Stability Test, Best Methods?

[Codesmith]

I just built a new system for a friend and it keeps locking up (video freezes, soft power off won't work) after 1-6 hours. Unfortunately the infrequency of the lockup is making it difficult to troubleshoot.

Athlon 64 2800, Asus K8VSE Deluxe, Antec True 430 Watt, 512 MB Corsair Value PC3200, Zalman CPNS700A AlCu Cooler.

My current testing methodology is to run memtest 86 for at least 12 hours to test the memory then run SiSoft Sandra's Burn in Wizard (CPU test) to keep the system running at 100%.

I am concerned that this methodology might not be accurate enough. I am also would like a method that yields quicker results.

So what is the most efficient and comprehensive plan for testing complete system stability?

What programs should I run, for how long and in what order?

Also I sometimes use a Live Linux CD to test systems, anyone know if there are any test programs included with KNOPPIX?














<P ID="edit"><FONT SIZE=-1><EM>Edited by Codesmith on 05/05/04 02:55 PM.</EM></FONT></P>

 

[Coyote]

I use memtest86 first, but only for an hour unless I'm OC the ram. Then Prime95 (torture test) for the cpu. Since I only do this when OCing I go for 12-24 hour. I also run Motherboard Monitor 5 to check voltage and temps.

What's a AMD 65 2800? Do you mean AMD 64?

I don't think there's any software that is a silver bullet for diagnosing in this situation. Most seem to focus on one (or a few aspects), thus seem more helpful if you already suspect a problem. Or, will show you system/component performance relative to similar components to help ensure you/ve got it set up right.

Reveiwing BIOS settings, ensuring chipset and other drivers are installed. And swapping parts seems to work best for me.

Oh, run the DXdiag too.

Good Luck

Barton 2500+
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
512MB Corsair Twinx 3200LL
9600 Pro
Enermax Noisetaker 420 watts
Win98SE

 

[Codesmith]

I found the problem. The case was the last part to arrive so I was testing the motherboard with it sitting on an anti static bag.

Now that it is grounded to the case via the mounting post it is running perfect.

I have run other motherboard out of cases before without a problem, not sure if this board is extra susceptible or the previous ones extra resilient.

I will give prime 95 a try. I remember someone recommending a different program that tested more of the extended instruction sets (SSE MMX etc). Unfortunately I can't recall what forum.

 

[Coyote]

"I found the problem. The case was the last part to arrive so I was testing the motherboard with it sitting on an anti static bag.

Now that it is grounded to the case via the mounting post it is running perfect."


Damn, that's weird! I always setup my new mobo's out of the case to test that they work fine b4 putting them in the case. Too much of a PITA to put in/take out for testing if something ain't working right.

BTW, I hear it's not a good idea to set mobo's and such on the antistatic bag. They say the static on those resides on the outside of the bag. Don't know if that's true, as I've done b4. Course, could always turn the bag inside-out.

BTW, your mobo ain't grounding to the case or it wouldn't work.

I've got a link to a site with a large amount of links to utility/testing proggies for all components. But it's on my gaming pc at home. I'll try to remember to bring into office tomorrow and post the site for you. If I forget pm me.

Barton 2500+
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
512MB Corsair Twinx 3200LL
9600 Pro
Enermax Noisetaker 420 watts
Win98SE

 

[Codesmith]

It has also been my experience that a motherboard will work just as well out of a case as in.

However when this motherboard was outside the case the voltage would drop from 3.3 to 3.2 when I connected a fan to the motherboard, and it would ALWAYS crash before Sandra's CPU test completed 10 times.

After mounting it to the removable motherboard tray in my case I could attach as many fans as I wanted to the motherboard and the voltage stayed at 3.3. Running the CPU test for over 8 hours didn't result in a single CPU crash.

When accidental contact between two points on a circuit occur it is "shorted" not "grounded". When a wire has a break in it, its an "open" not a "short".

Electrical compo nets always have a shared ground line that completes the circuit.

Its common practice for PCB's to connect this common ground to the metal housing they are mounted.

This is in benificial in some way that that my friend with the ET degree tries to explain whenever I ask, but I never quite grasp.

I am honestly just guessing that the metal ring around the mounting hole on the motherboard is used to ground it to the case via the posts. I just know that this is sometimes done with some electronics.

Setting a component on a conductive surface does in fact create an easier path for the EDS and increases the risk of damage. Its like the difference between touching a live wire and toughing a life wire while standing in a puddle.

The metallic once I believe are conductive on the outside and non-conductive on the inside with the idea being that any ESD or EMI would move around the bag rather than through it. Whats in the bag is safe, whats sitting on the bag isn't.

The bag I was using is the non-conductive bag motherboards ship in. Using it protects the finish of my table, but does nothing against ESD. Thats why I wear a dorky wrist strap grounded to a metal object.

 

[Crashman]

I've never used a wrist strap, worked over carpetted floors for years, done over 1000 system builds, and low and behold, no static damage...POP! DOH, I shouldn't have opened my mouth!

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[Coyote]

Try this link for utlity programs. Most are free. Perhaps some will be useful for diagnosing probs <A HREF="http://www.benchmarkhq.ru/english.html?/be_hdd.html" target="_new">http://www.benchmarkhq.ru/english.html?/be_hdd.html</A>

Edit: Still find it wierd that the mobo only workrd after putiing it in tray. I admit I don't understand it. I always thought the black (ground) wires in the molex connectors took care of any neccessary grounding to complete the circuit. But crap, I'm just an accountant.

Cheers

Barton 2500+
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
512MB Corsair Twinx 3200LL
BBA 9800 Pro
Enermax Noisetaker 420 watts
Win98SE<P ID="edit"><FONT SIZE=-1><EM>Edited by Coyote on 05/10/04 02:37 PM.</EM></FONT></P>

 

[Codesmith]

I also find it very wierd and contrary to all previous experience.

But all the problems dissapeared when I mounted it and all the problems reappeared when I unmounted it.

It went from a crash happy system that would lock up while sitting idle, to me running an NVIDIA demo,a CPU benchmark and simultaneously formating two 120 GB hard drives for 8 hours without a hitch.

I talked to a friend with an Electronics Degree from Devry and he was the once who suggesting it might work find as soon as I mounted the motherboard to the metal tray.

All I know if that from now on, I won't start building until the case arrives.

But thanks for the stablity testing tips, though I was hoping for someone to say run program A, B, C and then you are 100% certain you have a stable system.









<P ID="edit"><FONT SIZE=-1><EM>Edited by codesmith on 05/11/04 09:03 AM.</EM></FONT></P>

 

[PukePile]

I did same thing when i frist built my computer

Oh yeah 40k
<A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=77119011
" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=77119011
</A>
P4b2.4@2.8||1024mb DDR@392mhz||9800pro@314/366||Viewsonic UltraBrite 19"||Kensington 800 dpi mouse.