Here's a mystery for the masters

[gso94190]

I figure I'd throw my problem out for all you guys to chew on. I'm all out of ideas about whats going on.

Ok, Heres my situation. I put my computer together around last christmas.
specs:

MSI k8n neo
AMD ath 64 3400+
1 gig kingston ram
BFG geforce 6800 nu (now GT, i'll explain later)
coolmax 400w p.s.
win xp

I had no probs for a long time. I played tons of World of Warcraft with no probs. I got half life 2 but did not play it as much as WoW. One day, I was playing half life 2 and out of the blue, my whole system shutdown. no warning no nothing. It had never done that before. So I toggled the power supply swith on and off to reset it as I thought it had been overextended. The comp started right back up and I just returned playing WoW. Time passed with out problems. I got busier and stopped playing games for a while.

Recently, one of my friends started playing HL2 on my system again. The random crashes started happening again. He'd play for about an hour and the system would go down. On one occasion after a crash, we couldn't get the pc to respond. I cracked open the case and started swapping parts to identify the prob. First, I swapped out the Power supply. No luck. Then I checked out my video card. Low and behold both fans were visably cracked. One of the two fans was cracked in half. Hard for a fan to move air when its broken. I immediately assumed that was the problem and sent it off that day to get RMA'd. Just yesterday I got the card back from BFG. They sent me a 6800 GT instead of the NU. Very nice customer service. I poped the bad boy in, cleared and reloaded the drivers (77.77) and turned on HL2 to test it...and Yup, it happened again. I was still in the main menu screen of the game (like actually 5 min. after the game loaded) and the whole comp went down w/ no warning.

Has anyone had this happen? any suggestions?

Its only happened when playing HF2. WoW, Quake 3, UT2004, KOTOR all have no probs for hours. The video card has its own dedicated line from the power supply, no other connections on that line.

Please, if anyone can think of something i've overlooked, i'd be very grateful.

 

[mzkhadir]

What does the event viewer tell you, if you look into the application tab ?

 

[gso94190]

I checked the event viewer and nothing of consequence is listed.
thx for the quick reply

 

[mzkhadir]

so you don't even get a blue screen when you are playing the game ?
when you computer goes down, does it make like a clicking sound and then shuts down ?

 

[gso94190]

nope, I don't get any warning. Like I said I was just on the Main Menu of the game looking in the options. Then all of a sudden, the lights on my tower go off, the fans shutdown and the monitor clicks off (loses signal from the tower). Its just as if I had told it to shut down. Thats why I thought it was the Power supply. Like the Video card was asking for too much juice. But then why would it run for an hour or more before while playing the game, and now crap out after just sitting in the main menu for a couple minutes... I don't know.

 

[gso94190]

Do you think it could be the heat sink fan on the processor? I was reading on another board and came across that. I have the stock heatsink that came with my 3400+ installed. Are there typically problems with stock AMD heatsinks? Just a thought...

 

[mzkhadir]

Originally posted by: gso94190
Do you think it could be the heat sink fan on the processor? I was reading on another board and came across that. I have the stock heatsink that came with my 3400+ installed. Are there typically problems with stock AMD heatsinks? Just a thought...

if you never had a problem before with the heatshink and this started because of something else like NU videocard.

 

[mzkhadir]

can you try to see if you can update you video card drivers to 78.01 and see if the problem goes away ?

You can also check this thread about random 6800 freezing ?

http://www.nvnews.net/vbulletin/showthr...=32759&highlight=77.77+driver+problems

you will need to create an account.

 

[mountcarlmore]

you should track your cpu's temp by creating a log in whatever monitoring program you use, so you could see if the comp goes into thermal protection mode. hl2 is one of the more cpu demanding games out there.

 

[DAPUNISHER]

Monitor all temps&voltage rails under heavy load for starters. What is the +12v on that 400w RAIDMAX, and what PSU did you try when swapping? I know everything else runs fine, but if you have turned off auto-restart and it is a hard shutdown, that is always the place to start IMO, the PSU, then heat.

 

[monster64]

Well if it was a PSU issue, it would make sence that now with the 6800 GT it crashes alot faster as a 6800 GT uses ALOT more power than a 6800 nu. Did you check your ram?

 

[gso94190]

Well, I think we've narrowed it down to a heat or power issue. How can you make a log of CPU temps and rail voltages? Does windows have a program that can track this info for you or is there some utility that you guys could direct me to. Thanks.

 

[Parkre]

MotherBoard Moniter 5 might help you

 

[Stonesoldier]

OK here is a shot in the dark so to speak
check for bad capacitors (leaking\bulged\ect ect )

known problem for msi boards
look for a brownish residue around your caps


bad caps can cause all kinds of problems
might look like memory or PSU or video or drivers ect ect
your gonna have to look closely
so get a good light for it

 

[gso94190]

Wow, Mother board monitor is great! I never knew about it before.

I ran an interval log in the background while I booted up HL2. I told it to make a log every 5 seconds. After about a minute in the main menu screen of HL2, my comp pooped out (as expected). I turned my comp back on and found out some interesting info...

This is where I wish I knew more about the workings of my comp. I consider myself computer semi-savvy. For example I know the +12 V rail on the psu sends power to the video card (i think...) but I don't know the acceptable varience in voltage. Also, I know that if the proc gets too hot that there is a "breaker" so to speak that shuts the comp down to protect it. But, I don't know where that is. Stuff like that. So, I'm just going to put down all the data from the log that I think was suspious so that more intelligent eyes (yours) might be able to desifer it.

Ok, first, cpu temp. when I started the log, my cpu temp was at 48 degrees celcius. After starting HF2, the temp gradually increased until it reached 55 degrees. Mobo monitor only recorded 2 logs (10 seconds) of this cpu temp before the comp shut off.

Now in Mobo monitor there are 3 fan readouts. On the log, both of the Fan 1 and Fan 3 readings read 0 rpm's. I assume that the Fan 2 reading refers to the fan on my proc but I'm not sure. I don't know how to find out which fan Mobo moniter assigns to 1, 2, or 3. Interestingly, (to my novice eyes anyways) as the temp of the proc increased, the rpm's of the one fan that mobo monitor was recording didn't really change. At 48 degrees it was 3515 rpm's and at 55 degrees it actually droped to 3479 rpm's. But like I said, i'm not sure this is the heatsink on my proc. I also have an exhaust fan in the back of my case. One of the 2 fans isn't being recorded, but I assume the cpu heatsink is the one being recorded on mobo monitor. Whichever one it is, Mobo monitor shows 3500 rpm's to be in the red in the "dashboard" reading.

Ok, on to power. I'm not really sure what the "core 0" is monitoring, but at the beginning of my little HF2 experiment the reading was 2.44v. It ended up at 2.44, but fluctuated as low as 2.37v and as high as 2.51v. For most of the time though it was around 2.44v. The thing that caught my attention though was that in Mobo monitor the "dashboard" reading shows 2.44 to be almost in the red...

The 12+ rail stayed fairly consistant at 12.10v. It dropped once in a while to 11.98v and only a couple of times to 11.92v. I'm not sure if this is acceptable varience or not.

Well, that looks like all the pertinent info to me. everything else seems to be "normal". To me it looks like the stock amd heatsink isn't pulling its weight, but that core 0 voltage also might be a problem.

Thx for helping me with this. I'm determined to get to the bottom of this.

 

[gso94190]

This is really frustrating.

I ran a mobo monitor logs on quake 3 and star wars: knights of the old republic just like I did on HL2. The results were not consistent.

Quake 3 did not crash. the max cpu temp while it was running was 53 degrees celcius, but the fan was going 4623 rpm's. that's like 1000 rpm's faster than when it was at 55 degrees playing HL2. Meanwhile the core 0 rail (still need to find out what that is) stayed fairly steady at 2.44v.

Knights of the old republic did crash. The max temp before the crash while playing this was only 45 degrees. The fan was also doing much better than during the HL2 test: 4623 rpm's. The weird thing was that the core 0 rail, which normally stays at about 2.44v (which mobo moniter says is high), was down around 1.39v.

Just more info to add to the picture. I wish there was a pattern...

 

[gso94190]

no ideas huh... well, I flashed my bios to 1.90 (msi k8n neo) thinking it would update something and the problem keeps happening. It even happens when surfing the web sometimes...

Oh well, no more self bumps for me. I appreciate all your imput, if no one else has any ideas, i'll let this thread go to the thread graveyard.

 

[Parkre]

Hey , I am glad you like MB5. CPU 0 should be your first core (you only have 1) and CPU 1 is the second core.

2.44 sounds pretty high, but i am only about 10% sure of that. 3400 Sempron?? (i didnot see 64 3400 athlons at newegg)

As for any suggestion...well if it voltages try another PS.
IF it continues
find a replacement MB (Fresh Install too)

oh, and your 12 rail is just fine