Sign in with
Sign up | Sign in
Your question

New Build - Stable on and off (mostly off), yet passing all tests?

Last response: in Motherboards
Share
October 4, 2010 7:04:10 PM

Okay so I’m hoping to get some feedback on this issues I’ve been having. I am pretty sure I need to RMA something, I just want to make sure I RMA the right part (I’m thinking mobo right now)

Story:

I built a new computer a week or two ago…

AMD Athlon II X4 620 Propus 2.6GHz
G.SKILL 4GB (2 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600
GIGABYTE GV-N460OC-1GI GeForce GTX 460
ASUS M4A87TD/USB3
CORSAIR CMPSU-650TX 650W

Everything seemed to be going fine with all BIOS settings on auto for a day or two. Ran some stability tests, received no errors (save for one error generated on the Linpack OCCT Test). Played some games the next day and it froze. This first freeze seemed to start a process where this machine would freeze more and more frequently after this first one. After 5 or so freezes it wouldn’t even get into windows.

Thinking that my ALL Auto’s were to fault, I reset the CMOS, set these settings, and again the computer ran well for 1-2 days. In this uptime I ran much more stability tests to verify that things were stable. It passed hours of 3 hours of various Prime95 tests, a 12-hour cycle of the OCCT CPU test, several hours on Furmark, and this time it also passed a 1-hour test of the OCCT Linpack. So I thought “yay! It’s fixed!” Played some more games, then it eventually crashed and started the above cycle again—crashes became more and more frequent. This time, however, it would sometimes not even POST. I had to reset CMOS again, reconfigured the bios settings (this time with some extra voltage thrown on the ram and CPU)… Ran fine for another day, then all of this happened over again. Froze in a game, got more frequent, etc… again to the point of it freezing before windows, sometimes not POSTing, sometimes even freezing in the BIOS settings screen.
This is where I’m at… when the computer is “up” it’s passing stress tests… when it starts to freeze, it quickly chokes and completely won’t go anywhere. Right now, I can’t get it stable again… it just keep crashing while trying to boot up.

Basic question: does this sound like a mobo problem?

Asus tech support says it isn’t their mobo’s fault and that it is probably a faulty memory controller in my CPU or my BIOS settings, or ram compatibility. (While my ram is not on the QVL for this mobo, it’s siblings are (6gb version… I have the 4gb version). The thing is, wouldn’t a faulty memory controller not test well in Prime95? Etc.

Thanks for any light you guys can shed on this bizarre problem.

Cheers, Chocko.
a c 198 V Motherboard
October 4, 2010 7:45:15 PM

I can't tell or not from your posting, but are you overclocking anything (CPU, GPU, memory)? If so, stop until you stabilize the system.

Second, how are your system temps when it was playing? You seem to be describing a heat problem. Are you using an aftermarket or stock heatsink/fan (HSF)?
a c 715 V Motherboard
October 4, 2010 7:47:53 PM

In looking you are barely in the green with the PSU if everything is OC. Please post either the part-number or link to your DDR; there's Certified {QVL} and there's Tested.

What EXACT OC settings are you using? And a bigger pain what other BIOS modifications? What CPU cooler - how are the temps? CPUz + Prime95 10+ minutes would suffice. Any RAID or other unusual settings? Did you breadboard to see if "maybe" you have a partial short? The more info the better.
Related resources
October 4, 2010 8:04:30 PM

This system is 100% stock. No OC, stock coolers. Lots of fans. I was monitoring temps while doing all my testing. Load temps are ~50c on the cpu.

I have not taken it out of the case. No unusual settings as far as i know... here's a screen of what i was trying in my bios:

2 things you can see in this screenshot: my timings are at 9-9-9-24 manually, and i have turned off "Cool n' Quiet".

..and here's a screen i took when i was able to get into windows:


Also, thanks for the quick-ass replies!
October 4, 2010 8:06:25 PM

oh, to note... in the BIOS screenshot... the latest thing i'd tried is disabling ALL spread spectrums and enabling ALL Load-Line Calibrations.
a c 715 V Motherboard
October 4, 2010 8:42:03 PM

CPUz add this to the mix, I'm looking at them all.
October 4, 2010 9:39:12 PM

jaquith said:
CPUz add this to the mix, I'm looking at them all.
http://www.cpuid.com/medias/images/en/softwares-cpuz-05.jpg


that one is out of my reach now, i'm afraid. I've been unable to get back into windows since this last time it lost stability.

Oh, this is the ram: http://www.newegg.com/Product/Product.aspx?Item=N82E168...

...and here's a link my board on asus's website where you can view the QVL: http://usa.asus.com/product.aspx?P_ID=nlWYrI9wlNIYHAaa
a c 715 V Motherboard
October 4, 2010 9:42:59 PM

^ Posts crossed - The DDR part-number will be fine. My initial thoughts are: non-compliant DDR, Mild Short, bad OC, BIOS version and/or settings. Once I have all of the info I will research DDR/MOBO/BIOS.

Researching now.
October 4, 2010 9:55:52 PM

jaquith said:
^ Posts crossed - The DDR part-number will be fine. My initial thoughts are: non-compliant DDR, Mild Short, bad OC, BIOS version and/or settings. Once I have all of the info I will research DDR/MOBO/BIOS.

Researching now.


to rule out at least one of those, i have flashed to the most recent bios upgrade.

This is actually the first time i've looked at a QVL... that'd be lame if my this ram simply wasn't compatible... i feel like an idiot for not checking before hand.
a c 715 V Motherboard
October 4, 2010 10:05:56 PM

^ I was typing while you ere posting again - noted the BIOS update.

RAM:
Not that the {QVL} is the end all choice, but I know that there are over a dozen memory controllers for DDR memory and some work, might work, create stability issues, or simply don't work.
RAM Replacements:
http://www.newegg.com/Product/Product.aspx?Item=N82E168...

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

BIOS:
Verify the you indeed have 0801 or 1005 BIOS installed, if not update.
M4A87TD BIOS 0801 Improve system stability.
The {1005} though may only list a K/B issue, but trust me there's more...typically CPU related if unlisted.


BIOS Settings:
Load Optimized for now.

Short Testing:
I realize breadboading is a pain, so in lieu of that unscrew all of the MOBO screws and dangle and/or support with a towel.
October 4, 2010 10:21:13 PM

jaquith said:
^ I was typing while you ere posting again - noted the BIOS update.

RAM:
Not that the {QVL} is the end all choice, but I know that there are over a dozen memory controllers for DDR memory and some work, might work, create stability issues, or simply don't work.
RAM Replacements:
http://www.newegg.com/Product/Product.aspx?Item=N82E168...

http://www.newegg.com/Product/Product.aspx?Item=N82E168...



So to verify... incompatible ram, when it's working at the time, will not throw any memtest errors?
a c 715 V Motherboard
October 4, 2010 10:41:57 PM

^ Not necessarily - there's a difference between bad RAM {errors & compatible}. My reasonable assumption is that your chosen memory is causing the instability {works sort of okay and then it won't}. You can try to dummy-down {failsafe} the RAM using the MEMOK button on the MOBO. /note my first post./

Incompatible memory, as I mentioned, will do all sorts of oddball behavior. Enabling XMP for unsupported DDR is a guaranteed post failure/unstable system.

Some time ago, I purchased some Corsair the was 'supposed' to be XMP compatible so I enabled on a GA MOBO - yee haw - Post Failures galore, but what I failed to realize is the BIOS update included the memory profile. After that I don't even bother, I set all memory manually.

If a short is ruled-out then I would order a new set and/or exchange.

Refer to pages 14 & 28 MEMOK, 34 PSU, 55-(56)-59 {56 unstable}.

Footnote: when it comes down to it, it is difficult to determine or R/O -> CPU, MOBO, or RAM in many situations without replacing 2 of the 3 items. The RAM may or may not be the reason for your issue, but for now it is a reasonable assumption.
a b V Motherboard
October 5, 2010 1:51:12 AM

I maybe wrong but your problem sounds to me very like a power supply issue. Faulty memory would give consistent errors perhaps depending on temperature. Only a faulty power supply or perhaps (unlikely) the motherboard could give rise to the symptoms you describe.
a c 715 V Motherboard
October 5, 2010 2:30:38 AM

^Anything is possible, I noted the PSU is barley in the green which simply means that it is adequate. Corsair's are single rail + Japanese capacitors and are for the $ is an excellent PSU.

When you get oddball behavior it can be a lot of things {footnote}. I am use to building Certified rigs so I am in the habit of scrutinizing parts.
October 5, 2010 2:44:39 AM

I actually had a very similar problem a few weeks ago. Built a new system for a friend. Gigabyte mobo, Phenom2 x4 black, Patriot Sector5 ram. It would freeze every few hours for no reason, somtimes not even using it, just with the screen saver up. Then eventually wouldnt even hit POST. I thought right away the PSU. No luck, tried different ram, running the board bare, 2 PSUs. Got my friend to take it to a local shop, paid 25$, for them to say it was the mobo for sure. Was definitly worth it.
October 6, 2010 6:57:20 PM

Well, i tried different ram and a different PSU in the machine, but still could not get it to post.... son of a b.

Your footnote
jaquith said:
Footnote: when it comes down to it, it is difficult to determine or R/O -> CPU, MOBO, or RAM in many situations without replacing 2 of the 3 items. The RAM may or may not be the reason for your issue, but for now it is a reasonable assumption.

is spot on imo...

So I RMA'd the mobo yesterday. If that is a no go, i guess the ram or CPU is next. I chose the mobo because i'd tested other components from my buddy's build in my machine, except for CPU and mobo. The ram didn't help it post at all. In thinking back though, i should have tried testing the various ram slots. Alas, nothing i can do now :) 

Will keep thread posted!
a c 715 V Motherboard
October 6, 2010 7:15:29 PM

Thanks for the update - trouble shooting is soooo much fun! :pt1cable: 
October 14, 2010 6:01:13 PM

So I got the motherboard back yesterday... and guess what? The problem was not the motherboard... AHH!!

Upon assembling everything, i experienced the same exact problems i was before. First attempt on turning it on--no POST. Second attempt--Froze in the Bios.

So now i have it out of the case, to rule out any shorts going on. So i threw in a stick of my buddy's ram (also not on the compatibility list for my mobo, though his machine is working fine), which at least got my system to boot up. However, i think i've kinda ruled out the ram because even with his stick in there, the computer would freez/restart/go black-screen during OCCT/Prime testing. (though it is to be noted that the system likes his ram better)

Not knowing what to do i began messing with Bios settings again (because i could at least get there now). I set everything manually and tried my ram again. Boot problems again. Threw in his, again--was okay until it froze again.

When i'd throw in both sticks of my ram, it would tell me that my "Overclocking has Failed!"

So i went into the ram settings, set the timings for my ram again (9-9-9-24, set voltage to 1.6, and told it to set the frequency at 1600), but this time also changed 1T to 2T, because i'd seen a G Skill rep tell someone to use 2T with this ram somewhere. (should note, however, that i've seen people run it stable at 1T though)

So far on 2T, it is stable... wtf lol. Just passed 12 hours of memtest. I'm going to reassemble and test it for a day or two, see if this crops up again.

I find it wierd that i'd see such a huge stability difference between 1T and 2T...

Anyone have any thoughts?
October 14, 2010 6:26:13 PM

also... this thread i found: http://www.gskill.us/forum/showthread.php?t=4059 though very long, touches on my issues exactly. They're hypothesizing that Asus' Bios software is to blame in that it is not holding on to values very long. As such, you'll make a change, it'll run okay, and then it'll forget or something and have the problems all over again. ...sounds plausible, although its focus is cold booting, which seems irrelevant in my case.

At least i've got it down to CPU and RAM if i need to RMA something else.
October 15, 2010 4:44:50 AM

well, thing looked good for awhile. went 16 passes in memtest with 0 errors. Survived some internet browsing and a match of LoL, then about 20 minutes of Crysis. I took a break and watched tv for awhile, came back and played for about 2 minutes and it froze. Which led it to another freeze after a short while in windows. After that it wouldn't post, then it told me my Overclocking had failed, like it has before.

So, i guess it's down to the ram or the memory controller in my cpu. Should i flip a coin on which to RMA or do you guys have a recommendation?
a b V Motherboard
October 15, 2010 7:07:13 AM

I doubt very much if it is a CPU problem, processors are exhaustedly tested by the manufacture before being shipped. You have tried different memory and timings without any luck, so it looks more and more like a power supply problem to me and that is the next thing to try.
November 2, 2010 6:41:03 PM

Hi Guys, long time no post--but yes, it's taken this long to get concrete news.

I'll sum all this up quickly. I've found the problem and have had a stable computer for 3 days now.

Neither my motherboard replacement nor my CPU replacement fixed it, so i think it's pretty safe to say I completed 2 wasteful and useless RMA's, unfortunately.

Obviously at this point, I don't give *** about the above---I'm just thrilled to FINALLY have a working machine.

What was the trick?

My northbridge was underpowered. Thanks to a G.Skill rep for suggesting an increase of .1v, which made it a little more stable... so i decided to keep messing around with the available range of voltage--+.2v ended up working like a charm.

At least i hope so! My fingers are still crossed, but i'm pretty sure i've got it this time.

Thanks to all for your help, insight, and mostly your TIME!

Cheers!
!