New Build Up and Running Today -- WindowsXP Spontaneously Reboots Periodically

[Fluidstream]

Build:

Asus Crosshair V Formula Z
Kingston HyperX DDR3 @ 2133 MHz
AMD FX-6300 Vishera six-core
BFG Tech GeForce 7950 GTOC
Thermaltake Smart-M Series 850-watt PSU (Purchased in July 2012)
WinXP SP-3 (no updates installed yet)

I just put this system together on Friday, and after getting an invalid product key for Windows 8 from Newegg.com, I decided to install WinXP with a slipstream SP3 disc. Everything works nicely except for the periodic reboots. I disabled "Automatically Restart" in the advanced system settings so I could get a blue screen, but I get no blue screen... just a sudden reboot.

Now here's what happened. The memory I'm using (Kingston HyperX 4GB x 2 - KHX2133C11D3K4) is listed in the QVL and is in the manual that came with the motherboard.

I had the BIOS set to run the memory at 2133MHz, timing was 11-12-11-30 as per the book and QVL, and I had the voltage set to 1.65v. After the spontaneous reboot, I went back into the BIOS and noticed that all my memory timings and frequency had been set back to the defaults I remember seeing when I first turned the computer on (1333MHz and 9-9-9-24). And the voltage had been lowered to 1.60v, which is still higher than the 1.38v or so that was the default.

I'm wondering what the BIOS is doing here. Is it running some kind of analysis and lowering the voltage on my behalf? The book specifically lists the voltage when running four DIMMS. I have four DIMMS, but I only have two installed because the book said I should get a premium memory cooler if I want to run a full memory load. I can't get a memory cooler right now because I spent all the money I had to spend getting the other hardware and Windows 8. Should I change the voltage to something else with only two of the four DIMMs installed? Or maybe I should go ahead and stick the other two DIMMs in and see if that works? I've never had a memory cooler before, but I've also never built a computer with all four DIMM slots filled. I'm not sure how to proceed.

So I'm wondering if these BIOS settings are possibly causing my reboots. Since I can't get a blue screen for an error code, for now, this is the only thing I can think of that could be causing problems. When the BIOS was set to the defaults, spontaneous reboots happened far more quickly, like within ten seconds of trying to install Windows 8. (By the time I got to installing WinXP, I had configured the frequency, timing and voltage for the memory so I didn't have a reboot problem.)

What suggestions do you guys have at this point for stopping the reboots and/or what explanations do you have for these strange changes to the BIOS?

Thanks,
Fluidstream

[TenBlade]

Try with only one stick of memory at a time, place it in the slot farthest away from cpu.
Try to run memtest on each stick.
Have you tried XMP profile?

[Fluidstream]

Okay, I tried a few different things. I tried one DIMM in A2, then I tried one DIMM in the slot you said, furthest away from the CPU. Then I tried changing the DIMM I was using. Then I cleared the CMOS completely and the only thing I changed was turning off the RoG logo that shows when booting. It just seems to be getting worse. Now it sometimes reboots before Windows even starts loading, or it reboots during the Windows load. I would think that all-automatic would be the BEST settings for stability, but I had more luck running the memory at 2133MHz than 1333MHz, and 1333MHz is its standard speed.

This is driving me insane. Added to that, the motherboard I got (Crosshair V Formula Z) came to me in a Rampage IV Formula box! one of the pins was bent where the Q-Connector goes, and one of the SATA ports looked a little dirty, like a mechanic had touched it.

I'm starting to think that I got a motherboard that was handled by someone else in their house, not a nice, clean one from the factory, untouched by human hands. Maybe the one who handled my motherboard ruined it and send it back to Newegg, and then I got it. I don't know how the Newegg packaging department somehow knew that a Crosshair V motherboard would be inside a Rampage IV box.

Is what I'm experiencing typical? Shouldn't this thing work best with all automatic settings in the BIOS? That's what I was expecting, you know.... just something that works when I turn it on, without having to run through and manually set everything. If I were interested in overclocking, I might be willing to play with those settings, but first I wanted to get stability, which I thought was the first step in even trying to overclock: start with stability, and start bumping it up until you DO encounter instability. But it seems like the slower the settings, the higher the instability. It's so backwards.

As long as I stay in the BIOS, it doesn't seem to complain, but I want to do more with a computer than look at BIOS settings all day.

[Fluidstream]

I haven't tried an XMP profile. Another poster said I should set the memory to run at 2133MHz, but in later research, I found out that was its maximum tested speed. Its standard speed is 1333MHz. Having learned that, cleared the CMOS and made sure everything was on Automatic that had an automatic setting. That's what I want, anyway, for everything to work at its recommended speed. I'm after stability here, not overclocking. XMP seems to run things at a higher clock rate than JEDEC.

So anyway, when i cleared the CMOS and had everything on auto, I had four sticks of RAM in there (before I only had two) because I wondered if the QVL was saying that 11-12-11-30 and 1.65 volts was for running all four DIMMs. Those ratings, according to the Kingston PDF for this memory, are for running the memory at 2133MHz. 1333MHz is apparently the speed I'm supposed to be shooting for, so it's all on automatic now.

The only thing I haven't tried yet (but I'm about to) is setting the BIOS option that says "Load 4GB settings". Well, my modules are 4GB, so perhaps this is the setting for me. It raises the auto-configured 9-9-9-24 slightly, but that's okay.

I'll try just one stick of RAM. For a minute, I thought you were telling me wrong to put the single DIMM in the slot furthest from the CPU, so I took a look at the second manual because I thought it was saying to use slot A1 and B1 when using two DIMMS, but it says to use A2 and B2. The little arrows didn't help on the diagram, and I didn't notice that the text and pointers were pointing to A2 and B2. Still, the diagram doesn't say that having one DIMM in slot B2 (which is furthest from the CPU) is a recommended configuration.

What about disabling the ECC mode in the BIOS? It's funny that the manual says to use non-ECC memory, yet the BIOS for it apparently has settings for ECC. Why have a setting FOR ECC when the motherboard requres non-ECC memory?

The book also says to use a maximum of 3GB when using 32-bit OS. LOL I don't have any memory modules smaller than 4GB. However, I did notice that I have not been able to stay up and running as long with 16GB installed as I was with 8GB installed. With 8GB, I could keep it up and running for about an hour. With 16GB, I could keep it up and running about five minutes.

Maybe that's the issue, then...having more than 3GB of memory installed. I figured if there was more, Windows 32-bit would just ignore it and pretend it isn't there. I didn't think it would cause stability issues. 😞

[Fluidstream]

Here's another tidbit of information: while the computer would reboot at any time, it happened more often when I tried to load Farmville 2 or Stormfall: Age of War. I wonder if my graphics card could be causing some of this. I'm expecting a new one tomorrow (GeForce GTX 650)....at least I hope it'll be new. I ordered it from Newegg.com, too. But after getting Windows 8 with a product key that won't work, and a motherboard with a bent pin and a greasy finger print on the SATA port (and the CMOS battery (don't those guys wear gloves for assembly at the factory?)) all in the wrong box, I'm wondering if I'm going to get a Radeon card in a 3dfx Voodoo2 box.

[TenBlade]

If that were to be a brand new mainboard, you´ve been had. It´s supposed to be dust free and super clean when you get a new board. Sending you a board in the wrong box says it all. Smudges and bent pins, RMA immediately, and make it clear to them, that this is unacceptable, their reputation is at stake here.
And you send them some clear digital close ups of those faults, and the wrong board lying in the box (or vice versa)

[Fluidstream]

Yes, I agree, TenBlade. I made the mistake of not RMAing the Abit Fatal1ty AN9 32x I had before this one, and as a result, I didn't get full enjoyment out the six years I had it. It was remarkably stable for a motherboard with a bad capacitor the whole time I had it, but I thought the little black spot on the capacitor was just ink from a magic marker or something. I thought maybe they marked one capacitor in whole trays of capacitors as some that had been inspected, and I just happened to get the one on my motherboard that was the one marked, right next to the PCIe slot I was using for my graphics card. I didn't realize until I upgraded my monitor from a 22" CRT to a 27" LED in November 2012, and began running 1920 x 1080 instead of 1600 x 1200, that that was, indeed, a bad capacitor. Two days after I upgraded the monitor, my computer started turning itself off suddenly. Switching to a smaller monitor (15" 4:3 LED) didn't help. I took the power supply back to have it tested where I bought it in July 2012 (Best Buy) and it tested good, but the machine they normally use to test it was broken. He used some hand-held tester that still said it was supplying all the right voltages and whatnot. When I gutted my case was when I noticed that the capacitor with the black mark on it was then bulging. A computer repair technician at a local shop confirmed that it was bad just by looking at it with me.

The only other possibility is that the bad capacitor may have done something to my GeForce 7950 GTOC. A new graphics card is coming in today from UPS. I'm thinking about sticking it into test it just to see if it helps the problem. I was determined to get THIS motherboard to work so I won't have to wait for a new one to be shipped back to me, but if it won't even run with default/automatic settings in the BIOS, I feel like I'm not going to have any luck. Plus I'm worried about THIS motherboard doing something nasty to my new graphics card. I won't be able to tolerate any more than one spontaneous reboot before I give up on the motherboard and send it back.

The only good part is that the little sheet of plastic was on the Southbridge heatsink with the RoG emblem on it. Seems like anybody trying that board out would have pulled that off, though that person could have stuck it back on there before sending it back.

Anyway, if it's not normal for the motherboard to behave this way with all automated settings in the BIOS, I'll consider it a bad motherboard and RMA it. It runs better if I boost the memory to its maximum tested speed of 2133MHz, along with the proper voltages and timings, but it's supposed to run at a standard speed of 1333 MHz and it won't even complete a launch of Windows at 1333MHz. At 2133MHz, it'll run for an hour, maybe an hour and a half at best, before it reboots.

[Fluidstream]

Well, here's the deal. I just got my new graphics card (MSI GeForce GTX 650) and I installed it. Amazed I am, but I feel like it's too soon to relax and breathe a sigh of relief. With this new video card, I have had the computer up and running WindowsXP with default/automatic BIOS settings, and so far, it has not rebooted on me. With my old GeForce 7950 GTOC, I could barely get Windows started at all before I'd get a reboot. I had to move the RAM way up to 2133MHz, and adjust the voltage and timing, to get Windows to stay up for an hour. This is a good sign. The 7950GT was in a motherboard that died because of a bad capacitor, and it looks like that capacitor might have damaged the graphics card. I'm crossing my fingers, toes, eyes, legs, arms, and I'm trying to cross my nose holes in hopes that this finally fixes the problem. I'm still not thrilled about having a motherboard that came in the wrong box, but if it works and it's stable, I would really rather not send it back because I might end up with a new one that actually IS bad.

Oh, one other thing I changed, which I'll check in a bit...after installing the video card, I disconnected the extra rear USB ports I installed. I suppose it COULD be that, but I'm more inclined to believe it was the video card at this point. When it was rebooting, the fan on the graphics card was revving like a race-car.

I haven't yet tried to load anything that was previously causing a reboot every time, and that was Farmville 2. I'm almost afraid to because I don't want to be re-introduced to disaster. Eventually I will... but right now, I just want to enjoy gazing into my own monitor so I won't have to look at my roommate's damaged LED with all the scratches and the not-nearly-as-bright-and-brilliant screen image.

[Fluidstream]

Oh wow. I just checked HWInfo32 and it shows my GPU idling at 26ºC. My GeForce 7950 GTOC idled at about 57º-60ºC.

[Fluidstream]

Ugh... The reboots have returned! it just rebooted on me about three minutes ago. I let Windows XP load again, and then tried to launch a browser to come here. Before the page even loaded, it rebooted again! I don't get it. But this time, it stayed up for 3.5 hours, which is a record high for this rig. Now I don't think it was the video card anymore, which is good because my GeForce 7950 GTOC was a heck of a lot better performer than this nasty GTX 650. I now wish I hadn't bought it, but maybe its DirectX 11 support will improve it if I can ever get Windows 8 installed.

So fast memory speed and timings makes it run longer with a 7950GT, automatic configuration of everything in the BIOS makes the whole system unusable since Windows can hardly load before it starts another reboot cycle, a new GeForce GTX 650, at least at first, made the auto-configuration work much longer. I haven't tried changing the BIOS because I know it SHOULD run best with automatic configuration. I shouldn't have to adjust everything just to get it to run and stay up, especially when my old Abit Fatal1ty could stay up indefinitely running with everything auto-configured. I only had to turn it off when there was a power failure more than 15 minutes long. I guess I'm back to thinking of an RMA on the motherboard unless someone here has other ideas.

By the way, getting the thing to turn off is actually kind of a challenge. Sometimes just tapping the power button turns it off, sometimes I have to hold it down to get it to turn off, sometimes holding it down also causes it to immediately power back on, sometimes tapping the power button causes it to immediately power back on, and sometimes pressing the power button does nothing at all. My case has power leads for the Q-Connector, of course. It's hard to know what colors stand for positive or ground, but after connecting everything the best way I could understand, all the labels for the leads were facing in the same direction.

I'm just sick.