Windows 7: Windows 7 Random Restart Issue

Hey guys,

I recently purchased a custom machine, and I've been dealing with this random restart issue in Windows 7 for several weeks now. Basically, after using the internet for a certain amount of time whether it is browsing or downloading, the machine suddenly freezes for about 10-15 seconds and reboots, giving me a Kernel Power error. When I check the event logs, this particular error seems to coincide with it:

And for details:

- EventData

SessionName Microsoft-Windows-Setup

FileName C:\Windows\Panther\setup.etl



ErrorCode 3221225485

LoggingMode 5


I've already done two complete formats of this HDD I have my OS on, and it doesn't seem to make a whole lot of difference as the problem keeps reoccuring. I also tried installing a new PSU going from 500 watts to 550, and that didn't change anything unfortunately. All of my drivers are up to date and I made sure to only install official drivers.

I tried disabling automatic restart in hopes of getting a BSOD I could see and a crash dump, but I haven't gotten any so far in this format. In a previous installation of Windows 7 when I had this problem, I had a BSOD that mentioned a tcip.sys error of some sort along with that same stop code. I haven't had that tcip.sys BSOD yet with this installation.

Lastly, I've taken this machine back to the shop where I bought this from numerous times and he has run many stress tests that would seem to eliminate a hardware problem outside of the HDD. I've done a lot of web research on this issue and have yet to come up with any kind of solution for this aggravating problem. It was mentioned that deleting the file in the Panther folder would fix it, but it only seems to prevent the issue from happening for a brief amount of time.

I'm really at wits end trying to resolve this and could use suggestions for rectifying this problem.

Here are my system specs as well:

Don't worry about the "Panther" thing or the "kernel power" event. They're not causal - they're side-effects of the problem.

It sounds like the shop is making you jump through hoops for what is very probably a hardware problem. Convincing them may not be easy if they're going to be pig-headed about it. A few suggestions:

1) Try going into the BIOS config menu and staying there for hours if necessary (don't boot into Windows at all). If the same unexplained restart happens while you're in the BIOS, it would really stretch credulity to claim that the cause is anything but hardware.

2) Make sure the BIOS itself is up-to-date. That can sometimes resolve similar symptoms.

3) Test what happens when the OS is reinstalled from scratch and then used without any updates, and without connecting to the internet. Literally no updates, no apps, no games ... nothing that's not on the Windows disc. Painful but potentially necessary as a troubleshooting step.

4) If all that fails to clarify what's going on, I can suggest a method to conclusively determine whether the reboots are caused by software or hardware, though it won't necessarily pinpoint which component in either category is the culprit (it may identify a specific software cause). You'll need another machine close by, and both of them will need to have either firewire (1394) or old-fashioned serial (RS232) ports. It's "last resort" stuff, so I'd suggest you try the other steps first.

Good luck with it.

Thank you for the suggestions. I've done a couple things since last posting: I did a repair OS install of Windows 7, and I updated the Bios to the latest version. We'll see if this occurs again, and I'll let you know.

By the way, the stop error I mentioned in my first post, could that be a way of detecting what the issue is more specifically?

Unlikely. A stop error ("bugcheck") is _always_ deliberately triggered by a driver or an OS component in response to an error condition that is considered irrecoverable. Something has to call the "bluescreen" function - KeBugCheckEx - in order to initiate that mechanism. That is fundamentally different to a situation where a machine reboots without a bugcheck.

Just to make things more complex, it is possible for certain types of problems to result in a bugcheck which "cannot display itself", for lack of a better description, the way that (for example) an error log concerning a dying disk cannot be written because the disk is dying. Ruling out that obscure possibility is the intent of the 4th point I mentioned above. Otherwise, mysterious reboots without a bluescreen are caused by hardware problems in >90% of cases.

Having said all that, if you upload the minidump corresponding to that bugcheck you experienced, somebody is sure to take a look. No harm in trying that approach.

I checked my Windows folder and did not see any kind of minidump file generated.

I should also mention that when I did take this machine back to the shop, he did end up replacing every part at some point with a brand new component. Same model, but a new part. He did this for the motherboard, CPU, PSU, etc.

I'm almost leaning towards a bad HDD and I'm tempted to purchase a new one and install the OS to see if that fixes the problem. The reason why I didn't think it was the HDD to begin with was that I had it in my previous machine with Windows XP and did not encounter any issues whatsoever. I believe that drive is somewhere between 8-12 months old as well.

EDIT: After updating the bios and doing a repair install of the OS, the random restarts have come back. So far have had two which happened after normal internet browsing. It really does seem as if this issue mostly happens under some kind of network load.

It looks like you're methodical. That always helps in these situations

Do you have another machine close by? I may be able to suggest some steps to prove/disprove the "hardware" hypothesis if you're OK to follow some esoteric instructions.

Ya, sounds like hardware problem to me. I had that kernel-power error with random reboots, which continuously worsened with time, but quickly.

Pulled the motherboard and saw blown caps with tops open. Bought a replacement, installed it and system has been perfect since.

During my research with that, I did read of many different potential causes though. Almost all were hardware. My example is definitely not the only one.

I only have a laptop nearby at the moment. I also just did a SFC scan and that did not turn up anything.

Given the fact that the major parts have already been replaced, is there any chance that a hard drive could suddenly give me random reboots and what not? I find it difficult to believe it could go bad after such a period of time, but who knows. Additionally, when I did take this machine back to shop, and had the guy run stress tests on it, he used his own hard drive and each time I took it back for these tests since I kept having problems, he claimed nothing happened during that time.

Another bit of info I recall is after the first time I formatted the hard drive and encountered the problem again, I put Vista 32 bit on a separate partition and had the same issue.

The things you say while testing make me think it is the HDD.

The error itself does not.

Are you completely sure he replaced each component to test as you've said? Like, did you see him with your own eyes? A lot of times, especially in busy stores, people could get shady and just tell you they did this or that while not actually doing it. Think car mechanic scenario.

I have a feeling it's the motherboard. Or perhaps he really is an honest businessman. How do the voltages look in the bios or from a Windows app that reports that? Close to what they should?

Do you have any other HDD greater than like 20 GB to try for now from an old machine or something?

In the meantime, open elevated command prompt and type chkdsk /f followed by the drive letter and :

i.e. chkdsk /f C:

Hit enter. Type y, hit enter. Then reboot.

Let H2S04 guide you with that laptop if it both it and your pc have firewire. He can get the pc to report an error to the laptop that you'd normally not be able to see.

Final thought: I have to ask this to get it out of the way. I know you'd probably have mentioned it if you were, but are you overclocking at all? Perhaps you are and even forgot about it? When I was finding my highest stable overclock, I would also get the kernel-power error and reboots because it was too high. Also check memory settings in bios. Make sure that timings are not too tight, voltage is proper for specific memory modules and frequency is stock too. If all that is, you could raise the memory voltage by .05v and/or .1v for testing.

I think this guy is pretty honest, my friends use him often for purchasing computer hardware and parts, and have known him for years. I was also present at the time when he was replacing these parts.

The other thing is that I just finished running Prime 95 and did the blended torture test, and it did not turn up any errors. I let it run for about 10 hours with my CPU and memory pretty maxed out.

I also did not overclock this machine in any way. I'll also give that chkdsk scan a shot and see if that helps. If H2S04 could give me instructions for setting up that test between my desktop and laptop, I will give that a shot.