Another ex-58-ud3r power on power off cycle problem

[dbee]

hello, new here. was directed to check you guys having gigabyte support. Nice to meet you all of course, so heres the problem... i have seen many many threads on Gigabyte boards problems with continuous reboot/power cycling. This is a brand new machine. Now, when i first put it together it would power on ,off, then on and stay on. Stopped doing that. i get my OC to just under 3.5 using 165 blck x21 with only 1.25 vcore. I have managed a couple of times to get it upto 172 blck with little voltage increase. After that all i get is continuous reboot. believe me, i have read and eatin every guide out there. So i back down the BLck to 165 again, continuous reboot over and over, and it has never ever done this before at 165. Temps are not a problem at mid 60's under load. here is my setup. And another thing, my brothers identical machine does the same identical thing. I have tried it all. have my uncore 2x+1 of my mem multi, correct mem timings, voltage, many different voltage combos, even upto 1.4vcore just to try for 3.6, upto 1.35qpi , IS SOMETHING MISSING?
I know alot of people have this problem, lets try to help each other. This is driving me mad. I dont give up but now im ready to. Most people have been saying its the memory, but it tests out fine . IDEAS please. Thanks ,Dan.

i7 920 C0 stepping
gigabyte ex-58-ud3r rev.1.0 , just updated f3 to f6 bios, didnt help
OCZ 3x2GB 1333mhz ram at 1.64v
noctua u12p cpu cooler
CORSAIR 750tx power supply
CM RC690 case
WD Caviar Black 500

[dbee]

And it seems like when i save my bios settings to cmos , you know press f10, it shuts off and power cycles instead of continuing loading windows. It used to cycle occasionally, but now it seems to do it nearly all the time. Any thoughts , comments, will be greatly appreciated.

[Grnkjr0]

Hey again dbee

I am hoping someone can help this guy since I am the one who suggested he posted about his problem in here. I have tried helping him out, but with no luck and it seems that something funny is going on with his PC rebooting. I thought it was an Uncore problem but it seems that is not the issue here, but maybe someone has an idea.

Dbee I suggest that you post a full list of your hardware aswell as what settings you have tried and are currently running when it reboots. Further a picture with what it looks like when you are running stable under 100% load, like the one I have posted here now, will help too. Please use the same programs like I did and make Everest show your temperatures when under load. You will find that in the sensor setion.

Dbee if you do not all ready have Everest Ultima Edition, you can get it here:

Lavalys - Comprehensive IT Security and Management

And please show CPU-Z with memory and CPU as well.

Thanks

Poul

[Tyreman]

Don't know your board but...So if you unplug the PS, short out the cmos for a few seconds at the jumper, then additionally take out the battery for a half hour you still can't get in to cmos?

[genec57]

You may need to increase your Vtt voltage.

[dbee]

hey guys, i do have a feeling its a case problem , i saw a couple posts where people have taken everything out the case and boot it up on cardboard making it work fine, and some have the same case as me. i might have bookmarked it somewhere. i have not removed the cmos battery or cleared it from the board yet. didnt want to do so if unless i have to. VTT voltage is fine. when i get to my machine later i will post a full screen showing EVERYTHING. If we can get this to work, it may solve alot of other peoples problems too so i will try anything. im at the point where i do not care if the mobo burst into flames and melts into a pile of gooey capacitors and stuff. Tyreman, i can get into the Bios. Heres what happens, when i save my bios settings and it works, the computer stays on , loads windows. Other times after i save to cmos and exit, the computer completely powers down and powers on off continuously, but it eventually gets to dual bios and i can reconfigure, i now have my settings at the normal 165blck i been using originally , i will post later my screenshots. ANYBODY ELSE USING A COOLERMASTER CM690 CASE??

[Achelon]

Hi,

This is my first post (be kind )but.. I wanted to share some experience and observations that may help you.

When I first assembled my ex-58-ud3r 1.0 system i shorted the board on a metal case stand off. The board went into a 5 second on 1 second off power cycling loop. Removing the short corrected the problem.

When I went into the BIOS to configure settings I noticed a similar failure mode intermittently occurring when specific settings were altered. The failure mode is characterized by power cycling plus all PHASE LED lights on the board burning brightly. After many restarts the board posted and accused me of overclocking when i did no such thing.

dbee, you dont mention the state of the phase led lights, but i bet you will see they are all on.

I believe the explanation for these behaviors you have observed is as follows, although only gigabyte could confirm this:

1) This board has been designed so that some BIOS setting changes require the board to do a simple soft reset, whilst others such as memory timing changes require the board to actually power off and then on again which the board does automatically for you when you save your new BIOS settings.

2) The board contains a hardware watchdog feature. If the board is in an error state for too long (short circuit, CPU missing, memory failed etc) and cannot pat the watchdog before a specific timeout, the watchdog forces a hard reset.

3) The board ignores correct SPD RAM timing values and likes to use more aggressive values, even when the bios is set to "Standard" and not "Turbo" or "Extreme". On some occasions these out of spec timing values cause the RAM to fail to initialize, which is interpreted by the watchdog as an failure condition that must be cleared. Ironically, in some cases the very fault the board is trying to clear, is caused by the board, which defaults to out of spec timing values, pushing components too hard.

The result of this three combination of factors is that in certain cases (hardware fault, or certain BIOS settings changed) the board goes into this power cycling loop, because the watchdog is desperately trying to clear the fault it has detected.

I know this explanation doesn't help you much dbee, but I wanted to provide some background to ppl on what causes the power cycling effect on this model board.

Apologies if all this is already known to you.

Regards,
Achelon

[dbee]

echelon... no need to apologize. I knew none of this, this is all new info to me. Thank you kindly very very much. Even if i did know, your help as well as everyone else' on this forum is priceless. If the phase LED's we speak of are the ones in the top right hand corner of the board, then YES, they do all lite up . im gonna reboot and see if they are or which ones are coming on now since my machine is booting fine. i am also going to try for the heck of it , disconnecting anything conecting the board to the case. This would be looking for a short, but if there was a short, i would not be here on the machine we speak of as we speak since it would not boot period if there were a short. I will also work on my screenshots for you all to view. I know what everyone needs to see. THANK YOU ALL SO VERY MUCH. i could not thank everyone enuff. Now lets get this fixed .

[Achelon]

Originally Posted by dbee

echelon... no need to apologize. I knew none of this, this is all new info to me. Thank you kindly very very much. Even if i did know, your help as well as everyone else' on this forum is priceless. If the phase LED's we speak of are the ones in the top right hand corner of the board, then YES, they do all lite up . im gonna reboot and see if they are or which ones are coming on now since my machine is booting fine. i am also going to try for the heck of it , disconnecting anything conecting the board to the case. This would be looking for a short, but if there was a short, i would not be here on the machine we speak of as we speak since it would not boot period if there were a short. I will also work on my screenshots for you all to view. I know what everyone needs to see. THANK YOU ALL SO VERY MUCH. i could not thank everyone enuff. Now lets get this fixed .

Hi,

Yes, those leds are the phase ones.

To clarify, I do not know for a fact that my explanation is absolutely 100% accurate but I believe it to be the reason why so many people end up seeing power cycling issues on this model even though their problems might really have different root causes. As you say, a short is probably unlikely in your particular scenario, but certainly I would suggest check ram timings. Leaving all the timings set to auto in BIOS is recipe for big headache from what I saw. I mentioned the instabilities on my pc caused by the board's aggressive use of RAM timings to Gigabyte product support. I got a reply stating they didn't have same RAM as me to test (BL3KIT25664BN1608) unfortunately, but told me to try F7a or F7c BIOS. This is fair comment, of course since they can't test every ram.

Hope you solve your problem.

Regards,
Achelon

[dbee]

sorry if i ask too many questions but, would does this mean then exactly when they all light up right when my machine starts up. They all light up green to red everytime , optimized defaults , OC'd setings, all the time everytime. is this normal? I used windows mem diagnostic tool, no errors. i unplugged case connections to the mobo and gonna start clocking right about now. let ya know- THANKS !