Irritating cold boot issue

[Radikal]

Hi guys, I'm definitely in need for some help here, before I loose my mind!

My specs:
Asrock M3A770DE [flashed to 1.4 AMI BIOS - the problem was there with 1.20 too]
Corsair HX520W [replaced an Enermax Noisetaker 470W, thought it might have been it, no difference]
Athlon II X4 620 [auto vcore 1.4V - Cool&Quite disabled]
OCZ 3X13334GK 2x2GB kit [manually set to 7-7-7-20 1.75V - manufacturer's specs]
XFX Radeon 4770
WD 640GB HD
Sony SATA DVD-RW
Avermedia TV tuner
Logitech MK300 wireless keyboard&mouse
[I also have plugged-in but turn off an external HD and an EMU 0202 - both USB]

Everything is stock. No attempt to overclock whatsoever

My PC just won't play nice after being turn off for a couple of hours (in the morning for instance). The system will power up, then the BIOS will "beep", lately it will usually restart itself right away, then pass POST while displaying "your system has failed to start, change some settings etc" and then Windows 7 64 will lock up in the loading screen. This is with the SATA controller set in IDE mode. Before I had it on AHCI and it will get stuck on the screen where it detects the devices, endlessly searching "..."

Now, I used to have an issue where it will get stuck before it passed the memory test displaying 6B38 on the bottom right. I changed the RAM settings to the manufacturer's specs [7-7-7-20 1,75V] and it never got stuck there again. I also run memtest86+. No errors after 4 passes.

The frustrating thing is that this behavior is only demonstrated after the computer is off for some time. In fact the problem has deteriorated, from some failed boots, to every time it's off for a couple of hours.

When it loads into windows it's rock solid for hours (8-10 hours), doing everything from gaming, to heavy DAW usage, multitasking, you name it.
Also I can restart it and it will allays reload fine. And the final nail in the coffin of my sanity is that even if I do a full shut down and cold boot [leaving it off for about an hour] it will still load fine, quick, no problems!

I cleared the CMOS, loaded default settings, everything is on auto, except for the RAM that was set by default to 7-7-7-16 1.5V, and I change it to spec.

How can it be any hardware or software issue, if after the first miserable boot everything is working fine and reboots and cold boots...

I'm loosing it here! The only other factor I can possibly think is that the room is pretty cold, about 12-14C [around 55F], when I turn the pc on... The temperature is actually in correlation with the amount of trouble it has when first booting... Since the temperature got this low, it always fails to boot on the first attempt. But I don't see how that can affect anything save for the fact that the motherboard is not electrically sound...

Any help and suggestions will be gladly appreciated guys!

Thanks

ps. The motherboard has a EuP jumper, that supposedly lowers the power consumption to EU standards. Can there be some kind of power instability/inefficiency happening with this and the low temperatures?
ps2. The CPU temperatures hover around 20C when booted. Maybe it's some sort of CPU instability with low temperatures?

 

[Texun]

I'm no expert but it sounds like a capacitor problem or possibly a cold solder joint. I had a MSI board a few years back that did the same thing. Wouldn't start on a cold boot but hitting the reset worked 99.9% of the time. Capacitors need to charge up which may be why a warm boot works - it's already charged.

A cold solder joint will also do funky stuff. As the temp increases the parts expand and close the gap. Depending on the joint the reverse is also possible - it can expand away from the joint causing it to fail when warm. I don't know which it is, if any, but if it's temp related my money would be on one of the two. My MSI board got a little worse over time and I finally replaced it. I hope someone else comes along with a miracle fix for you. I would like to know as well in case I get hit with it again. :\

 

[daniel49]

sounds like early stages of motherboard problems.

 

[Radikal]

Thanks for the replies

I did some timing and temperature sampling

Yesterday morning. Same problem. Temperature around 14C.
Left the PC off for 4,5 hours. Temperature 14,5-15C boots at first go...

Today morning, temp 15-16C, same problem.
Left it off. After 3,5 hours, temperature about the same boots fine...

So WTF?!!! I mean, we 're not talking about extreme temperatures here! And it seem that there's a time ceiling after which the problem occurs...

As a matter of fact I completely switched off the power strip just to see if it made any difference. Nop.

How long do the capacitor take to discharge? Are we talking hours or minutes here?

My last attempt will be to test some other timing and voltages on the RAM, in case that's the issue, although how it works fine [and passes memtest] after that first boot is beyond me...

I'm fearing something is wrong with the mobo though...:twisted:

**little update**
I visually inspected all the capacitors, there are no signs of bulging, leakage, or venting. Not that that means anything. Also there are no visible marks on the motherboard circuit traces (in case a scratched something)...

 

[Blazer]

have you looked at event viewer to see if windows is reporting problems starting up ?

also is 1.75v the default voltage setting for your ram, if so try 1.80 or 1.85.

 

[Radikal]

No, there's nothing important there. No errors. Besides windows don't even get a chance to load. The freezing happens just as the loading screen appears (before the light swirling thingies)

I asked at OCZ forum about this particular kit and they did not suggest going that high on voltage. They did suggest maybe relaxing the timings and I'll try this next, if the morning boot fails and it probably will :twisted:

 

[Blazer]

relaxing the timings may solve it.

 

[Radikal]

I sure hope so! I'll keep my fingers crossed! :biggrin:

 

[Texun]

No, there's nothing important there. No errors. Besides windows don't even get a chance to load. The freezing happens just as the loading screen appears (before the light swirling thingies)

I asked at OCZ forum about this particular kit and they did not suggest going that high on voltage. They did suggest maybe relaxing the timings and I'll try this next, if the morning boot fails and it probably will :twisted:

My bad. I thought it was hanging at the POST. You might try burning a live Linux CD, like SuSe. Insert the CD when you turn it off at night. Make sure it's set to boot from the CD and see what happens when you start it up the following morning. That should tell you if it's anything related to the HD or OS. The downside is that it really tells you nothing if it happens to boot clean on just one test. You might need to try it over a period of a few days to see if the problem you are experiencing can be replicated.

I'm not an engineer but capacitors do need to charge up to work right, and if you have a bad one there is no way to know how long it may take for that one cap to charge. It would depend on how bad it is - if that's the problem. The first order of business is to isolate the problem to the OS, HD, board, etc. The Live CD should help with that.

 

[Radikal]

Well yeah, basically that's where it hangs. When you turn it on [after lots of hours] it immediately resets, goes through POST and then hangs on windows loading screen immediately. Actually Windows don't even begin to load.

When I had the drives set to AHCI mode [reverted to IDE just in case] it would do the same, but hang at the screen where it finds the AHCI devices.

RAM timings could be it, but if not then the theory of a faulty capacitor somewhere sounds the most probable...

I was thinking of trying a LiveCD too. It's on the to-do list Let's hope I won't have to.

Again thanks for all the ideas and suggestions guys!

ps. I already set the RAM timings at 8-8-8-24 as the guys at OCZ suggested. So I'll know by tomorrow. This must be the slowest troubleshooting ever!

 

[Radikal]

Ok so it booted today, first go with the relaxed timings!

It's nothing conclusive, I'll have to see what it does after work, when it's gonna be off for a while. But maybe we got something here.

Again I relaxed the timings of my OCZ RAM from 7-7-7-20 <-> 1,77V to 8-8-8-24.

 

[Texun]

Ok so it booted today, first go with the relaxed timings!

It's nothing conclusive, I'll have to see what it does after work, when it's gonna be off for a while. But maybe we got something here.

Again I relaxed the timings of my OCZ RAM from 7-7-7-20 <-> 1,77V to 8-8-8-24.

I agree. Before you can confirm memory timings as the problem or fix, I think you will probably have to repeat the test a number of times with the board in the same state as previous boot failures.

But hey..... at least it's a good start. No pun intended.

 

[Radikal]

After another 8 hours, another good boot! Looks promising!

 

[StrangerGuy]

Good Luck!

Side story: I got a Gigabyte Geforce 6100 IGP mobo that works flawlessly as long as I kept the PSU energized...When I cut off and switch on the power, there's only a 50&#37; chance to boot successfully. Free motherboard, so I ain't complaining.

 

[Radikal]

Thanks!

You know, we now have BIOSes where we can tweak every little parameter known to man, yet there isn't a standard to create some sort of system level failure log. Like "reboot because this or that happen in that device etc"...

Instead we are left trying to troubleshoot mystifying errors. Atleast mystifiyng to those who don't really care about the insane tweaking but just want a goddamn stable system.

Ok, rant is over! :biggrin:

 

[Blazer]

i believe your problem is corrected, this is why its important to buy memory that can operate at default voltages and mobo timings, if sometime in the future your mobo resets to defaults you will have to do this again.

 

[Radikal]

I sure hope so!

Believe it or not, I wasn't planning on buying high performance RAM, but when I ordered the parts they didn't have the one I wanted [I think it was a Kingston or Corsair DDR3 value RAM] so I went ahead with this one...

Not that it's OCZs fault [if it turns out that the issue is curred], but it looks like motherboards are picky with these pushed modules.

Let's see what it will do tomorrow morning

 

[Blazer]

OCZ has a good reputation in the OC's world, dam good ram but not needed outside of ocing, a good name brand performance memory that meets mobo defaults for timing and voltage will provide a good stable system day to day.

anyway glad you solved you issue, good luck.

 

[Radikal]

Oh yeah man. I definably don't put the blame on OCZ, they 're a great brand. It's the value motherboards that are picky with this RAMs.

I have left it off since yesterday. If it boots later today then I guess we have found the culprit (the motherboard that can't ran the modules stably)

 

[Radikal]

Soooo... did it again... failed to boot, restarted, crashed on loading...

Relaxing timings to 9-9-9-30...

 

[Texun]

:\

Time to break out the live Linux distro

 

[Radikal]

Yeah, I'll try it with the new timings for a day or two. Then the distro. Although it won't help since it restarts immediately on boot at first...

I noticed that the Vcore [left on auto] is strange. On BIOS is reported around 1,357V (I think), AMD OD says it's 1,4V, CPU-z shows it at 1,425V...

Hmmm... anyone knows what is the suggested vcore for an Athlon II X4 620? maybe I should try setting that manually... [I do have C&Q disabled].

Could it be that the bios is overvolting it at first - hence the immediate reboot? Although the CPU temperatures are extremely low, around 22C...

 

[Texun]

No updates in a few days so I guess the problem has been resolved? If so then it's good you didn't end up RMA'ing the board.

 

[Radikal]

I'm holding the update until I can conclusively say that I have something, cause I kinda have!

When the 9-9-9-30 timings failed to fix the problem (no surprise there) I turned into something else. I had noticed that the Vcore voltage was pretty high (1,4V in AMD OD, 1,425V in CPUz) so when it failed to boot I entered the BIOS. The Vcore [on auto] was saying 1,376V, yet that's not what I was seeing. So I manually set the value to 1,365V [I think that was the closest value available]. Now I had a correct reading in both programs.

It has booted fine since then, for two days now. But I want to make sure, before I jump the gun again.

I hope this fixed it and would really like if someone could explain to me why.

Did the BIOS overvolted/undervolted the CPU on boot causing it to restart? If so why sporadically?

Is there a relationship between the (pretty high) Vdim voltage [1,75V] required by this particular kit [2x2GB] and the Vcore voltage, causing the BIOS to do something wrong?

I really don't know. I sure hope this fixed the problem though!

 

[Radikal]

Unfortunately neither the manual Vcore settings did it. It worked for 3 days, today failed to boot again.

Only difference? Today is much colder... This does not look good... :\