Build will not POST after running great for nearly a year. :(

Mike2015 · Jul 6, 2015

Hello...I'm hoping someone can help me with my issue as I've run into a brick wall and don't know how to proceed further. Let me start by listing my system components:

ASRock Z97 Fatality Killer Motherboard
Intel i5 4690K CPU (not overclocked)
8 GB G.Skill DDR3-2400 Ram (2 x 4 GB)
XFX ATI Radeon HD 6770 GPU
Corsair H90 Liquid CPU Cooler
Samsung 840 Pro 256 GB SSD (Running OS)
Seagate 1 TB HDD (Storage)
WD 500 GB HDD (Storage - from an older system)
Antec EarthWatts 650 Watt Power Supply
OS - Windows 7 Pro (64bit)

I built this system last August (2014) and it has been running great with no issues up until about a week and a half ago. All of a sudden it will not POST. The power light comes on and I can hear the fans come on. My display doesn't detect a video signal and enters sleep mode. My keyboard (which was used to turn on the system) lights go dark and I can hear the GPU fan spin up high, then low, then high, then low. Then the system shuts itself down. A few seconds later, it tries to start up again, then shuts down. After several iterations of this, I will ultimately get a display stating that the system has failed to boot after several attempts, press F2 (or DEL) to enter BIOS, along with a 10 second countdown to boot into the Windows OS. If I let the countdown complete, it boots into Windows and everything seems to run fine...

So here is what I've done to try and diagnose what's going on so far:
I swapped out the HDMI cable.
I removed the GPU from the PCIe slot (suspected it was bad) and switched to onboard graphics.
I removed the HDD power and SATA connectors
I removed all USB connected devices
I opened the case looking for any shorts or lose cables, double checked all connectors
I unplugged power and cleared out the BIOS settings via jumpers
I replaced the CMOS battery
I removed and re-seated the RAM
I even pulled the SSD SATA connector ( I got no POST messages at all after doing this )

While in Windows, I downloaded Intel's CPU diagnostic and ran that. Everything passes with a green light.

This is totally frustrating. Has anyone else ever encountered this type of issue? It seems like it has gone from stable to totally unstable overnight... I did not make any recent hardware changes. There haven't been any power outages or electrical storms that I know of. There were a few updates which were applied by MS shortly before this happened, but I've never known those to cause problems in the past. I'm stuck....I don't know how to proceed to narrow it down to the cause of the issue now and I'm getting close to my one year of purchase date. Any help anyone can offer would be greatly appreciated!

Thanks!

Darkbreeze · Jul 6, 2015

What are the +3v, +5v and +12v system voltages showing in the BIOS?

Have you recently updated the BIOS? Are you on the most recent version?

How old is your OS installation, 1 year? It really kind of sounds like it's either a PSU issue, corrupted boot partition or a drive with failing sectors.

Run Seatools for windows, short drive self test and long generic test, to see if there are any issues with any drives. Run it on all connected internal drives.

Download, create bootable media and run Memtest86. Run it for several passes. I'd say 3-4 passes should tell you if there is an issue. 7 passes for full testing of RAM.

Visually verify whether or not the CPU cooler fan is running during the initial POST process when it shuts itself down. Many systems will almost instantly shut back down if the CPU fan tach signal is lost or below a specific RPM.

Mike2015 · Jul 7, 2015

@darkbreeze: Thanks for your input!

Here are the voltages displayed in the BIOS:

+3v = +3.344
+5v = +5.088
+12v = +12.096

CPU Input voltage = +1.760
CPU Vcore voltage = +0.976
CPU Cache voltage = +1.104
System Agent voltage = +0.824
Analog IO voltage = +1.008
Digital IO voltage = +1.008

CPU and Motherboard temps show 32C

No, I have not updated the BIOS. The BIOS is version 1.3 which is how it was shipped and shows that it's compatible with my CPU. The current BIOS version is 2.2. I didn't feel a need to update it because everything was working just fine. Plus it has been my experience that a re-install of the OS is sometimes necessary after updating a BIOS, and I did not want to have to go through that hassle. (If it ain't broke, don't fix it

)
Do you think the BIOS may be the issue?

The Windows 7 Pro OS was installed when I built it, so yep, it's nearly a year old. I'm running the OS on a Samsung SSD. Samsung's Magician shows that everything is fine. Will Seatools also diagnose issues with an SSD?

When I first put together this build, I ran Memtest86 for several hours. I believe it went through 11 or 12 passes with everything being okay. Should I repeat?

I cannot tell whether the pump on the Corsair H90 liquid cooler is actually running or not. It's very quiet. The fan attached to the radiator is running. When I touch the top of the cooling block, it remains cool to the touch, so I'm assuming it's running and doing its job. The liquid cooler is the most recent addition to my build (Dec 2014). I was planning on giving overclocking a try since many other i5 4690K owners have reported good results doing so.

I'm thinking about picking up a PC speaker (don't have one installed) and a power supply tester as well.

Darkbreeze · Jul 7, 2015

Reinstall of the OS is never necessary after a BIOS update. At least, not the four or five hundred times I've done it.

Yes, Seatools is a way more involved testing procedure than Magician, which I also have. Is the SSD a recent addition to the system or did it come with it?

Yes, repeat the Memtest86 testing. Nothing, car parts, computer parts, your tv, was ever bad at some point but then at some point, was, so between then and now something could easily have changed.

You should DEFINITELY have a PC speaker connected for relaying any error codes. A power supply tester is a good idea, but again, may not show an issue if it's intermittent. If it's doing it pretty regularly, it should show something though.

I'd definitely update the bios. Changes to drivers or the OS itself can become an issue on older bios versions in some cases. It isn't common, but can happen. It wouldn't normally cause a unit to not at least POST though. Always good to be on the latest bios unless you know for certain that a particular bios version is problematic with some hardware item you have installed, which would also be rare.

Darkbreeze · Jul 7, 2015

Is this same behavior happening when you're running on the integrated graphics as well?

Mike2015 · Jul 7, 2015

Okay..... I will flash the BIOS and give that a try. Can't hurt anything at this point

If the problem persists, I'll run the Seatools and Memtest86 diagnostics and see what happens. (I installed the SSD when I first built it. I did a fresh OS install from the Windows 7 install CD)

I'll also pick up a PC speaker and install that as well and get the PS tester. Is there a preference in PS testers? I see some where you plug in the 24 pin connector directly and some which test the various output leads seperately (5V, 12v, etc).

Oh...and yes. The same issue is happening when I remove the GPU and switch to onboard graphics.
I did notice my GPU runs very hot when gaming now that the case is open. It can get up to 76-78C with the fan running at 65% of max. I've always assumed this was normal however. I've had the video driver crash on occasion as well. Any attempts to overclock the 6770 have not been successful, so I pretty much leave it at the default settings.

Thanks again for the assistance.

Darkbreeze · Jul 7, 2015

You really don't need a PSU tester, they don't test load conditions anyhow. Just get a cheap multimeter and test according to the steps here:

https://www.youtube.com/watch?v=ac7YMUcMjbw

If you want to get a PSU tester, then I suppose the Thermaltake Doctor power II is as good as any, which probably isn't saying much.

Mike2015 · Jul 8, 2015

Okay....the advice about having a PC speaker was spot on!

I wasn't able to get much done yesterday but I did pick up a speaker and plugged it in. As soon as I powered up, I was getting all kinds of beeps.

Here is what it's doing after I turn it on.

3 short beeps, 3 short beeps, <interval>
3 short beeps, 3 short beeps, <interval>
3 short beeps, 3 short beeps, <system powers down>

about two seconds later, the system restarts (lights come on and fans start to spin again). Then I'm getting...

2 short beeps, 4 short beeps, 5 short beeps, 3 short beeps, 3 short beeps, 3 short beeps, <interval>
2 short beeps, 4 short beeps, 5 short beeps, 3 short beeps, 3 short beeps, 3 short beeps, <interval>
2 short beeps, 4 short beeps, 5 short beeps, 3 short beeps, 3 short beeps, 3 short beeps, <interval>
2 short beeps, 4 short beeps, 5 short beeps, 3 short beeps, 3 short beeps, 3 short beeps, <interval>

this continues until eventually I get the "Your system has failed to boot after several attempts. Default BIOS settings have been selected. Press F2 or DEL to enter UEFI BIOS or continue to wait and Windows will launch in 10-9-8-7-6.....1".

At which point I get one long beep and then OS loads.

So, I checked ASRocks support site for the beep codes all all they seem to have is this:
-----------------------------------------------------------------
The following are AMI BIOS Beep Codes that can occur.

Beep Code Descriptions
1 short Memory refresh timer error.
2 short Parity error.
3 short Main memory read / write test error.
4 short System timer failure.
5 short Process error.
6 short Keyboard controller BAT test error.
7 short General exception error.
8 short Display memory error.
9 short ROM BIOS checksum error.
10 short CMOS shutdown Read/Write error.
11 short Cache Memory bad.
-----------------------------------------------------------------

So.....my question is, are the beep codes accumulative? Do I really have all of these issues going on? According to these codes, my memory is suspect, the system timer is suspect, and the CPU is suspect. Is that correct?

Should I still proceed to flash the BIOS to the latest version?

Darkbreeze · Jul 8, 2015

No. Do not flash the bios until all the errors are sorted out. The only time I recommend flashing the bios in situations like these are when the beep codes indicate an unsupported cpu that is supported with a newer bios.

Darkbreeze · Jul 8, 2015

Pull all the RAM except a single stick, and try it again. Make sure the memory module is seated in the DIMM slot that is designated for single module operation as per the motherboard user manual population rules. If it still acts the same, try a different module in the same slot. If it fails in on all modules, try the modules again in a different slot, generally whichever slot is designated for the second module when two modules are used. Run through this until all modules and all slots have been attempted. Then we can go from there.

Make absolutely sure the RAM is fully seated and that there is no debris or foreign matter in the dim slots. Unplug all drives and external devices. At this point we're only trying to get into the BIOS to configure the XMP profile of the memory. You might also want to remove the CMOS batter for about thirty seconds, then reinstall it, with the power to the unit disconnected while doing that. Then try the memory as outlined above.

Double check everything outlined here as well: http://www.tomshardware.com/forum/261145-31-perform-steps-posting-post-boot-video-problems

And here: http://www.tomshardware.com/faq/id-1753671/bench-troubleshooting.html

Mike2015 · Jul 9, 2015

@darkbreeze

Okay. I pulled the graphics card out, pulled the ram, disconnected power to all drives and SSD (so basically just power to the motherboard). The PCIe slot and the RAM banks are all clear of any debris or dust.

I picked up a Thermaltake Dr. Power (they were out of the Dr. Power II) to test the PS anyway. I connected the 24 pin connector to it and turned it on. All lights are green except for the -5V light. I believe this is okay since this PS doesn't have a -5V output. Tested the 8 pin CPU power connectors, they showed +12V. Tested some of the molex connectors +12V and +5V.....SATA connectors also. So it seems like the power supply is good. Switched off the PS.

Removed the CMOS battery for a while and cleared the BIOS again just for kicks and then plugged the 24 pin connector back into the MB and replaced the CMOS battery. Switched on the PS and powered up. I got 3 long beeps, which is what I was expecting since there was no RAM installed at this point.

My MB supports dual channel memory and has four RAM banks. The manual recommends that RAM be installed in banks 1 and 3, or 2 and 4, or all 4 filled. So I started with 1 and 3. When I switched it on, it powers up and I hear:
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps
Then it powers down. Two seconds later, it powers up again. Then I hear:
6 short beeps, 9 short beeps, 3 beeps
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps

It continues through this cycle of beeps several times while I'm pressing F2 trying to get it to enter the BIOS. Finally I get the same message about the boot failures and am able to get into the BIOS. The system time and date need to be set so I do that, then go to the memory config page. Now I notice that only 4096MB of RAM is recognized (in bank #1) and it's running in single channel mode. So it's not even recognizing the other RAM in bank #3. :-(

I powered down, pulled the RAM (oh...and I'm wearing one of those electro-static wrist straps, by the way, while I'm doing all this) and re-positioned them into banks 2 and 4. Then powered back up. It powers up and I get the 3 cycles of 2 x 3 short beeps and it powers down. Then it wakes back up, and I hear:
2 short beeps, 4 short beeps, 5 short beeps, 4 short beeps, 3 short beeps
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps
3 short beeps, 3 short beeps

(Why this time is it different???) Again it goes through several cycles of this same pattern, before I am finally able to get into the BIOS. This time I see it has recognized the full 8192MB and is running in dual channel mode. (Uhhhhhg!) I load the XMP profile for the memory, save and exit.

So.......just to satisfy my curiousity, I reconnect my optical drive's power and put the MemTest86 disc in. Sit through the annoying beeps again 2-4-5-4-3 3-3 3-3 3-3 2-4-5-4-3 3-3 3-3 3-3 over and over.......finally the sucker lets me chose the drive to boot from and I launch the UEFI MemTest86.

I let it run all night (actually by now it's like 3 in the morning anyway)......
So, when I check it several hours later, guess what?! The test completed and it went through 4 passes without a single error. CPU cache memory & all 8GB all passed.

I went thought the checks in the links you suggested as well. Thanks by the way.

So I'm slowly going nuts here....and I've lost all confidence in ASRock and this motherboard :-(
How can a system go from being perfectly stable to something like this? (faceplant)

Anyway I digress........, should I continue testing at this point? What's next? A sledge-hammer maybe? :-/

Darkbreeze · Jul 9, 2015

Did you happen to notice if the time and date, and XMP settings, stuck, when you next when into the BIOS after changing settings and running Memtest? If not, check that to be sure. Could be a failed CMOS battery, doubtful, but always one of those things that when you find out what it is you go, F#$%, why didn't I think of that in the first place.

Did you verify that the pump is working and has a reference signal. If you have the original CPU fan that came with the chip, temporarily unplug the pump, plug the CPU fan into the CPU fan header and attempt to boot. If it boots normally, shut down right away, and then deal with the pump issue.

If it does the same thing as before, shut down and plug the pump back in.

Then try booting with only 1 stick of RAM in slot A2, as the population rules show slots A2 and B2 as the priority 1 population guideline.

If no luck, try it again with each individual stick. Don't attempt to boot with multiple sticks installed at this point. If that fails, try it again with each stick in A1.

If possible, take a look at the POST screen or in the BIOS and tell me WHO the bios manufacturer is. Phoenix, AMI BIOS, etc.

Mike2015 · Jul 9, 2015

Yes, the time and date and XMP settings stayed. I just replaced the CMOS battery with a new one a couple days ago. (beat you to that one.)

How do I tell about the pumps reference signal? I can feel it vibrate slightly and there's a barely audible hum, so I think it's working. I still do have the stock cooler which came with the CPU.
But.....if I remove the pump..... I'm going to mess up the perfect layer of thermal paste that came with it..... Bawwww :-(
Did I mention I have OCD?....That's really going to irk me. That means I'll have to remove the big ass radiator and fan as well so that there's no chance of the cooling block slamming into the board. The sacrifices we have to make

The BIOS is an AMI BIOS. A M E R I C A N M E G A T R E N D S

Darkbreeze · Jul 9, 2015

You don't need to remove the pump. Just unplug the connector and plug the CPU fan connector in temporarily. It can just hang there or hold it or whatever. Just so the bios sees the RPM signal from the fan. You shouldn't need to remove anything.

Darkbreeze · Jul 9, 2015

I'm seriously almost convinced this is a motherboard issue, but it could possibly be a failure of the internal memory controller on the cpu. That would be MUCH less common than a failed motherboard though. If the CPU fan issue doesn't resolve it, I think you will need to RMA the motherboard. The codes seem somewhat random and don't make a lot of sense. Code 3 indicates a lack of installed memory or failure to address the first 64k, which we know is not accurate. I don't know though, since it boots after a few attempts, I'm also wondering if there is a backup bios it's reverting to after several failed attempts and perhaps the primary bios ROM image is corrupted.

Mike2015 · Jul 10, 2015

I'm thinking you're right. It's the motherboard.......

I will do the fan test and see what happens and report back.

This POS motherboard does have a backup BIOS.

It also has something in the BIOS called 'Bootguard' which is enabled.
Says that it will revert to default BIOS settings after the specified number of failed boots has occurred.
The specified number is set to 3.

Darkbreeze · Jul 10, 2015

Try setting the memory speed to 1600mhz in the bios and see what happens as well. If that works, try setting it back to the default XMP profile manually and see what happens.

Mike2015 · Jul 13, 2015

@darkbreeze
Okay, I disconnected the pump and connected the old cup cooler/fan for the reference signal. When I powered up, no change. Same beep codes.

So I shut down and reconnected my pump and restarted. Once I got into the bios, I set the RAM to DDR3-1600 and did a save and exit. Same thing. Beep codes.

Mike2015 · Jul 13, 2015

I'll be damned...on a hunch from your comment about a backup bios. I moved the jumper over to utilize Bios B.
Turned it on and it started right up like normal with the splash screen and all....no beep codes.
Now what do you think? Bios A is corrupted?
I disabled the splash screen so that I could see the normal post screen and yep, just one beep and it posts fine.
Here's my dilemma....I've already requested an RMA and gotten approval to return it.

Darkbreeze · Jul 13, 2015

I would return it and let the RMA department know that the main bios is corrupted but that it boots fine with the backup bios. OR, you can try and see if there is a more current bios version and flash the bios then retry it with the main bios. It may only be the bios image that's corrupted, not the actual ROM itself since the backup image works. Either way, problem solved. At least you KNOW it is an issue with the bios image, which is what I was kind of leaning towards anyhow which is why I mentioned it.

If there is no newer bios image, you can try just reflashing the same bios version as it already has. Check your motherboard manual, there may even be an option to recover the main image using the backup image.

Mike2015 · Jul 13, 2015

@darkbreeze

Here's what I did. I flashed the backup BIOS to the most recent version (2.2). The update went fine and it POSTed again fine.
Then I decided to rebuild by primary BIOS using the newly flashed secondary BIOS. You were right, there is a tool within the BIOS to do exactly that. Well, that went fine as well. I moved the jumper back over to utilize the primary BIOS and started it up. Walla!!!! It works!!
Just to be sure, I pulled the RAM and moved them around switching between banks 2 and 4 (priority 1) and banks 1 and 3 (priority 2) and now each time the system is recognizing the full 8Gigs of RAM. I set the XMP profile for the Trident DDR3-2400's and the system is now running the RAM at 2400. I ran a few passes of MemTest86 with both configurations. No errors.

So then, I started putting components back in starting with my GPU. That worked. Then SSD and hard disks. Now it boots right up into Windows within 4-6 seconds just like it did before. Yeah!!!
Thanks so much for all of your help and for sticking with me!

I just have one question.....can you enlighten me on how my BIOS image became corrupt in the first place? I can see that they exist on separate ROM chips and was under the impression that they are pretty well isolated from malware attacks and such and are pretty well protected and stable. I've seen horror stories about people corrupting them by switching off or losing power in the middle of a flash update, but I wasn't doing anything in the BIOS at all. My system is not on a UPS....I'm just using a power strip with a built in surge suppressor. There haven't been an recent storms I can think of either. Just looking for ways to prevent this in the future.....
Anyway, I'm going to be carefully observing my system for the next few days now.
Thanks again for you assistance!

Darkbreeze · Jul 13, 2015

It could be anything. A slight voltage spike or getting too near a magnetic object. It could have just been a poor flash from the factory. I'm not an engineer so I can't get too specific, but I've seen it more than once although it is rather uncommon for the image itself to simply be corrupted unless it's as a result of a bad or incomplete flash. Usually it's the ROM chip itself that's bad in these situations so if you begin to experience similar problems, I'd right away contact the OEM and explain the situation. If your ROM chips are replaceable you can usually get them to send you a replacement and if the primary and secondary chips are identical you may be able to switch them until the replacement arrives. Hopefully, that's not the case and your problem was just a faulty flash job. Hope that was the case. So good luck and I'm glad we were able to sort this out.

Build will not POST after running great for nearly a year. :(

Reputable

Retired Mod

Reputable

Retired Mod

Retired Mod

Reputable

Retired Mod

Reputable

Retired Mod

Retired Mod

Reputable

Retired Mod

Reputable

Retired Mod

Retired Mod

Reputable

Retired Mod

Reputable

Reputable

Retired Mod

Reputable

Retired Mod

Share this page