to it. After that, I reinstalled Java, but I still get the same freezing problem after closing out Java applications. So, I suspect it has something to do with that stupid old version of Java installed in some ungodly location that I cannot extricate it from. Anyone have an idea?

Make sure to check your PATH variable so it doesn't point to the old Java installation. I think I once encountered the problem of an unupdateable Java when trying to install some Oracle stuff, and IIRC that was the solution.

You can also check the currently used version by opening the command window and typing "java -version". Should there still be some old version, use the above mentioned method to fix it.

PS: to access environment variables go to: WinKey+Pause (=System Properties) > Advanced > Environment Variables

Android, I just installed Java 1.5.0_05. Then I check the 'CLASSPATH' and 'Path' variables. They were pointed to the wrong direction (different version). I changed that, but I stil have some concern. This is what both variables look like:

CLASSPATH: C:\Program Files\Java\jre1.5.0_05\lib\ext\QTJava.zip

Path: %SystemRoot%\system32;%SystemRoot%;%SystemRoot%\system32\WBEM;C:\Program Files\QuickTime\QTSystem\

Firstly, there is no such thing as QTJava.zip in that same place of my current Java installation

Secondly, I don't know what to make of the Path. Is Java supposed to be associated with Quicktime like that?

Anyone have any further input? I have had no luck fixing this, and I kind of want to avoid a clean Windows install.