Question on PC power and cooling

[wintermutemd]

I recently purchased an EVGA 680i mobo, e6600, and an OCZ Gamexstream 700W PSU. The PSU +12 is sitting at around 9V and the -12 around -10 as per Everest listings. I've decided to switch to a PC Power and Cooling Silencer 750. This has a single 12 volt rail at 60A. I'm looking for any possible problems with powering my rig with this new PSU.

Full rig

EVGA 8800GTX (If I decide to run SLI I will get another PSU)
EVGA 680i Mobo
e6600 (don't plan on OCing)
CORSAIR XMS2 2GB (2 x 1GB)
Plextor 18x DVD burners x2

[Oklahoma Wolf]

I doubt you'd have a problem with the PC P&C, but I also doubt you're having a problem with the OCZ - software is an inaccurate way to tell what the voltages are. Best to use a multimeter. Don't worry about -12... it's very unlikely the system is even using it.

[GalvanizedYankee]

Either PSU is way more than enough.

You must take voltage readings with a DMM (digital multi-meter) because software generated readings don't mean squat.

[wintermutemd]

Stupid question...where should one be measuring the voltages with a DMM? I have access to one just no idea where I need to actually be testing at.

[GalvanizedYankee]

Black is ground
Red is 5V
Yellow is 12V
Orange is 3.3V

There are some good links at http://www.jonnyguru.com/ > links = Lots of info ;)

[ChingChang]

You can measure from any of the connectors.. 4 pin molex, or even the 24 pin mobo connector. One end on black (ground) and the other on any color (yellow is +12v, orange 3.3, red 5..)

[jonnyGURU]

I recently purchased an EVGA 680i mobo, e6600, and an OCZ Gamexstream 700W PSU. The PSU +12 is sitting at around 9V and the -12 around -10 as per Everest listings.

And you'd be a fool to believe those readings.

-12V is always wrong since most motherboards don't even use that rail and don't even have a trace coming from the -12V pin on the ATX connector.

And +12V is so critical that if it were actually 9V, your PC wouldn't function properly.

I've found the eVGA 680i motherboard's health monitoring is the worst I have EVER used. With a 1kW PSU on it, I was getting 11.4V on the +12V while only pulling about 300W from the wall. When tested with a DMM, I read 12.2V!

NEVER... EVER... trust software readings.

[wintermutemd]

Thanks for the info and fast replies. It'll be monday before I can get into our lab and snag a DMM and will check then. I fully agree with you on the 680i I wish I had waited instead of jumping in. My problem is that anytime a graphics intensive application is run (Battlefield 2142, WoW, Doom3, F.E.A.R, etc) The application will randomly crash usually giving a memory exception fault somethign similar to "The instruction and 0x006EB7C0 referenced memory at 0x100015A8. The memory could not be written/read" the addresses change of course. Temps are all within normal ranges. I've run Prime95 for 10+ hours, Memtest86+ for 5+ hours, 3dmark passes no problem (can't loop it though).

I contacted EVGA and they advised it sounded like a PSU issue since the OCZ is dual 12V rails at 18A a piece stating the card requires a minimum of 30A. Does this sound like a reasonable diagnosis or am I chasing a flaky Mobo for no reason?

[jonnyGURU]

I contacted EVGA and they advised it sounded like a PSU issue since the OCZ is dual 12V rails at 18A a piece stating the card requires a minimum of 30A. Does this sound like a reasonable diagnosis or am I chasing a flaky Mobo for no reason?

The card does not require 30A. The card requires that you have a PSU with at least a total of 30A on the 12V rails. That's 30A for the entire system. The card can't physically use more than 18.75A and I've used 600W PSU's to power similar builds with TWO 8800GTX cards.

Could it be a PSU problem? Sure. But it could also be a bad mobo, RAM, video card....

[EsaT]

...the OCZ is dual 12V railsGameXstream is quad (virtual) rail PSU.

I think there's good chances it's EVGA which is "throwing wrenches in the works".
http://enthusiast.hardocp.com/article.html?art=MTI0NCwxLCxoZW50aHVzaWFzdA==



And +12V is so critical that if it were actually 9V, your PC wouldn't function properly.Or propably wouldn't even boot.

[ianm2]

you mean deliberately messing things up for people?

strange...motivation for that?

it sounds like the same old thing to me of simply rushing things out/bad implementation/bioses, without proper testing, and letting consumers do the fault finding for them, appalling practise.

[jonnyGURU]

Also keep in mind that there is a BIOS and patch for your board.....



NVIDIA nForce® 680i SLI “Designed by NVIDIA” Motherboard

Some EVGA customers using the EVGA nForce 680i SLI motherboard have reported experiencing disconnect or write error issues with SATA disk drives. To address this, we have worked with NVIDIA to release a BIOS update for this motherboard that eliminates this bug.

EVGA and NVIDIA strongly recommend that all customers upgrade their motherboards to this new BIOS, regardless of whether or not they have experienced this bug. For the best interest of your customers it is also recommended that you email blast your customer base who purchased the EVGA part# 122-CK-NF68-xx, notifying them of the below “hotfix” and encourage them to upgrade the BIOS, which is as simple as clicking the “Check System Now” link below and running the utility.

This NVIDIA hotfix will automatically check your system to determine if an update is required. If an update is required, it will prompt you to install the hotfix, and upon completion it will shutdown your system. The hotfix will preserve all of your current computer settings, and will likely improve your system stability.

This update is related to SATA disk drives and system instability. System instability is observed in the following ways (not a complete list):

• Random application shutdown
• Corrupted boot drive
• “Blue screen of death”
• Corrupt data

This update will improve system stability and prevent future stability issues related to SATA disk drives on those systems requiring this hotfix.

Step 1: Check System Now (http://us.download.nvidia.com/downloads/evga/nvidia_hotfix_121906_evga.exe) to determine if your system requires this hotfix.

Step 2: Click Here (http://www.evga.com/support/680i/680i_instructions.asp) for instructions on how to use this utility


Thank you for your support






Please find below directions and download link for the NF68 P23 Beta 2Bios from EVGA Floppy Method

* Download the following file:ftp://ftp.evga.com/BIOS/NF68_P23BETA2.exe (ftp://ftp.evga.com/BIOS/NF68_P23BETA2.exe)
* Insert blank floppy disk into drive and run executable
* Restart your PC and set the floppy drive as the primary bootdevice
* The BIOS will automatically be updated, after the update restartyour PC and clear CMOS NVIDIA System Update Tool
* Download the following files: http://us.download.nvidia.com/downloads/evga/updatecenter.exe (http://us.download.nvidia.com/downloads/evga/updatecenter.exe) and ftp://ftp.evga.com/BIOS/p23beta2.bin (ftp://ftp.evga.com/BIOS/p23beta2.bin)
* Select browse and locate p23beta2.bin
* Select load bios
* Power off PC completely, on reboot load defaults in BIOS


I haven't checked to see if this BIOS is no longer beta or not, but it's the BIOS I'm using and I have zero issues.

[wintermutemd]

Yeah I'm flashed up to P23 final did that before even plugging in a hard drive :)

I'm going to try the new PSU when it gets here (not really sounding like my problem anymore but we shall see) and if the issues persist I'm torn between trying for a refund on the board with Newegg and getting a Bad Axe 2 or RMAing for a replacement with EVGA and praying :(

This site is awesome as well as the boards for EVGA and I really do appreciate the info provided so quickly by you guys

~Wintermute

PS.
One day I'll learn how to use punctuation just not today =)

[wintermutemd]

OK so I grabbed a DMM off a friend today and checking a Molex and the PCIE connectors both reported 12.4V so unless I'm getting an unstable amperage somewhere the PSU is not the problem. The DMM I have can only go up to 10A unfused so can't check that without carrying my entire rig into the lab :o

I'm thinking known reliability is going to be a bit more important to me at the moment and with no current plans to use SLI I'm just gonna RMA the 680i for a refund if Newegg will let me and grab a Bad Axe2 for the time being until the 680i gets ironed out. On a side note A-Bits 680i is looking good and should be out soon.


EDIT: Stupid question...Could I not attach the probes to the molex connector and watch the DMM while "gaming" which is where my problems occur. If it is an unstable current it should reflect in the voltage reading correct?
~Wintermute

[Oklahoma Wolf]

EDIT: Stupid question...Could I not attach the probes to the molex connector and watch the DMM while "gaming" which is where my problems occur.

Go right ahead.

[wintermutemd]

OK so it is not a power supply problem under graphics load dropped from 12.42 to 12.40 on the molex. Would this be indicative of what is happening with the PCIe power as well? If so its time to RMA the board and see what happens.