The Corsair User Forums

The Corsair User Forums (https://forum.corsair.com/forums/index.php)
-   Memory (https://forum.corsair.com/forums/forumdisplay.php?f=198)
-   -   M2N-SLI Deluxe and TWIN2X2048-8500C5D (https://forum.corsair.com/forums/showthread.php?t=74568)

IvanPro 11-25-2008 08:41 AM

M2N-SLI Deluxe and TWIN2X2048-8500C5D
 
3 Attachment(s)
Hello!

I'm having problems running my configuration.
The system seems to be stable, it POSTS, boots and works normally but then, for no apparent reason it freezes or shuts down in Windows.

It does that in 2 specific ways:

1) it freezes completely the OS while working, mouse led is on as are keyboard leds, case and CPU fans rotate properly, only solution is to reset, then it works as normal.

2) it shuts down out of the blue. I've never seen this happen directly because it always happens when I'm far from the PC or sleeping (=.=). However I'm inclined to think it may occur after several hours the system is on (mainly when I leave it on during the night to download things from the internet). What is strange is that the mouse led is off (it's USB) but the NumLock led on the keyboard stays on (so is as well for the motherboard led, the power supply led, and the ethernet red and green leds).

It doesn't happen always, but then when it happens it can do it even 2 or 3 times in a row, then work properly for days.

I think it may be worth to mention that the system is connected to an APC Back-UPS RS 800 so current shortage or fluctuation is out of question (besides my other system doesn't run on UPS now and is just fine, and anyway I tested it on this UPS and works fine, not to mention that that old system was connected to this UPS before I assembled this new one, so the UPS is fine too).

Memtest completes succesfully multiple times with Standard BIOS addressing, however if I set the addressing to BIOS All or Probe, the program spits tons of errors immediately and crashes.
I've tried to switch the memory from one channel to the other.

BIOS is 1405. Not too new I know, but the newer one doesn't list added compatibility with RAM, only CPUs. Settings are left to auto, I don't need neither want overclock, just a rock solid stable system, so AI Settings are left to Standard.
System devices I don't use such as RAID controller, integrated audio, floppy and IDE, LPT ports, Firewire ecc. are disabled in BIOS, Cool'n Quiet is enabled with drivers in place under Windows and the proper power scheme is activated. Latest chipset drivers are installed.

The system runs pretty cold: 18° CPU, 27° MB, 32° GPU, 22 HD°.

I don't know really what to try. Should I tinker with the RAM settings manually? Update the BIOS? Please let me know.

Attached are three images from CPU-Z results.

Thanks in advance.

Ivan.

Yellowbeard 11-25-2008 02:02 PM

Try the newest BIOS for sure. The MOBO makers are not going to list all changes they have made publicly so you may benefit from having the newer BIOS.

Also, if you are running the memory at DDR800, manually set the voltage to 1.9v-2.0v, and manually set the timings to 4-4-4-12, 2T command rate.

If you are trying to run the memory faster than DDR800, manually set to 2.1v, and 5-5-5-15 timings, 2T command rate.

IvanPro 11-25-2008 04:35 PM

Thank you very much specmike! Very fast support!

Well thank you for your advise. I will try that.
Still, I've got a question unanswered: I've set the voltage to 2.1 while leaving timings on auto, and have disabled USB legacy support (I've disabled USB entirely :-P). Now the Memtest completes on Standard BIOS addressing and on Probe, but still it crashes on All Bios addressing setting. Is that normal?

Anyway I will try your advise and see if that helps.

Best regards, and thank you very much for your support again!

Ivan.

RAM GUY 11-26-2008 02:34 AM

Yes please read the read me file that come with the down load or from the site where you down loaded memtest. This is common and its not suggested running all test or the Probe with some system especially X86 based systems. Just use the default tests unless you fully understand the tests you are running.

Synthohol 11-26-2008 04:19 AM

1 Attachment(s)
just my .02
your CPU should be running at 2.7Ghz your multi is WAAAAYYY too low and the RAM should be running at 400 at least.
your rig is tremendously UNDERCLOCKED!!
this can be the issue as well.
i suggest resetting the bios and entering the RAM settings manually and hopefully the PC will run at stock speed at least.

i have the M2N32-SLI board and i know there are a billion settings to tinker with.

personally i would start there.
your bus speed should be 200 and your memory frequency should be at least at 400!!


good luck!!

13x200 is stock IIRC not 5! heres a screeny of a slight overclock but yours should be similar at 13x200
http://www.houseofhelp.com/forums/at...achmentid=4669

IvanPro 11-26-2008 08:58 AM

Hello to you all!

Again, thank you very much.

Yesterday I did the following:

update M2N-SLI Deluxe to BIOS version 1502 (latest)
set timings to 2.1v, and 5-5-5-15 timings, 2T command rate
everything else on Auto, and other settings still as mentioned in my first post
Cool'n Quiet enabled (Synthohol, I believe my system looks underclocked because of this, I will post CPU-Z images of my system under load if you wish)

Unfortunately the system shut down 4 hours after all that. I left it on downloading, and it simply shut down as described before (no fans, only keyboard, power supply, and mobo leds on). I was there this time and I can assure it was not an electric shortage (the router attached to it was just fine).

I'm desperate about this. What to do? Any suggestion really appreciated. Thanks to you all again BTW!

Ivan.

Yellowbeard 11-26-2008 11:04 AM

Hmmm.....odd. Can you check the Event Viewer and see if there were any events/errors reported at the time the system quit?

This is starting to sound possibly heat or power related but, you said your temps were nice n cool.

IvanPro 11-26-2008 01:30 PM

Ok. Thanks specmike.

I've tried to look at the Event Viewer but no, there doesn't appear to be anything suspicious there. In fact the logs generally end at 23:50 PM of 25 November (the time I booted in the OS). The crash happened at 02:00 AM of 26 November.

What I'm to try now? I won't give up easily on this issue, so anything is still welcome! (RMA is always a PAIN here in Italy, and my dealer won't be collaborative I'm afraid)

Synthohol 11-26-2008 03:19 PM

ok, i was never fond of any of that cool&quiet or throttling stuff. i find it difficult to buy a cpu and not have it run at rated speed or better. also power mgt is off on my stuff.

maybe as a test you can turn all that stuff off as well for a while?

fwiw, i run BIOS 0706. yes its older but i have tried them all and this one is the most stable of them all and allows me to overclock 24/7 with no problems whatsoever!

i wish the newer ones didnt crash my system like they did.

there is the other avenue to try if all else fails too.
you will need to force the older bios on there with the /force command if i remember.

at least you have 2 new things to try:biggrin:

IvanPro 11-27-2008 04:59 PM

Hello guys.

Just a quick entry to tell you I'm on Cool'n Quiet disabled now as suggested and will come back with the results ASAP. I think downgrading the BIOS is a little risky as of now and I will resort to that if nothing else is left (I've got this friend that maybe can lend me some Corsair low latency PC800 memory to test the board...)

Now the (a little) OT part.
I'm intentioned to buy the HX Series Corsair PSU, just to be on the safe side (I don't trust that Thermaltake and besides I want to go modular).
The 520W version would be enough for future proof upgrades considering my system? Or should I go teh 620W way FTW :-P?
I've got the Antec P180 case. Would the cables be long enough for it? Particularly about the mainboard and CPU cables (the latter on the Thermaltake is too short so I can't make it pass through the back of the case... awful...)
End OT.

More later!

Ivan

IvanPro 11-27-2008 08:23 PM

1 Attachment(s)
Ok, I've tried to disable Cool'n Quiet from BIOS, disabled power management from Windows and every other possible gimmick.

When I was back I found the system off. Shutdown as always (at least this night it did the little variation of freezing... so it actually sucked energy the whole night... duh...). So I guess it's not CnQ.

Besides I'm noticing that maybe the system is beginning to shutdown itself faster, but this may just be a coincidence or an impression.

I wanted to show a pic of CPU-Z in the current status (RAM timings and voltages set manually to manufacturer recommended values, minimal BIOS peripherals and settings activated, no Cool'n Quiet, defaults and auto settings wherever available) so that you can check if there is anything wrong (comparing my system to Synthohol's is that HT Link 1004.6 MHz value normal?).

I'll wait for your comments on this.
Thanks and have a nice holiday :-)

Ivan.

Synthohol 11-27-2008 09:45 PM

nah, your not far off, id give a bit more voltage to the cpu but thats me.
do you have another PSU to try? its the common overlooked item sometimes.
does the bios report good voltage +-10%??

here is my mild OC i bumped the multiplier up because its unlocked and a tiny bit on the FSB.

http://www.liquidninjas.com/photopos....php?file=2833

IvanPro 12-01-2008 07:36 AM

1 Attachment(s)
Hello to you all again.

I've been doing a little more research on this issue.

Well, I'm sending you another pic were you can see something I consider very interesting.

There are some readings:

Vcore from Asus Probe is at 1.38
VCore from CPU-Z is at 1.232

This is not a slight difference.

I left (again) my PC on this night too, to see what would happen and if Asus Probe could help me in some way. It didn't crash (a miracle) but Asus Probe recorded VCore being "abnormal" to a whopping 3.33V?!?!?!?!?

I don't know what to think.

Then there is something else. As I stated earlier I set manually the voltages for the memory banks at 2.1V. I took a look at Speedfan under the Tab Info - Read DIMM Info and this is the result:

Scanning SMBus at $1C00...
Decoding DIMM #0
Memory type is DDR2
Module Ranks : 2
Levels : 1.8V
Parity : NO PARITY
Refresh Rate : 7.8us
Total Size : 1024MB
Decoding DIMM #1
Memory type is DDR2
Module Ranks : 2
Levels : 1.8V
Parity : NO PARITY
Refresh Rate : 7.8us
Total Size : 1024MB
Scanning SMBus at $1C40...



It says "Levels: 1.8V"? For both the banks? Is that normal? Otherwise I'd say something is seriously borked here (which you couldn't even tell because the PC works like a champ... besides it randomly freezing LOL). Maybe RAMGuy can throw his 2 cents at me? I'm crazying over this, I'm no Pro at this (I almost stopped keeping informed on PC technology around the A7N times) and all these readings cores voltages and whatnot are getting me paranoid (I dreamt of my PC locking while I was saving a SUPERVERYIMPORTANTWTFBBQ document LOL).


Thanks for your attention, I'm eagerly awaiting for your answers and considerations.

Ivan.

Yellowbeard 12-01-2008 01:47 PM

Software voltage monitoring is notoriously unreliable. If you are concerned about this voltage, you'll need to check it with a meter.

IvanPro 12-01-2008 05:44 PM

Thanks specmike for your time :-)

I know software voltage monitoring can be dodgy. Still my question was of a slightly different nature, it's just that I didn't explain myself very well. This is a setup that currently has big issues, it's not like I'm on a smooth working system and then I alarm myself because of some voltage warning. I'd like to know if this voltage peak that has happened (and that maybe has always been there but never showed up because of the PC freezing and then being unable to record it), could be caused by some specific component. I'm thinking PSU here, but then I can also think about RAM, but don't know if other peripherals can have influence on this.

I'm very concerned as you can imagine because now I fear that everytime I use the PC I risk to burn it. And I need it for work, so I can't avoid using it being that it's the only machine I own.

Besides, what do you think about Speedfan's results regarding RAM, which you can find in my previous post?

P.S. As per your suggestion, I've talked with someone who has the skills to use a volt meter on the mobo. Where can I find info about the solder joints I've to probe with it to measure the VCore?

Many many thanks really!

Ivan.


All times are GMT -4. The time now is 10:43 PM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.