Unstable Ivy Bridge system - voltages?

[MIDIman]

I've had an Ivy Bridge setup for a few months now and it just started giving me problems.

Silverstone GD05
i5-3570k w/ Scythe Big Shuriken 2
Kingwin Lazer Platinum 650w
Gigabyte Z77MX D3H
SAMSUNG 4x4GB DDR3 1600 lowprofile
Vertex3 120GB
Vertex 60gb
Seagate 1TB
WD 2TB

System has been working great for two months. I installed steam recently and I updated the Intel USB3 and HD4000 drivers this evening. Ever since then I've been getting random reboots. When the reboot occurs, typically the system won't even make it past the BIOS and I need to reset the CMOS

I have since updated to the latest bios and put it at the Optimized Defaults - no overclocking. I have also removed the new Intel graphics driver and replaced it with the old one. The system was a bit more stable after that, but I just got another reboot after about 30min. I'm uninstalled and reinstalling the USB3 one now.

Any thoughts?

Temps are 28C idle according to CPUID. Temps jump to 50C when I start Prime95. 60C after a few minutes.

Here's the voltages according to CPUID - are these OK?
Vcore 1.07v
VIN1
+3.3 2.05v
+5 3.45v
-12 -4.1v
-5 6.1v
+5 vcch 2.84v
vbat 1.57v

 

[pantsaregood]

If your 3.3v and 5v rails are really running that low, your PSU is amazingly bad. If not, run memtest.

 

[Prey2big]

Monitor your voltages in your BIOS instead. I don't belieave your voltages are that off.

 

[Rifter]

Screw software dude get a multi meter and find out what your voltages really are. Because if they are that low then thats your issue.

 

[ShintaiDK]

Its not a temp issue.

Are those 1.07V the CPU voltage at load or idle?

Tried downgrade the drivers again?

 

[MIDIman]

I have uninstalled the Intel graphics. I was still getting crashes after only reinstalling the graphics driver from a previous version I had before. I now have the USB3 drivers completely uninstalled (unknown device in device manager). No problems yet. So maybe that's the culprit. If after a week or two I have no crashes I'll try reinstalling the USB3 driver I was using previously.

This is the USB3 driver I had installed:
http://downloadcenter.intel.com/Det...ional, 64-bit version*&DownloadType= Drivers

The version before was either 1.0.3.214 or 1.0.4.220 - I'm fairly sure it was the latter.

For Intel 4000 graphics I went from 15.26.8.64.2696 to 15.​28.​0.​64.​2792. I'm now back on 15.26.8.64.2696.

I had also update to the latest Chipset INF - forgot about that.



I checked the BIOS concerning the power voltages and it seems to be OK. It only reports a few:

CoreV 1.140 (is set to this when using Auto)
DRAMV 1.368 (which is what its manually set to - was reporting 1.2 when on auto. My RAM is 1.3v)

+3.3 = 3.363
+12 = 12.312

It seems that the core voltage reported in the OS is definitely a bug according to this:
http://www.anandtech.com/show/5850/gigabyte-gaz77mxd3h-review-z77-and-microatx/3

Even if the CPU was set at 1.200 volts, I would still get the 1.068 volts reading (even at load). This means that Gigabyte are artificially putting in a layer between the CPU and the OS and manipulating the results as it passes through to the OS.
Read more at http://www.anandtech.com/show/5850/gigabyte-gaz77mxd3h-review-z77-and-microatx/3#bZQRPUWWCluAvdQT.99

 

[ShintaiDK]

USB driver error then it seems. Who would have thought.

 

[Termie]

Do you have any USB3 devices connected? I'm surprised it would give you errors if you weren't actually engaging the driver, even if it was installed.

By the way, Gigabyte is complete fail. I remember reading that review and thinking Anandtech should never have given GB a pass on that. Big flaw in my opinion, and their last round of boards was equally messed up.

 

[ShintaiDK]

I use the latest USB3 driver btw. No errors.

But I only got mouse+keyb attached to USB3. By pure accident as well.

 

[BonzaiDuck]

The advice already given or the diagnosis of likely causes is probably sound.

I'm just not all that happy with your PSU choice. Kingwin -- I bought one of those in late 2002, to build a P4 533FSB system. It was a 500W unit I'd picked up at the periodic computer-fair at the Pomona Fairplex.

It didn't take long before I absolutely had to replace it. Which didn't help much -- I replaced it with an equally "inexpensive" ULTRA model.

Then I caught up on the pros, cons and virtues of some PSUs over others, and never again tried to "save money" by using a $35 PSU.

But . . . . jus' a minute . . . . I'll see if I can find your PSU with a web-search . . . OK . . . there's a $50 650W model, and the "Lazer Platinum" which I see at XOxide for $150+. Even if I had $150 to spend (and I usually do when it comes to PSUs), I wouldn't buy a "KingWin."

For that price, you might have had a 750W "Gold - 80plus" Seasonic, or any number of PSUs veterans here would recommend. Albeit -- the XOxide page does show a 5-year warranty -- a partial indication of reliability. But . . . . well . . . I won't say any more . . .

And . . . for the 2nd and 3rd gen processors and chipsets, Gigabyte hasn't taken prizes recently, despite the fact that I've got three Gigabyte boards running in this house as I speak . . . .

As per another's earlier post, I'd drop back to a 2x4GB RAM configuration and do some serious memory testing -- HCI Memtest would do it, and I'd use the CD-bootable version, to eliminate a corrupted OS as the problem. Perhaps -- test them in pairs, then test single modules if you find a pair that fails after so many iterations. Even Samsung and G.SKILL occasionally ship defective units -- it's a statistical likelihood over thousands of kits and even the most stringent but practical QC practices.

 

[VirtualLarry]

Even Samsung and G.SKILL occasionally ship defective units -- it's a statistical likelihood over thousands of kits and even the most stringent but practical QC practices.

I have it on good authority, that Gskill ships RAM that has up to TWO errors during their burn-in phase, as "good" RAM.

After finding that out, I wonder for how much longer I will purchase GSkill RAM. No wonder their prices are often some of the cheapest around.

N.B. - I already purchased 80GB+ of GSkill DDR3 over various newegg sales months ago when RAM was really cheap. Didn't have any problem with any of it I've used so far.

 

[BonzaiDuck]

I have it on good authority, that Gskill ships RAM that has up to TWO errors during their burn-in phase, as "good" RAM.

After finding that out, I wonder for how much longer I will purchase GSkill RAM. No wonder their prices are often some of the cheapest around.

N.B. - I already purchased 80GB+ of GSkill DDR3 over various newegg sales months ago when RAM was really cheap. Didn't have any problem with any of it I've used so far.

Never worked in the silly-cone chip business, but they would have to use statistical samples for burn-in tests, I'd think.

You would, of course, remember the Great Crucial Crisis, when we bought Ballistix or Tracers, and they went south after six months. I think Crucial had a crisis with the barrage of RMAs. After that, I first wondered "Who is G.SKILL?" I was pleased, if only for the contrast with the Crucials. Oh-- I think I picked up a 2x2GB kit of Corsair Dominators, and those, too, were fine -- OK -- reliable -- reasonably "adjustable" at the voltage spec, and the voltage spec had a "warranty tolerance."

But it also gives me comfort to get helpful, swift e-mail responses from a G.SKILL tech-rep when I have questions.

It's a very competitive market, and sooner or later people will migrate to this or that . . .

I DID have a G.SKILL module -- a DDR2-800 -- go bad after six months. The fast RMA replacement was a contrast to an experience I had with Crucial.

 

[MIDIman]

I'm fairly certain the USB3 driver was the problem. After uninstalling the driver, all has been working perfectly.

After this weekend I'm going to reinstall one of the old ones I have in my backup.

I do use an external USB3 Seagate 3TB, but it has its own issues and was not hooked up at all last night when I was having my problems. I pulled all of my HDs except for the two SSDs.

re: the Kingwin PSU "Lazer Platinum" series is actually very highly rated. Look around at reviews and you'll see what I mean. I particularly bought it for both stability and noise. There is a switch on the back of the PSU for a Normal and Eco mode - when in "Eco" mode, the fan on my model will actually not turn on until a given temp is reached. That said, all of that has been disabled while I've been troubleshooting and all of my fan settings are on max.

 

[2is]

Screw software dude get a multi meter and find out what your voltages really are. Because if they are that low then thats your issue.

This... Software readings are worthless, BIOS isn't a whole lot better either. Use a multimeter and you can also test the voltages under load and idle states.

I highly doubt your system would even POST if those were your actual voltages.

 

[BonzaiDuck]

I'm fairly certain the USB3 driver was the problem. After uninstalling the driver, all has been working perfectly.

After this weekend I'm going to reinstall one of the old ones I have in my backup.

I do use an external USB3 Seagate 3TB, but it has its own issues and was not hooked up at all last night when I was having my problems. I pulled all of my HDs except for the two SSDs.

re: the Kingwin PSU "Lazer Platinum" series is actually very highly rated. Look around at reviews and you'll see what I mean. I particularly bought it for both stability and noise. There is a switch on the back of the PSU for a Normal and Eco mode - when in "Eco" mode, the fan on my model will actually not turn on until a given temp is reached. That said, all of that has been disabled while I've been troubleshooting and all of my fan settings are on max.

That's interesting news about "Kingwin." My experience with that brand occurred in 2002/2003, so it may not be very relevant today.

 

[Termie]

I'm fairly certain the USB3 driver was the problem. After uninstalling the driver, all has been working perfectly.

After this weekend I'm going to reinstall one of the old ones I have in my backup.

I do use an external USB3 Seagate 3TB, but it has its own issues and was not hooked up at all last night when I was having my problems. I pulled all of my HDs except for the two SSDs.

re: the Kingwin PSU "Lazer Platinum" series is actually very highly rated. Look around at reviews and you'll see what I mean. I particularly bought it for both stability and noise. There is a switch on the back of the PSU for a Normal and Eco mode - when in "Eco" mode, the fan on my model will actually not turn on until a given temp is reached. That said, all of that has been disabled while I've been troubleshooting and all of my fan settings are on max.

Great job trouble-shooting this. I'm finding that more and more people are having trouble with USB3. Lots of driver issues, lots of incompatibilities, evidence of speed much lower than it should be, etc.

Perhaps USB3 wasn't actually ready for prime-time. Given that over USB3 your average hard drive transfer speed only doubles (getting nowhere near the speed of even SATA2), it seems there's really very little argument for using it.

 

[BonzaiDuck]

Great job trouble-shooting this. I'm finding that more and more people are having trouble with USB3. Lots of driver issues, lots of incompatibilities, evidence of speed much lower than it should be, etc.

Perhaps USB3 wasn't actually ready for prime-time. Given that over USB3 your average hard drive transfer speed only doubles (getting nowhere near the speed of even SATA2), it seems there's really very little argument for using it.

It's not something I would miss much, but since the 2nd gen mobos include the feature, I thought I should have it "operational." I found a four-port USB3 hub for a 3.5" drive bay, and hooked it up directly to the "two-port" plug on the motherboard. I never really had a problem with the drivers . . .