BSOD and Random Freezing/Restarts that doesn't create a BSOD (resolved)

[Brainonska511]

Specs:
Asus M4A785TD-M Evo AM3+ Board
Athlon II X4 635
8GB DDR3 1333 (4x2GB)
2 WD 500GB Caviar Green drives (one is just backup)
ATI HD5770
Windows 7 64-bit Home Premium

For almost a year, this computer was working without any problems. Mid-May, I added an internal card reader, 2 extra sticks of RAM (to go from 4GB to 8GB), and an aftermarket HSF. Everything continue to run without any problems as far as I can remember.

During Memorial Day weekend, I upgraded to Catalyst 11.5. I didn't use it much that weekend, but the following week it was acting really funky - freezing (such as when it was going to sleep or waking up), a few BSODs (0x0000001e and 0x0000003b errors), so I figured it was the ATI driver, since that was the most recent change. I uninstalled it and reverted back to 11.4 - still having problems.

To investigate further, I ran the Windows 7 memory utility and memtest86. The RAM passed with no errors in either one.

I then reformatted the machine - that went without any problems, what-so-ever. Installed the drivers again (using what I had originally before any upgrades + Catalyst 11.4). Then the problems appeared to start again. I saw freezing at random points (during photo screen saver, installing a printer...) and saw "Controller error on \Device\Harddisk6\DR7" in the event viewer, so I thought maybe the SATA cable was loose - replaced it with another one and made sure it was all secure.

Right after I did that, I got a STOP 0x0000001e (all 0x00000000...) error and a STOP 0x00000024. I'm running the Western Digital hard drive diagnostic tool, but I don't know if it was that.

I've also had it just cut out and restart like someone pressed the reset button with no BSOD. I don't think it's overheating though. After I installed the HSF, I ran Prime95 and it was fine - got up to ~43C (in CoreTemp) and idles at ~24C. Most Windows activities, it barely breaks 30C. And it did freeze once in the BIOS, but for the most part, it POSTS fine and I was trying to open the 'generic BIOS help' (F1 in BIOS); but I haven't seen that happen again either.

Any suggestions?

Things I'm thinking about doing:
-Run memtest86 overnight instead of just for 2 hours.
-Remove the new RAM and the card reader and see how things go
-Reinstall Windows on the second WD drive or at least on the first (if it passes the WD diagnostic)

Edit:
So far, it seems to be running crash-free in safe mode as the WD diagnostic is running. Could it still be a device driver issue?

 

[Brainonska511]

Hmmm... maybe it is a RAM problem? I just started running memtest86+ again, around ~8000M, it shot out 73000 errors in just 2% total pass.

Though, I think I need to take out the RAM and test the newest sticks first, one by one.

 

[Ultralight]

Checking one stick at a time is the way to go. Also, you might want to check the memory voltage set in the BIOS and compare it to the manufacture's specs. Do they match?

 

[Brainonska511]

According to my motherboard manual, DDR3 runs at 1.5V and A-Data lists the RAM specs as 1.5V as well. I'll open up the BIOS after another pass to confirm that it says 1.5V. I've never actually looked though since it was running great with 2x2GB.

 

[denis280]

Just do like you said.run one ram at the time and see. the way you telling us with memtest it as error already.

 

[Brainonska511]

Just do like you said.run one ram at the time and see. the way you telling us with memtest it as error already.

Yeah, that's what it looks like to me too. I guess I should have let it run more when I originally ran it. Would have saved me some of this run-around.

I'll just test the sticks: newest to oldest - first one is going right now: 2 passes, no errors, on pass #3.

Pain in the butt. I'm surprised it didn't manifest sooner - like shortly after I installed the extra sticks.

 

[Brainonska511]

Ran the first new stick overnight - no errors in DIMM slot 1. Running the second new stick now in the same slot and I'll see what it says when I get back home today. I'll then test the original 2 sticks (though, I'm confident that those are okay, seeing as they ran problem free for 11 months on their own). Probably have to go slot-by-slot as well to rule that possibility out. It takes so long...

I've seen on a few other forums that it might be a RAM voltage issue. If it isn't giving problems with individual RAM sticks and only throwing up errors when all 4 slots are populated, could bumping up the voltage from 1.5V to 1.65V solve the problem, even though the board's and the RAM's rated voltage is 1.5V?

 

[Brainonska511]

Both new sticks check out fine. I'm fairly confident that the two old ones are also fine - I'll test those later if need be. Right now, I'm testing the two new ones together in slots 1 & 2. Then I'll move them to 3 & 4 to make sure that those are operating properly as well.

Edit:
The newest dual-channel kit was good on slots 1 & 2 (almost 3 passes). Now I'm running it on slots 3 & 4. Maybe I'll try lowering the clock to 1066 when I put 4 sticks back in before trying to play with DRAM voltage or NB voltage.

 

[denis280]

With asus you should have.amd overdrive use it too do your setting.sure a little bit more voltage wont hurt.but not to high.

 

[Brainonska511]

With asus you should have.amd overdrive use it too do your setting.sure a little bit more voltage wont hurt.but not to high.

Yeah. I'm going to try downclocking the RAM to 1066 (and maybe then 800) first before I overvolt to 1.55-1.65V on the DRAM. Running memtest86+ now with all 4 slots populated (ran fine with only 2 slots populated and the newest kit checks out okay).

I saw an errata for the Athlon II/Phenom II where 2 Dimms per channel at 1333 could cause problems and it relates to C2 stepping AM3 CPUs, which the 635 is part of .

Edit:
Looks like there are C3 stepping Athlon II X4 635 cpus, but I got a C2 one:
Ahttp://detonator.dynamitedata.com/c...com/Product/Product.aspx?Item=N82E16819103702DX635WFGIBOX: http://www.cpu-world.com/CPUs/K10/AMD-Athlon%20II X4 635 - ADX635WFK42GI (ADX635WFGIBOX).html (C2; what I have)
ADX635WFGMBOX: http://www.cpu-world.com/CPUs/K10/AMD-Athlon II X4 635 - ADX635WFK42GM (ADX635WFGMBOX).html (C3 stepping; errata I'm experiencing doesn't apply)

 

[Brainonska511]

Downclocked my RAM to 1066 and ran memtest overnight: no errors this time either. I'll put my hard drives back in tonight and reinstall Windows and see how it goes from there.

 

[Brainonska511]

It was a little scary for about an hour:

1) I started to reinstall W7 on my hard drive, memory clocked at 1066 (4x2GB DDR3 1333 sticks) - it froze.
2) Opened up computer, figured I might just pull the graphics card too and use the integrated graphics. Tried to restart, nothing on screen. Used the jumper to reset the BIOS. Posts and I reset the BIOS to what I want.
3) Restart again - just makes lots of beeping (I think 5, which suggested CPU problem)
4) Shut everything down again. Disconnected card reader from USB header; pulled processor, checked for damage on socket/chip (none), cleaned off the paste, reapplied paste to my HSF and reseated it all. Cleaned graphics card "gold finger" (asus' words) with some isopropyl alcohol on a coffee filter.
5) Set the BIOS back to what I wanted it all to be and reinstalled Windows.

So far, so good. Got SP1 installed and putting on some drivers. Keeping my fingers crossed.

 

[denis280]

Sounds Good so far.

 

[Brainonska511]

So far, it's been good for the last day and a half. No freezing entering or leaving sleep, no random restarts, no BSODs, no freezing while installing stuff... Hopefully, it stays this way. I am going to keep that card reader disconnected though - just on the off chance that it was also causing screwy things. I have an external USB one for the few times I need one.

Just my luck to get stuck with a BL-C2 stepping Propus. It's been running fine though at DDR3-1066. I don't notice any difference between DDR3-1066 and DDR3-1333 speeds.

Probably post back in a week to close this off if nothing else goes wrong.

 

[JimKiler]

Checking one stick at a time is the way to go. Also, you might want to check the memory voltage set in the BIOS and compare it to the manufacture's specs. Do they match?

You mean only have one stick in and run the memory test or have one stick in and see if windows crashes?

 

[Brainonska511]

You mean only have one stick in and run the memory test or have one stick in and see if windows crashes?

I did 1 stick at a time while running memtest, then added back more to make one full kit - ran again; moved it to the next slots, ran again, then had all 4 in and ran a final time.

 

[Brainonska511]

One last update (in case someone ever stumbles upon this thread in their own search): been running now for almost a week and a half with no weird restarts/BSODs/failure to POST. Whatever was wrong seems to have been fixed by what I did...

 

[denis280]

Sometime it is weird. you can get a big problem fix in no time.but with all new mobo out now. it always the little one that make you say all those BAD WORDS.