PC keeps becoming unable to boot - help me diagnose please

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520
I have a HP P6710F desktop system with Windows 7 that had been stable and has had no hardware changes or significant (like reinstalling or changing OS) software changes for at least 8 months. The configuration is 8GB memory and a SSD drive as the boot/OS drive and a hard disk for other storage.

Recently it started crashing. It would just suddenly start rebooting itself. It would run for a few hours and then reboot. Eventually it would not boot and Windows startup repair would not fix it.

My diagnostic steps so far:
I ran memtest through a full cycle and it did not find any errors.

I was thinking that it might be a failing SSD drive, so I did a fresh install of windows on the hard disk. The install went fine, so I tried using prime95 as a CPU stress test to see if there was a temperature issue or a power supply draw issue. It did not cause any crashes. I could access the SSD partition fine and I made a complete backup of it and I could read it fine. I tried copying a bunch of large files to it and didn't see any errors.

Next I did a fresh install of windows 7 onto the SSD. No problems during the install and it booted OK. I started doing normal post install tasks. Installed Norton Internet Security and updated it. Ran the Windows Update to get all the updates. Seemed to work fine. Then I don't remember exactly what I was doing, but I rebooted and booting got stuck at the splash screen. Power cycle and it won't boot and startup repair won't fix it.

So now I'm thinking it is probably the SSD, but I'll give it one more try. This time I actually disconnect the hard disk to make sure it cannot be affecting anything. Complete fresh Windows install to the SSD including format of both the system and OS partitions. Same thing as above. I got through most of the Windows updates and during one of the reboots, it freezes on the starting windows screen and won't boot and startup repair won't fix.

Now I'm sure it is the SSD and I disconnect it and do a fresh install onto the hard disk. Get through the install and Windows updates and during one of the reboots, it happens again. Won't boot and startup repair won't fix it.

Now I can't think of anything else but that it is some kind of hardware failing and I need to buy a new computer. Can anyone offer me any other possibilities or diagnostics I can to to narrow it down further?

 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520

Thanks, both of you. I would have been thinking about the power supply if it just kept rebooting, but something is trashing the data on the disk. Would a power supply problem do that?

I don't have another power supply to try swapping. Any suggestions of diagnotic steps I could take to prove it is the power supply before I go out and buy one? Thanks.
 
I dont have any suggestions for testing other than trying another one. Can always buy one at a local brick and mortar store and take it back if it doesnt fix the issue. A dying PSU can cause a host of problems with your PC, some things youd never think it would effect, it does.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I found an old PC that was not being used anymore and took that power supply and replaced the one in the problem system. It is a 300 watt power supply and the original one in the HP P6710F is 250 watts. The current rating is higher on all voltages.

Same problem. I fresh install Windows 7 and during one of the reboots required when doing the windows updates, it won't boot and startup repair can't fix it. So I think that rules out the power supply as the issue. Anyone have any other ideas?

 
bad ram, electrical short, temps.

start with cleaning your computer, top to bottom, take it all apart. if that doesn't work, take out all but 1 stick of ram and run memtest86 on it for a few hours. then try with the other stick then try with them both. see if the ram is alright.

finally keep the side of your case off and use a room fan to blow cold air into it. see if the problem goes away.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I don't think that it is the RAM. I've done some more experimenting since replacing the power supply. I took out the 2 4GB RAM sticks (they were an upgrade since I bought the PC) and put back the original 2 2GB RAM sticks it came with. Still had the same problem.

I guess it could be temperature related, but I think not likely because it started when the weather was pretty cool and thought the weather has warmed up a bunch, it doesn't seem to happen faster now. And if it is temperature related, what is the sensitive component? I think it is down to the motherboard or processor and if I have to replace them, I might as well replace the whole PC.

 
there are a lot of parts that can overheat.

i remember helping a guy with an overheating problem that completely vexed him. he was having a similar data corruption and crashing issue, but none of the temp sensors said their was an issue and the "finger test" didn't turn up anything (nothing was amazingly hot to touch).

turned out he had a dead spot in his case air circulation around his hard drive, the temps around the drive were almost blast furnace like... 70C-80C; he actually cooked the hard drive and replaced it before calling me in to look at the problem. the problem went away when i stuck a room fan in the openning of the side of his case, so i knew it was an air circulation issue, just took me 20 minutes to figure out what was happening and another 30minutes to adjust the cables and rig a small 80mm fan up to fix the problem.

so don't just dismiss the heat issue until you are certain. the best way to be sure is to pop the side of the case off, and stick a room fan in the opening. that will blow enough air over the parts NOTHING should overheat, even if a fan is malfunctioning (like in a psu). if your problem vanishes you know for sure it's a temp thing.

As to the parts that can overheat (or atleast that i've personally seen overheat)
CPU
Northbridge
Southbridge
MOSFETS
GPU
PSU
Hard Drives
SSD
Ram

in short most stuff can overheat and fail. thats why good case airflow is so important.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


Yes, I see how that makes sense. I was thinking that if it was a heat problem caused by a bad component, there would not be a cost effective solution. But if it is caused by bad air circulation, another fan could fix it inexpensively. I early on blew out all the dust from the case and heat sinks and anywhere it was settled. I'll try the room fan idea and report back. Thanks.

 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I tried the above and still have the same problem. I fresh install windows and then rebooting during a windows update reboot, the system won't boot and startup repair won't fix it. One thing I've taken to doing is running sfc /verifyonly before rebooting to see if the boot/startup files are corrupted while ruinning. It always comes back with no errors, although it won't run after the windows update process because it says something about a restart being pending.

Now the failure has become pretty reliable after a windows update, I am starting to wonder if I really have a hardware issue at all. Maybe there originally was a power supply/heat/hardware problem, but now I have done something during the debugging that is now triggering a Windows software problem. Maybe something got messed up in the boot partition that isn't getting fixed with my fresh install. I'll explain what I am doing and maybe someone can explain if there is a cleaner way to do it to make sure.

I have a WIndows 7 bootable install DVD that I made. I don't remember where I got the image, but it is the same disk I used when I fresh installed Windows when I got a SSD back in January and it installed fine back then and ran fine until I started having problems.

I boot from the DVD. Now the disk I'm installing to (either the SSD or hard disk, I've tried installling to both) has a system partition (partition 1) of 100 MB and a primary partition (partition 2) of 80+ GB for the OS. I format both those partitions. I'm not sure the format of the system partition makes sense, but I was trying to get rid of any old bootloader info. I assumed that the install process would write whatever is needed into that system partition. If there is a surer way of making sure the install gets a clean system partition to work with, please let me know.

I install the OS onto partition 2. Then after the system reboots and sets up the user/desktop, I install a network driver that is needed, install Norton Internet Security, run Norton Live Update repeatedly until it says no updates are found. Then I run the Windows updater. The Windows update eventually needs a reboot and then the system won't come up. I haven't kept perfect notes, but I know that sometimes the first reboot after Windows updates has worked and it fails on some subsequent reboot. Lately it seems like it has been failing to boot on the first reboot required by Windows update.

Is there anything in that process that could be causing a problem? Or anything I should do differently?
 
you've got me puzzled.

let me review basically what's happening and what you've tried. i want to be sure i understand this completely.

1) windows blue screen crashes on loading (SSD)
2) tried to reinstall, still same problem
3) tried a new hard drive thinking ssd was faulty
4) same problem, will install windows just fine, but on restart it will hang up and crash.
5) problem started randomly
6) checked the ram with memtest and in different configurations
7) checked for overheating
8) cleaned case
9) tried different psu

that about sums it up?

cause if that is all true, i think the real problem is your motherboard chipset drivers. I think the reason why its working for the windows install is because windows is using some generic drivers for the install, then at some point it's updating them with the "microsoft approved" drivers, which were likely updated during a recent windows update starting you on this horrible problem in the first place and are breaking something along the way.

What i would do is load into safe mode (you said that works right?) and get an older version of the drivers for the motherboard from the manufacturers webpage, update it all that way and see what happens.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I think your summary is basically correct. The initial problem was just the system rebooting without any error messages. Initially it rebooted OK and ran for a few hours. This happened several times. Then one time it wouldn't boot and startup repair wouldn't fix it. That is when I started reinstalling and all the other steps. I don't think that since I've started reinstalling (which is after opening the case, cleaning and leaving the case open during debugging), the system has ever rebooted on its own as it initially did. So maybe the initial reboots were heat related (there was a lot of dust including between the fins of the CPU heatsink) and now the problem is something else, perhaps driver related as you suggest.

I'll try to find HP drivers and try what you suggest. Thanks.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I can't get into safe mode once the system has become unbootable. WIndows tries to roll back updates and hangs.

I found two drivers that might be relevant:

Original AMD Chipset Driver
Original AMD SATA Controller Driver

I tried installing them after a fresh install, but before WIndows update in the hopes that update might not replace them if they were updated since fresh install. Still ran into the same problem.

I tried fresh install, the windows update, then install those drivers before rebooting. Still have the same problem.

I think that it is definitely windows update that is causing the problem, so I am going to keep playing with it and see if I can figure out how.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520
OP here. It appears to be a single windows update is responsible for my inability to boot after running Windows update. Believe it or not, it is the IE 10 update. If I fresh install and then select only the IE 10 update from the list of approx 120 updates, the system will not boot after the update. If I exclude IE 10 and install all the remaining updates, the system boots fine. Thank you Microsoft "engineers" for continuing t treat every application or feature as part of the operating system.

I still may have an underlying problem that started this whole process or it may have been fixed by cleaning the system or replacing the power supply. I'll have to run for a while to see if the system is stable. I did run prime95 for about 12 hours while monitoring temperatures with HWmonitor. The system didn't crash and the temperatures stayed pretty low.
 


that's one hell of a bug. i'd talk to microsoft about it. don't deal with the entry level tech support. you're way past their blanket suggestions like "re-install windows" or "load into safe mode". Insist on getting help from someone who knows what they're talking about and explain the situation.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


I may have been too hasty ragging on Microsoft about the IE 10 update. The underlying problem has surfaced again. The PC rebooted itself again with no provocation. No blue screen or anything - one second it was running and the next it was booting. And it failed to complete booting and startup repair will not repair it. So there is something that is causing pretty basic files or boot stuff to get messed up. It may be that the IE 10 update was not messing with anything that low level itself, but was somehow triggering the same problem.

I am once again at a total loss of how to find the root problem. The next thing I can think of trying is loading Linux on the system and seeing if it will run. I'm not going to do that until I get a replacement PC, so I can copy all my data. I guess I can try running Linux off the distribution disk and see if it runs for a while. Only problem with that is I've never seen the spontaneous reboot when I've booted off the Windows distribution disk, so it may require booting from the hard drive or SSD to trigger the problem.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520


Thanks for all your help. I tried running Linux off the distribution media for a day and it didn't crash, but I didn't really do much but leave it running. Next I'll try installing Linux, but I'm going to wait until I have a new PC set up and all my data moved. I'll be busy setting up the new PC, so I won't be debugging this system for a while.
 

pdxman1

Distinguished
Sep 10, 2013
22
0
18,520
OP here. I installed Ubuntu Linux on this PC and have run it for a few days without any crashes. So I think that I either have a hardware problem that Windows triggers, but Linux uses the hardware differently and doesn't trigger or triggers less often.

Or I have a Windows bug that I haven't been able to narrow down. I think this is more likely.

I'll be turning this PC into primarily a Linux system, but I'd still like to have the option of running Windows 7 on it occasionally, so I'd still like to figure out a solution to the problem.