PSU or Driver Issue?

[jackshy]

Hello, I have recently replaced the stock motherboard in my Dell XPS630i with the ASUS P5N-T Deluxe. I also replaced the video card with 2 GeForce 9800GTX 512M in SLI. I have two Seagate 500G drives in a RAID-0 config and a CD Burner,DVD Burner, and Bluetooth/Flash Card Reader.

Recently I find that when I'm gaming I can hear the fans ramp up and the video will freeze and then the system reboots. I have turned down the graphics options in the game to minimize and/or eliminate this.

My question is - is this power supply related since I still have the stock Dell PSU that shipped with the system or is it driver related?

I think that the standard PSU is 750W but I haven't checked this as yet. I'm trying to figure out where the $$ needs to go to fix this so I'm hoping you guys can point me in the right direction.

Thanks in advance for your help!

 

[dndhatcher]

Have you checked the GPU temps? Sounds like maybe they are overheating? Have you cleaned everything or is dust reducing your cooling? Has it ever worked with the current hardware?

 

[T_T]

Without the specs on the PSU, it's hard to say. Post the specs on your PSU.

Generally the PSU problem would be that there isn't enough voltage to power up the SLi mode, but considering your concern is "freezing", your PSU is powering up your cards.

When you replaced the motherboard, did you remember to reapply thermal paste (properly) over the CPU before attaching the heatsink? Your CPU may be the one causing the freezing due to overheating.

Also try running one card at a time, maybe one isn't working properly.

 

[ubernoobie]

I agree to dndhatcher, check temperatures then if it's under 80C load then it's your psu, get a psu like this:http://www.newegg.com/Product/Product.aspx?Item=N82E16817371035

 

[jackshy]

Thanks for the help. I went home last night and ran it with the case open for 2-3 hours without issue on the highest settings in Aion so I thought we had it narrowed down. This morning (after my car accident - yeah, it's that kind of week) I started running it with the cover off again and it did it's video lockup/reboot routine in less than an hour. Any suggestions on a PSU or temp monitoring software so that I can try to get a log of what is going on when this happens?

The PSU is a DELL Model #H750E-01 P/N HP-D7501A001

The other thought I had was this - the MBoard is 3-way SLI capable but I have only two - I put the existing cards one on top of the other slot but, to me, that doesn't look like it gives good airflow for the vidcard fan. Can I separate the two cards with an empty SLI slot in the middle? This is my first SLI rig and so I am not sure i put them in right.

 

[dndhatcher]

EVGA.com has software called precision that will let you overclock and monitor temps of any Nvidia GPU.

 

[jackshy]

Ok, I installed the evga Precision software and Coretemp to monitor both GPUs and the CPU - all of them stay in the range of 37-43 degrees Celsius. I also moved the two SLI cards apart so that they are in pci-e slots 1 and 3 to give better airflow. The system has twice gone into the video lockup/reboot - once the video locked up the pc and I had to manually reboot and once the video locked and the pc rebooted itself.

I then went into the BIOS on the M/Board and made some further changes - setting clock/cpu control to standard instead of automatic and the primary video card was set to pci instead of pci-e and set the QFan control from Performance to Optimal. I will update as soon as I have more information but it seemed to be more stable.

 

[dndhatcher]

Yiou should try the graphics cards separately to see if it is even a graphics card issue. Could something like bad memory bios settings.

 

[jackshy]

I ran the Windows Memory Diagnostics all last night (approx. 12 hours) without any errors or reboots. I downloaded a memory tester for the GPUs but haven't run that yet - probably overnight like the last one. I played about an hour of WoW without any issues on the highest settings. I'm starting to wonder if it could be Aion itself causing the issue but I am going to continue testing with both and then one vidcard at a time with Aion today as suggested. Also, the PSU itself is puttting out very hot air from the back. I don't notice that with any other part of the system but that could be completely normal.

 

[Alvin Smith]

Has anyone verified the output of that PSU yet ?!

It should ALL be on a sticker, on the PSU. DELL also has a very complete Dbase of all their sys/parts ... If you give your ser#, they can see your orig config, exactly.

I really do think that (A) verifying what the PSU is "supposed" to put out (if not what it is *actually* putting out) would be a pre-requisite, before even beginning to discuss temps.

If it is an old PSU ... it would be high on my suspect list.

Also, ... have you logged onto the GPU site, and updated all drivers?

Can you get a DELL tech to read the BIOS release (notes) history, for your rig?

Even if u r out of warranry, u should b able to logon to Dell sppt site and read the history and to check your current version, in POST.

Reseat the cards ... make sure they are ALL the way seated. You could swap'em, if identical. Make sure the interconnct is firm, etc.

= Al =

 

[jackshy]

I am pretty fanatical about keeping my drivers up to date and if/when I experience issues, it's usually the first step I take in getting them updated (if they aren't the latest already).

After running the system memtests overnight, I also used the machine all day with no issues until about 9PM when the video went black on both monitors and I couldn't get it back. I shut down the machine at that point and pulled the second vidcard, shut the case, powered up and played on the system for another 3-4 hours without issue.

I had posted this earlier but not sure you saw it: The PSU is a DELL Model #H750E-01 P/N HP-D7501A001

Also, even with one vidcard running, the NVidia PhysX screensaver consistently crashes if I try a preview although I'm not sure if this is a Windows 7 issue more than the vidcard.

My concern right now is that if I do have a defective card then it's probably under warranty. But how would I go about proving there is an issue with it so that I can get it replaced?

I will try pushing the system today to see if it will crash with only one vidcard running in the system.

 

[Alvin Smith]

And, you can run with one card (as only card) and then the other card (as only card) ...
THAT would prove a HW issue with one of the cards (barring intermittent PSU power problems).

If drivers are peachy, then, prolly one of the vid cards, or the PSU ... If we can at least "prove" the problem is HW and limited to one of these 3 parts, then we ARE actually getting somewhere.

Time to start taking acurate notes (really) and I think swapping the cards, for "whole days" and running them through the same problem tests ... both in SLI and as individuals, will eventually yeild a true solution.

Of coarse, if the cards change and the probs do not move (clearly) with the cards/configs, then the PSU is highly suspect (if the drivers are all peachy).

= Al =

 

[T_T]

Jack, you also mentioned repeatedly that your SLI set up is in slots 1 & 3. Is that how the manual instructs you? I'm not saying your theory is without merit, but you must consider that the R&D from the manufacturer must know what they are doing when designing the mobo to have 3 PCI-E slots. Yes, there will be a greater gap, thus 'potentially' allowing better air flow, but the cards are self-cooled and when working properly, don't get hot enough to cause your system to fail.

My point is make sure you configure your SLI properly to rule out the possibility of misconfiguration. Secondly, don't be shy about seeking an RMA from the manufacturer of your GPU's. I had no conclusive evidence that my PNY GTS 250 was defective, but I just recently had to send mine in for replacement, as I was experiencing similar 'video' problems.

 

[jackshy]

Thanks to everyone for your continued patience and help. I have always been the "software" guy not the "hardware" guy and some of this is a learning curve.

@T_T - I went through the ASUS site and looked up a few articles regarding the P5N-T Deluxe board. The 3 SLI slots are (1) Lime Green (2) Black (3) Lime Green and indications are that with a 2-card SLI setup I should be using the two similar colored i.e. 1 &3. In addition, I realized that ASUS shipped a solid board connector for 3-way SLI that can also be used for 2-way but I was using the ribbon that came with the cards instead. Last night I put the second card back in and put the ASUS connector in place.

I also realized that while I asked everyone here for suggestions on monitoring software, I never installed the ASUS AI suite or other software that came with the board (because I'm an idiot apparently). I did this last night and I plan on stress-testing the rig tomorrow.

FYI - I have been using the Precision software and also Coretemp since last weekend as well and the CPU temp rarely goes above 40 C while the vidcards when pushed may get up to mid-50s. I am anti-overclocking with my systems because I just don't have the cash to risk replacing anything so that does not have to be factored into any of the solutions.

I have a large front-intake fan, the CPU cooler (DELL XPS stock), and both vidcards have their own cooling but I have no rear fan to push the air out so I am going to pick one of those up today.

Thanks again to everyone for your continuing suggestions.....

 

[jackshy]

Also, I meant to mention above that the system memory max is 8G which I have as 4x2G modules. In the BIOS there is a message about "Due to chipset limitation, SLI-Ready Memory supports Max. 2 DIMM module combination" which I am taking to mean I should actually have a matched pair of 2x4G. With the current configuration, would the system memory config cause issues with the vidcard memory which might lead to this type of issue. If so, am I better off taking two of the DIMMS out until I can afford the 2x4G combination?

 

[T_T]

In short, yes. You have a configuration conflict. Is there an "Expert" mode in your BIOS? If so, set your BIOS to Expert so that you can configure the RAM timings.

"Due to chipset limitation, SLI-Ready Memory supports Max. 2 DIMM module combination" pertains to overclocking SLI-Ready memory modules. It means you are limited to two memory SLI memory modules when you want to overclock.

New questions: Are you using SLI-Ready memory modules? Are they OC from factory spec?

 

[jackshy]

It looks as if the answer to that is "no" - I removed two of the DIMMS when I got home and checked the BIOS but the same message was still in place. It's grayed out with an X in front of it.

The memory is a set of four matching - purchased in single box sets of 2. OCZ2G10662G - PC2-6400 (400 MHz) as listed in CPU-Z which is not listed as SLI-Ready on NVidia's site.

 

[T_T]

Here is a link with the specs of your PSU.
http://www.tomsguide.com/us/dell-xps-630,review-1064-3.html

With 18A on all of the 12V rails, you should have enough juice to keep your system quenched. That being said, you haven't addressed my first question.

"When you replaced the motherboard, did you remember to reapply thermal paste (properly) over the CPU before attaching the heatsink?"

 

[jackshy]

Yes, I did. I use industrial grade paste used on commercial tuggers for the batteries in the auto plant where my dad worked. Sorry for not mentioning that specifically but I did mention the temps of my cpu and both vidcards several times. There doesn't seem to be any logical reason for the behavior. Not to mention that I turned the machine on this morning and it happened twice within the first half hour and the day before I used the machine for 6-8 hours in gaming with no interruptions. It's very sporatic and difficult to track down. Thanks for the follow-up!

 

[Alvin Smith]

Just so someone says it ...

There are enough variables at play, here, on a fairly "pioneering" complex comnfig so now, that you know things are good and "squirrelly" . . .

It might be a sane thing to dedicate a pad of yellow paper to this process (because of the many permeatations of "casualty swapping" and the inevitable, "stip it down to the stand-offs and test (I mean verify and log, without "fuzziies"). And then, build and test and swap and log and build and . . .

You have to simplify, divide, and conquer (or not). You have to get to a point (screws?) where you are on solid ground. If you try a few things, test, then try a few more, well that works for the easy stuff but when you don't know if it is drivers of bios settings or slot-order oo bad-PSU, et. al. . . . ???

Let's get "Hardcore Boolean ! . . . HOO-rahh!

I can be SO obnoxious ... And, my real name isn't Alvin, anyway, so you can't beat me up.

... ok ... That was from "The Book of RTFM : Chapter 07 Verse 22" (Murphy & Moore's Edition 5th Printing January 2003)

This issue kinda reminds me of that section in College Algebra: "Systems of Equations with Many Variables" ... Add Derivatives and you got Engineering Physics.

=sorry=

 

[Alvin Smith]

try a new/diff psu

 

[T_T]

Jack, to add on to what "Alvin" is saying, doing a breadboard as you rebuild the system from the ground up will help you eliminate potential hardware failure/conflict. Take a look at the troubleshooting guide (link in my signature).
Using the suggested yellow pad (or any color, of course), write down the results, changes, experiences, observations, etc. as you continue down the list.
**Note** Breadboarding is technically not in the list, but always worth the time to try.

 

[jackshy]

It took me some time to find it but awhile back Dell sent me at PSU tester to use at work (I'm local pc/network/server hardware support) and I just tested the PSU a few minutes ago. Everything came back perfect so my next "venture" is making sure that BOTH video cards work without any issues when used by themselves. Then it's on to the RAM and other components. Thanks again to everyone for their help.

 

[Alvin Smith]

I guess you already know how much force it takes to fully seat one of those puppies (gpus) so, I won't go there ...

... absolute, simplest config possible.

... Test ... log ... swap ... test ...

= fun ! =

 

[jackshy]

And the winner is.....

Looks like the second video card is bad. Using it singly it creates screen artifacts, random jumps and glitches, etc etc - and that's just at the Windows desktop. I started the RMA process today for returning it to EVGA. Thanks again to everyone for their (continued) patience and helpful suggestions. Once I have the new card, I will drop a line to let everyone know how it turns out....