Unstable Asus M4A89GTD PRO/USB3

[SuperDuty]

Hi,

For the past few months, my computer has been very unstable, freezes, various BSODs all leading to a hardare problem, sudden black screens, all at random moments, and sometimes during Windows startup. These have worsened in the recent weeks. I've even had some corrupted psd files that I was working with recently (not at the moment of the crash, so that's weird).
I did not make any hardware change before it all started.

Here are the main components :
- Phenom II X4 965BE cooled by a Cooler Master Hyper 212 Evo
- Asus M4A89GTD PRO/USB3
- 2x2Gb Kingston HyperX 1600Mhz
- 2x4Gb Kingston HyperX 1600Mhz
- Samsung 850 Pro 128Gb
- WD Caviar Green 1Tb


I've tried :
- temperatures are all good according to HWMonitor and Speedfan.
- memtest86+ : no problem
- Prime95 for hours : no problem
- removing everything but the CPU, RAM and main hard drives, and outside the case
- using various video cards : GTX480 at first, then HD5970, then HD5850 and also the motherboard HD4290 chipset.
- swapping power supplies with a known good unit
- clearing all drivers and installing fresh ones
- I did upgrade the BIOS earlier this year but the latest version made the computer very unstable, so I switched back to the previous version and it worked perfectly until around August.

Sometimes the computer will freeze, and resetting it wont work properly : it will start but not boot, no beep, just a black screen and the fans spinning. It takes a few tries before it boots properly, or leaving it off for a few minutes.
Also, it happened several times that after a BSOD, Windows will BSOD again during the loading screen, sometimes three times in a row, and on the fourth try, start correctly.

The most common BSODs are 0x00000124, 0x0000003B and 0x00000050, but there were some other codes as well, and common descriptions are "PAGE_FAULT_IN_NONPAGED_AREA" or "MACHINE_CHECK_EXCEPTION".

So that leaves me to the only conclusion : the motherboard has an issue. It could be the CPU as well but it's quite rare these have problems.

I bought an upgrade kit (i7 4790K and Asus Maximus VII Ranger) because I need it for work, but it's still a pretty good system even after six years, so it's worth trying to repair it.
Where would you start looking ?

I've always taken great care of my computer, cleaning it regularly.

[mikan]

Re: Unstable Asus M4A89GTD PRO/USB3

More likely ram/ram timing issues than it is motherboard, but I have an asus board that does similar things, it freezes at a random time, usually 5 minutes but had it running as long as an hour, then it refuses to reboot, and its definitely not ram related. I think it's because of a broken ground pin in my socket though.

[mockingbird]

Re: Unstable Asus M4A89GTD PRO/USB3

Which PSU are you using?

[SuperDuty]

Re: Unstable Asus M4A89GTD PRO/USB3

I'm using an OCZ ModXStream Pro 500W that is just three years old, and my backup/multi GPU PSU is a HiPro 750W from a Dell Workstation.

I removed the motherboard and found a broken tiny capacitor near the lower PCI-E 16x slot, I probably broke it while installing a graphics card. So I replaced it.

But most of all, I noticed the "Core Unlocker" switch was turned on, I never used this feature and never even tried it so it probably got switched on while I was doing something around here. So I turned it off.

I can't scream "victory" yet but it been running trouble free for a couple days now. Not sure I'm really happy the problem would be so simple, especially since I've been scammed by the seller of the new parts I was supposed to get, meaning I lost money for absolutely no reason...at least I've got a working computer for now. This scammer will get a visit from the police.

[momaka]

Re: Unstable Asus M4A89GTD PRO/USB3

Try new SATA cables.

If problem remains, perhaps one of your hard drives might be going bad.

I am assuming the Samsung SSD is the system drive and the WD Green is for data storage. So perhaps, try removing one of the HDDs (the non-system HDD). The WD Green drives are known to have terrible reliability when WDIDLE is not enabled to keep them from parking their heads constantly. Not sure about the reliability of the Samsung SSD. Hopefully, someone with more knowledge can clue you in on that.

Finally, check your system's Event Viewer and let us know if you find anything you are not sure about or anything that looks suspicious.
On that note, I'm curious if you are running Steam (the game client). I just tried to install it on two older computers last week, and one of them reboots as soon as Steam is started, while the other reboots only when a Steam game is started.

I'm using an OCZ ModXStream Pro 500W that is just three years old

Might be time to replace the caps in that, especially if they are CapXon or Teapo. Also, the primary cap too, if it's not a Japanese one. APFC and CN/TW caps do NOT make a good team.

[SuperDuty]

Re: Unstable Asus M4A89GTD PRO/USB3

I know it's filled with chinese caps, these are made by Channel Well, but I think the primary capacitor is japanese in these. I've cleaned it recently and visually there was nothing wrong.

You're right about the hard drives, the Samsung 850 Pro has a 10 years warranty, those cost a bit more than the 850 EVO for that reason and are faster too. These and the Intel SSDs are probably the most reliable.

I'm running Steam ocasionnaly, but strangely I had a lot less crashes while playing games.



So far so good, the computer hasn't crashed since I turned off the Core Unlocker feature.

[TELVM]

Re: Unstable Asus M4A89GTD PRO/USB3

The Deneb Phenom II's have 4 physical cores at the most, and your 965BE came OEM with all four of them operational, so activating core-unlock in your case could achieve nothing but trouble.

For the record the M4A89GTD-Pro mobos are very good at core-unlocking (provided there are healthy cores to unlock onboard the CPU), mine successfully unlocked two extra cores from a 960T Thuban that has been running like a champ for half a decade @ 4.1GHz on 6 cores. I didn't fiddle with the mobo switches however, just unlocked from BIOS.

. . .

[SuperDuty]

Re: Unstable Asus M4A89GTD PRO/USB3

Intersting, I should grab one of those as a small upgrade. Worth a try if I find one cheap !

I have no idea how I activated the Core Unlocker, maybe it was when I was fiddling with the power cables. But I haven't had a crash since I deactivated it !

[SuperDuty]

Re: Unstable Asus M4A89GTD PRO/USB3

Well after three flawless weeks, the BSODs and freezes are back. BSODs are mostly 0x0000124 which is an hardware error.

Since I've already tried everything, my only options now is to flash the BIOS again or replace the motherboard

[momaka]

Re: Unstable Asus M4A89GTD PRO/USB3

Maybe it's a bad solder joint or BGA issue with the CPU socket or the NB/SB ?

Try running a video or benchmark (maybe Prima95), or something similar, and then firmly press on the motherboard with your finger in various areas around the CPU socket, NB/SB and RAM. See if this makes it unstable or crash.

[SuperDuty]

Re: Unstable Asus M4A89GTD PRO/USB3

Ok I'll try that if it happens again, because I flashed the BIOS to the latest version and the crashes remained.
So I just to try something, I deactivated the Cool n Quiet feature and so far it hasn't crashed again.

[kaboom]

Re: Unstable Asus M4A89GTD PRO/USB3

Ok I'll try that if it happens again, because I flashed the BIOS to the latest version and the crashes remained.
So I just to try something, I deactivated the Cool n Quiet feature and so far it hasn't crashed again.

If that drops the number of active phases in the VRM during periods when the CPU is supposedly idle, it'll generate excessive ripple on VCore.

For a given regulator duty cycle (ie- a given VCore, minus switching losses), as the number of phases drops, ripple magnitude and nodes of ripple peak vs varying levels of Iout increase.

What's worse is that constantly picking up and dropping phases can throw the compensation off if marginally designed. As in compensation-not-compensated for picking up/dropping phases.

So during periods of high-load (memtest, etc), all phases will be active and VCore should be as stable as it gets for a given board. Start screwing around with cyclical loading (normal computer use), and marginal designs show their problems.

This phase-dropping was a false-start. No one ever heard of dropping phases on large motors during periods of "low load." And for a VRM, I'd rather have a solid three phases active at all times, the load large or small balanced across the three legs.

When silly ASUS MBs drop down to one phase, not only does ripple/noding increase, but I2R losses in that singular active phase concentrate. These limitations are why we went to polyphase VRMs in the first place!

Only ASUS would charge money for, and implement, marginal designs as a "feature."

ASUS Xtreme Phase
ASUS 8+2 Phase Power Design

Anytime ASUS lists "X+Y Phase Design," I take this to mean it's a phase-dropping VRM, no matter what they're calling it at the moment... At the least, it's dropping VCore when CPU is supposedly idle; at worst and as far as I'm concerned, it's dropping phases.

Unleashes ultimate memory performances with independent power to core components, while providing fast transient response and stability for the CPU under heavy loading or overclocking modes.

IOW, keeps several phases off and "in reserve" until VCore starts falling below setpoint. When VCore falls too low, we kick the reserved phases in, attempting to bring VCore back up. Bye-bye compensation, hello ripple, crashes, BSODs.


Just keep all phases hot, and regulate VCore with duty cycle! And keeping all phases active keeps inductor current continuous for all phases. Don't throw phases into/out of CCM by picking them up/dropping them.

Another harebrained ASUS "feature."

Any SMPS, can "provide fast transient response" w/o dropout/pickup of phases. It's easier and better that way.

[momaka]

Re: Unstable Asus M4A89GTD PRO/USB3

@ Kaboom: I think they do the whole phase dropping to increase the VRM efficiency at low load. However, you are right - that concept just seems very wrong. What they should be doing instead is probably AEE (Automatic Efficiency Enhancement) Control. I just read about it today and think the idea is quite good.
Link on AEE here:
http://www.digikey.com/en/articles/t...Ma5j2LgNWBE%3D

[kaboom]

Re: Unstable Asus M4A89GTD PRO/USB3

@ Kaboom: I think they do the whole phase dropping to increase the VRM efficiency at low load. However, you are right - that concept just seems very wrong. What they should be doing instead is probably AEE (Automatic Efficiency Enhancement) Control. I just read about it today and think the idea is quite good.

It sounds like the marketing dept and "IT designers" got into the power electronics- a place neither belong.

Based on the input and output voltages, the switching frequency is automatically adjusted to enhance efficiency while maintaining control-loop stability and the output filter

That's what the newer SMPS wall warts do, even some +5VSB supplies do this.

As to what ASUS did/does, it's just turning an aspect of past design, no longer sufficient in today's designs, into yet another feature.

Like those "stainless appliances" that we discussed awhile ago; all show 'n no go, and disguising the same so-so stuff in a new package.

If anything modulates VCore, they should follow intel's lead, from what, 12 years ago: Drop VCore, but proportionally drop FSB, and/or multipliers. It worked for the "super P-3," aka Pentium-M and Core/C2D/C2Q, and even up to the latest I7.

It's not like you can't underclock/undervolt on AMD's platform... Wanna bet there's 100mV of ripple/over/undershooot on that ASUS board when "the feature" is enabled?

[RJARRRPCGP]

Re: Unstable Asus M4A89GTD PRO/USB3

"MACHINE_CHECK_EXCEPTION".

That code exists in XP and the like only. Because Microsoft decided to change the error code, starting with Vista.

With XP, it's 0x0000009C MACHINE_CHECK_EXCEPTION

With Vista and later: 0x00000124 WHEA_UNCORRECTABLE_ERROR