GA-P67A-UD5 : doesn't boot as ACPI S4_S5 LEDS glow red

[adnaanusa10]

damn.. still waiting for my system to work..with the new i7 2600k

purchased GA-P67A-UD5 yesterday.. and the mobo gives me problems..

The ACPI S4_S5 LEDS glow red.. any expert opinions what this means and how to fix it to get my system to work.

[Dark Mantis]

Hi and welcome to the Gigabyte Forum.

I would suggest clearing the CMOS as a first step.

Remove the power cable from the mains supply and then press the power switch on the case for a few seconds just to drain any residual energy in the PSU capacitors.

Once done remove the motherboard battery for at least ten minutes before replacing it.
 
Next plug back into the mains supply and boot.
 
You will now need to enter the BIOS by pressing DEL and load Opimised BIOS Defaults.

Make any other changes to the BIOS settings to suit your self and then press F10 to save and exit.

If you are still having problems then a few more details especailly of your hardware and symptoms would help.

[niekell]

Hello Dark Mantis,

I am also having the same problem with a GA-P67A-UD7. The ACPI S4_S5 LED glows red and the board just keeps stopping and starting with the Debug LED flashing a partially lit second character*. (* The Debug LED can show two character error codes but this just seems to be a three quarter circle in the second characters position with the left top section missing... its probably just a power-up flash which means nothing as it doesn't correspond to any code in the manual.)

In addition, before it powers down, the memory and cpu-vvt LEDs [both green and orange] all go on and the top two phase LEDs are red, with the sixth green. Before the ACPI S4_S5 LED goes red it is initially unlit whereas the S0 LED is green.

I tried what you suggested but nothing new happened. Still locked in the up-down cycle of doom.

I have managed to get into the BIOS twice (but no longer, it seems those two times out of 100 boots were flukes), and I even loaded the optimal defaults one of those times. It ended in a FF Debug Code on the on-board LED when it hung on the windows7 (64ULT) start screen.

Do you have any further suggestions?

I was using v1.65 RAM, but have since put in v1.5 RAM, made no difference sadly.  
There are a few posts around which suggest updating the BIOS... don't see how I can when it doesn't even start!

Best wishes, Niekell.

[Dark Mantis]

Hi niekell  and welcome to the Gigabyte Forum.

It is probably something fairly basic like a power supply issue or a short to earth.

I would suggest the next step, after checking all your power connections to the motherboard, that is the 24 pin and the 8 pin, is to remove the motherboard form the case.


Make sure that you observe anti-static precautions.

Lay some cardboard or use the motherboard box that is non-conductive on the worktop and remove the motherboard/CPU/heatsink/fan/buzzer/one stick of memory and PSU from the case and put it on the cardboard.

Add the keyboard and try and boot. The system should try and then fail emitting a series of beeps. Please post what sounds it makes.

[niekell]

Hello Dark Mantis,

PSU or short... makes sense, however the PSU (750W ThermalTake Toughpower) has not changed since the previous board so I think that unlikely. It is also a nuisance to remove from the case (so apologies but I left it in for the test you suggested). I removed the MB with 1x2G RAM in slot1, CPU w/heatsink and fan, pc speaker (which I didn't previously have connected) and placed it on the (upside down) packing cardboard frame the MB came in.
I connected the KB to the PS2, plugged in the MB power and CPU power, then turned it on.

No change, no beeps or noise either.

Here are some photos of the result and a small low quality (sorry) video of the way it fires up.

Powered off: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_001.jpg
Powered on / not booted: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_002.jpg
Powered on / not booted #2: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_003.jpg
Powered on / not booted #3: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_004.jpg
Powered on / not booted #4: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_005.jpg

Powering up video: http://users.adam.com.au/niekell/ga-p67a-ud7/20110120_006.mp4

I'm still at a loss.
I could remove the PSU as you suggested but I just don't think that is shorting as it was working fine with my old system.
I'll try it if you think its still a likely culprit though?

No warning beeps of any kind.
I tried using the Clear CMOS button, no difference.

Best wishes, Niekell.

[Dark Mantis]

Thanks for the pictures and video it all helps.

From the video it would seem that it is trying to start and then shutting down. This is indicative of a thermal shutdown on hte CPU. Double check the CPU cooler fixings and make sure they are fully home. This is quite a common problem with the stock coolers.

Also remove all the memory apart from one stick in the first slot DDR3_1.

[niekell]

Hello Dark Mantis,

Thanks for the advice, I spent all day troubleshooting this and I did remove and reseat the CPU and cooler as a part of that process.
What seemed to work (not reliably mind you) was something like what is mentioned here:

http://www.xtremesystems.org/forums/showthread.php?t=263598&page=9

This part in particular at least got me intermittently back into the BIOS which was enough to reset it and slowly things stabilized.

It is possible to make the BIOS Auto-Recovery kick in (Dual BIOS) and re-flash the MAIN BIOS with the contents of the BACKUP BIOS.

This is a simple and easy method for anyone to try before having to resort to other more difficult methods, or a RMA.

1. Shut off the power supply using the switch on the back of the PSU, wait 10-15 seconds.
2. Press and hold the case Power On swtich, then while still holding turn on the power supply from the switch on the rear.
3. Still holding the case power on switch, the board will start, once it does release the case power on switch and shut off the power supply via the switch on the read of the unit. (Do the latter two parts as quickly as you can once the board starts)
4. The board will shut down.
5. Turn the power supply back on using the switch on the rear of the unit.
6. Turn on the motherboard by pressing the case power on button.

Once the board starts this time you should see the Gigabyte splash screen, or POST page, then the Auto-Recovery from Dual BIOS will kick in. You will see a checksum error, and then recovery from BACKUP BIOS will begin. Once it is done reboot your machine and enter the BIOS and load optimized defaults then save/apply/reboot back to BIOS.

Now you are done, and will be using whatever BIOS was in your BACKUP BIOS, From there you can attempt whatever you were previously trying, or update your BIOS to the latest version.

Now I should point out that during the course of the day I tried the above along with the Clear CMOS button and sometimes it worked and sometimes it didn't. Eventually it booted and continued to boot as I slowly added each component one at a time.
Almost 12hrs later and it seems stable, at least it boots and recognizes all the components... even if windows did have conniptions about losing its "activated" status as the OS came into play.

Now I only have two things outstanding with regards to the board.
I cannot update to F6 of the BIOS using either @BIOS or Flashing it via the Bios Setup Screens. The file (via http or ftp) for F6 always reports as being corrupt in Flash-Bios and in @BIOS windows throws up a "stopped responding" message immediately upon the completion of the download and read phases. Nothing seemed to work to go to F6. I DID manage to go from F3 (the original) to F4 without any problems though. I tried F7a... same result as F6, corrupted file. (I downloaded multiple times from multiple servers located in Asia, China, USA, etc... always corrupt.)
I also don't seem to be able to get dual channel RAM working. I have 4 sticks of ram in there now, 2x4GB and 2x2GB, and was of the understanding that if the channels matched then there would be no problem with that (they have the same timings and speed, despite the size difference). I changed the BIOS Memory Management to be Expert and enabled channel and rank interleaving... not sure I should have done that though...?
Was I wrong to expect the use of dual channel to report the RAM to the OS as half of what it actually is? I just assumed...?

So it could have been the CPU seating or the cooler, or anything since I went all the way back to nothing on the board and worked forward from there. Still the only thing that seemed to have a sliver of success was what the above quoted section recommended (along with many applications of the Clear CMOS button, sometimes at the same time as the above procedure!). Once it began booting it had a bout of failures again, where I had to apply various and prolonged combinations of buttons and power switches, then it booted again and has stayed up since.

I've never had a MB this... finicky... fussy... or troublesome. Now I just hope it stays stable... from now on would be preferred.

Thanks for the help Dark Mantis.

Best wishes, Niekell

[Dark Mantis]

Well I am glad that you seem to be winning but things don't sound right yet.

Silly question maybe but i have been caught out before so I will ask, did you remove the plastic cover from the TIM on the bottom of the CPU heatsink when yoiu installed it ?

If so did you clean off and replace the thermal compound when you remvoed and then refitted the heatsink ?

I wouldn't mess with the memory at the moment until you have a stable system. Just use one set of memory. When you can i would suggest downloading and running Memtest86+. This version is for the P67 boards

Memtest86+ V4.20 Beta 13

.Bin :http://www.memtest.org/download/beta/420b13/mt420b13.bin
.Iso :http://armstrongcomputer.ca/mt420b13-jasonacs.iso

Insert one stick of memory in slot 1 and run Memtest on it for at least 10 loops and if there are no errors then swap it over with the next one and continue untill you have checked all  modules.

If you have any errors the module is faulty.

If you have any faulty modules you will have to return the whole kit as they are matched.

Post back when you have done that with the results.
 

[niekell]

Hello Dark Mantis,

I did remove the plastic from the CPU cooler prior to putting it on, but I did not clean off and replace the TIM when I later removed the cooler (and CPU) when trying to resolve the initial stability issues. In retrospect I should have done so to provide a smoother conductive surface with new TIM, but I guess I had other things on my mind at the time.
CPU Cores are staying around 49C which seems acceptable despite the less than optimal TIM surface.

I downloaded and ran MemTest86 v4.20 Beta as you suggested.
First I ran it on one of the 4GB GSkill sticks, it took over 7hrs to complete but after 10 passes there were no errors detected.
Then I switched over to the next 4GB GSkill stick, boot failure.
Switched to a 2GB Kingston stick, boot failure.
Switched to the other 2GB Kingston stick, boot failure.
Not believing I could have 3 bad sticks of RAM I stuck the original 4GB GSkill stick back in, boot failure.
Hit the Clear CMOS button. Booted okay.
Freakin' inconsistent motherboard!
Not wanting to spend another 14hrs (for a 4GB and 2x2GB) doing RAM tests on single sticks I put all the RAM back in and let it run Memtest86 overnight. 6 passes of 12GB RAM ran without error. I don't believe the RAM is the source of the inconsistency.
There haven't been any boot failures since the overnight Memtest86 run. It has run all day with no unexpected shutdowns or reboots.

Hmmm... it seems that this MoBo isn't as resilient to hardware changes as all my previous ones have been.

Oh, the issue with the @BIOS F6 not working has also been resolved. So the BIOS is now running the F6 version.

Also of note, booting into WIN7 ULT 64 takes over 2 minutes. This is commonly a driver issue, however booting in "Diagnostic Startup" mode (which only runs minimal services and drivers) did not quicken the boot time at all! I have updated all the drivers I could during this process, but still the horrible boot time persists. I even used the GigaByte Smart6 app to enable the OS Quickboot. No change.

You never said if I should be expecting to see the OS report 6GB RAM if it is running the 12GB of installed sticks in dual channel configuration, should I be seeing 6GB or 12GB in all the places the OS reports RAM sizes? (Not sure if its supposed to, but when I ran dual channel on my previous PC it did halve the reported ram when running in dual channel mode via the BIOS.)

On the topic of mis-reporting things, the i7-2600 (NOT a K) is supposedly 4 core hyperthreaded... so why isn't windows reporting 8 cores as it traditionally interprets the 2 threads per core as each being individual cores?

Best wishes, Niekell.

[Dark Mantis]

I did remove the plastic from the CPU cooler prior to putting it on, but I did not clean off and replace the TIM when I later removed the cooler (and CPU) when trying to resolve the initial stability issues. In retrospect I should have done so to provide a smoother conductive surface with new TIM, but I guess I had other things on my mind at the time.

It isn't just about making contact, the old TIM will have cured and so not be any good any more. That is the reason for replacing it. If it is failing to do it's job you will get thermal shutdowns which is what you could have in this situation.

Interesting result from the Memtest program. There is something there that is causing intermittent problems though. Maybe bent pins?

You never said if I should be expecting to see the OS report 6GB RAM if it is running the 12GB of installed sticks in dual channel configuration, should I be seeing 6GB or 12GB in all the places the OS reports RAM sizes?

You should see all 12 GB of memory in the OS. This again points to bent pins.

Very carefully remove the heatsink and fan. Probably a slight twisting motion is best to release the two as the thermal paste can act like a glue.

Once off remove the CPU taking extreme care both physically and electrically(static) and place somewhere safe.

Make sure that the lighting is very good and even.

Now scrutinise the socket for any pins that are out of alignment, even slightly.
If in any doubt whatsoever take a couple of macro (close up) photos of the socket and post them on here for us to check.

Also look at the bottom of the CPU at the lands (little copper circles) and make sure that there is a mark in each somewhere near the centre.
If there are any missing marks or any of the pins are obviously bent then that is your problem.


Check this out first and if we don't find anything we will start looking at the memory settings.

[niekell]

Hello Dark Mantis,

The "Bent Pin Search"
I found some slightly bent pins down the side closest to the rear of the board.
Straightened them as best I could, they really didn't look too bent in the first place!
Comparing the CPU lands indicated they were "off center" otherwise I would probably not have seen them.

(Macro shots, but with flash, so taken at different angles.)

Board Pins:
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_001.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_004.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_006.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_007.jpg

CPU Lands:
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_011.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_015.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_016.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_017.jpg
http://users.adam.com.au/niekell/ga-p67a-ud7/20110123_018.jpg

After tentatively straightening the pins which were landing "off center", I cleaned off all the old TIM and added new Zalman Thermal Grease and refitted the CPU and Cooler.
Boot failure! Clear CMOS and boot success. I still find that odd. 

Since then I've reverted the memory settings to all be "Auto". I don't think its running in dual channel mode while in "Auto".

It is also running about 6C hotter than before (avg 55C at low load), which I wouldn't have expected with the cleaning and reapplication of Thermal Grease.

I've had enough of tinkering with hardware for now.
Thanks for you help Dark Mantis.

Best wishes, Niekell.

[Dark Mantis]

I can appreciate how you feel at the moment and sometimes it is better to just walk away from it for a while.

It is weird that it is running hotter now. The new thermal compound will take a little while to cure but I wouldn't have expected it to make that much difference.

As long as all pads on the CPU have marks somewhere near the centre it should be fine. At least it is another possibility dismissed.

Next thing to check is the memory settings then. Can you get it stable enough to boot into the BIOS?

If not you are going to have to remove the motherboard etc from the case I'm afraid.

[niekell]

Hello Dark Mantis,

Seems to get up to about 69C under heavy load. Thats still under the max spec so I suppose its okay. I can live with it.
Yeah all pads on the CPU had marks on them, some (3-4) were near the edge before I straightened some pins. All were still on lands though.

Okay, memory settings... uh... boot to BIOS... uh-oh... I think we have a misunderstanding here.
Whenever I say I can boot, it means I can boot into the OS and WIN7 ULT64 seems to run fine (no BSODs).
I've had the system running for about 12hrs today and its been stable and without problems.
So getting into the BIOS is no problem.
What memory settings would I need to set to get the dual channel ram working properly?

Best wishes, Niekell.

[Dark Mantis]

Can you post the make and model (kit) number of your memory and how much of it you have installed please?

[niekell]

Hello Dark Mantis,

RAM Statistics:
Slot #1 : DDR3 4096MB PC3-10666 G.Skill (Part Number: F3-10666CL9-4GBXL) 9-9-9-25 1.5V
Slot #2 : DDR3 2048MB PC3-10666 Kingston (Part Number: 9905471-001 .A000LF) 9-9-9-25 1.5V
Slot #3 : DDR3 4096MB PC3-10666 G.Skill (Part Number: F3-10666CL9-4GBXL) 9-9-9-25 1.5V
Slot #4 : DDR3 2048MB PC3-10666 Kingston (Part Number: 9905471-001 .A000LF) 9-9-9-25 1.5V

For a total of 12GB.

Update on CPU Temps:
Temps have come down a bit in the CPU since originally reapplying the TIM.
It went from 49C up to 55C when originally applied, but today it is back down hovering at 50C after a few hours of running.
So I don't know, maybe the TIM needed some heat to spread out and mould to the surfaces more smoothly?

Best wishes, Niekell.