Desktop PC reboots randomly, sfc errors each time

[derstrom8]

Hi folks--

I have been having this problem intermittently over the past year or so. It had settled down for a while but then it started up again just recently and has become very annoying. As the title states, my desktop PC (specs are on my profile page) started shutting down and restarting randomly quite a while ago, and it continues to cycle that way for several times before eventually settling down on the Windows 7 login screen. I replaced almost all the hardware -- motherboard, CPU, video cards, RAM, etc -- except for the hard drive and the power supply. I ran various tests on the hard drive and I am fairly confident it is not the issue. I eventually decided the problem was the PSU. I replaced the main filter capacitors and the system seemed to run very smoothly, until just a few weeks ago. Now it has started doing the same thing--shutting off in the middle of a session for no apparent reason and restarting. I got a blue screen the other day complaining about the hardware, but after some tests I came up empty--none of the hardware seemed to be having issues.

Anyway, this is the reason I'm posting here:

Whenever the PC reboots, I try to boot into safe mode with command prompt and run the sfc /scannow command. It always comes up with errors--if the computer doesn't restart in the middle of the scan--most often complaining about shell32.dll (according the the CBS log). It says it repairs it, but then as soon as I start using my computer again it shuts down again. I restart the sfc and get the same error. Now I'm not sure if the system files are damaged because the power supply quits (if that is, indeed, the issue), or if the system shuts down because the files are damaged. This is the reason I'm asking here.

Before anyone asks, yes--I've tried re-seating the RAM, run chkdsk (no issues found), memtest turns up nothing unusual, and CPU-Z shows nothing out of the ordinary. Keep in mind though that I'm lucky if I'm able to complete any of these tests without the computer shutting down in the middle of it.

Anyway, if anyone could answer my question about shell32.dll, I'd be very grateful. I have ordered a new power supply and it should be here within a week, but until then I'm pretty much out of options. Any and all suggestions/recommendations are very much appreciated!

Regards,
Matt L.

 

[ICIT2LOL]

Hello and welcome derstrom8 when I started reading your post I immediately thought PSU but on seeing the shell32 problem am not real sure of what it is that has happened. Take a look at this for some ideas though.

How To Fix Shell32.dll is Missing or Not Found Errors

 

[derstrom8]

Hello and welcome derstrom8 when I started reading your post I immediately thought PSU but on seeing the shell32 problem am not real sure of what it is that has happened. Take a look at this for some ideas though.

How To Fix Shell32.dll is Missing or Not Found Errors

Thank you for the welcome, ICit2lol, and thanks for the link. I think I'd found that site once in my googling but I guess I didn't read it very thoroughly. It mentions starting in safe mode, and the problem with that is that it still shuts down when I'm logged in that way. So any testing I'd do would have to be very quickly. I'll explore some of the options discussed in that link and will get back to you.

Regards,
Matt

 

[derstrom8]

Well I haven't tried any shell32.dll recovery programs yet, but I thought I'd post this--

I finally got it logged in in safe mode and removed some windows 7 themes I had installed a while back. I figured they might be messing with the system somehow. I also found some unrecognized programs (that I don't recall ever downloading) that were most certainly not supposed to be there. I uninstalled them and ran SpyBot Search & Destroy, and it found two root classes -- HKLM\SOFTWARE\classes\sdp and HKLM\SOFTWARE\classes\sdp x64 -- that were rated as high danger. I removed them (through SpyBot) and will see what happens. If the issue continues I'll try a recovery program for shell32. I also ordered a new PSU so I'll be replacing that.

So you don't think an unexpected shutdown would corrupt shell32, or could that be the main issue?

I'll have to get out my oscilloscope and test the outputs from the PSU, see if there's any significant ripple or noise that shouldn't be there. Anyone here have any recommendations for that? All I have are a couple analog oscilloscopes from the mid-to-late '70s, one from the '60s, and one from the '50s (estimated). No digital ones, unfortunately.

Much obliged!

 

[derstrom8]

Well, no such luck. The computer died when I was running a MalwareBytes scan, and hasn't stayed running long enough for me to do anything with it. I'm just going to have to wait until my new PSU comes, and we'll see how it works after that.

 

[ICIT2LOL]

Yep Matt wait for it mate and then see what the go is - seems like the power is just dribbling away and leaving you nothing.

 

[derstrom8]

Hi folks--

The power supply arrived a couple of weeks ago and I installed it and didn't have any problems for a while. However, just a couple of days ago my computer started going into a boot cycle again. Thankfully it hasn't shut down randomly on me like it used to (yet), but it tries to boot three times before finally succeeding. i.e. Tries to boot, fails, restarts, tries to boot, fails, restarts, tries to boot, succeeds. It is very consistent--it always tries to boot three times before it actually works. I have now replaced almost everything in my machine--CPU, motherboard, power supply, video card, RAM--everything but the hard drive--and I still have the same issue. Normally I would assume this means it's a a problem with the OS or HDD, but I tried running from a ubuntu live CD once and it still had the same problem (think I mentioned this already).

So I am officially completely stumped. Do you guys have ideas what may be causing this problem?

Many thanks,
Matt

 

[ICIT2LOL]

Ok mate try this seeing as you have replaced everything I didn't notice the PSU.

PSU- Jump Start

for a reference
PowerSupply 101: A Reference Of Specifications - Power Supplies asyou will see there is a +/- 5% variance that can make all the difference so for example for the +12v rail/s 11.4v is way too low and 12.6v way too high.

Plus remember correct volts do not necessarily mean correct current (amps - inturn watts)
WARNING DO NOT attempt to take the PSU apart under any circumstances

 

[derstrom8]

Thanks for the response--

The PSU is brand new, and it's a good company--Corsair. It's rated for 650W, which should be enough for what I have (system specs are listed on my profile). I did use the jump process on my old PSU and tested it with a 'scope, but there's really not much I can do without an adjustable load. Which part of the link did you want me to follow, exactly?

Thanks,
Matt

 

[ICIT2LOL]

Yes mate I know I have exactly the same one in my Sandy Bridge and a HX 1050 in my Ivy but if I were having these problems if only for purposes of elimination I would just test those leads and sockets. Like I mentioned new doesn't necessarily mean it is good it is like anything you buy cars fridges whatever it can go wrong from day 1 to whenever.

Now in the 101 ref there is a part (Section 2 - Voltage rails) which refers to the Power_Good signal maybe worth checking that out to see what the volts are there because if they go outside t the limits then it is a problem with the PSU. That particular pin is number 8 (grey cable) I usually ground on the mobo fixing screws or one of the blacks I find the screws the easiest to use because of access.
Another thing to try is the resistance between the black lead/s and the fixing screws - should be zero ohms. Oh and by the way don't use an analogue multimeter because they inject volts for their operation you need a digital meter but I am guessing you know this already but thought I had better mention in case. .

 

[derstrom8]

Yes mate I know I have exactly the same one in my Sandy Bridge and a HX 1050 in my Ivy but if I were having these problems if only for purposes of elimination I would just test those leads and sockets. Like I mentioned new doesn't necessarily mean it is good it is like anything you buy cars fridges whatever it can go wrong from day 1 to whenever.

I understand, I'll give it a shot. But what is really confusing me is that it consistently tries to reboot 3 times before starting up. That doesn't sound like a power issue anymore, that sounds more like software telling it to try one thing, if that doesn't work, try another thing, etc. until it works. But then again, I'm just thinking out loud, I honestly have no idea.

Now in the 101 ref there is a part (Section 2 - Voltage rails) which refers to the Power_Good signal maybe worth checking that out to see what the volts are there because if they go outside t the limits then it is a problem with the PSU. That particular pin is number 8 (grey cable) I usually ground on the mobo fixing screws or one of the blacks I find the screws the easiest to use because of access.
Another thing to try is the resistance between the black lead/s and the fixing screws - should be zero ohms. Oh and by the way don't use an analogue multimeter because they inject volts for their operation you need a digital meter but I am guessing you know this already but thought I had better mention in case. .

Interesting, I'll definitely have to check those. I am an electronic engineer so I understand the bit about the DMM, but I appreciate mentioning it anyway. I know I probably won't be the only person reading this thread. It'll come in useful to someone

Regards

 

[derstrom8]

Just another followup:

My computer has bluescreened twice in the past week. It gives me the following STOP error:

stop: 0x0000001A (0x0000000000000403, 0xfffff6800003a4c8, 0xafc0000206c5b825, 0xfffff8a010cb64a0)

I did some research and it seems it has something to do with the RAM (as I was guessing). However, the RAM is practically brand new, from the same manufacturer, same model number, etc. I have 4x 4GB sticks. I wouldn't expect them to be incompatible. I'm running the Windows Memory Diagnostic Tool but it is at 90% and has not found anything yet.

I've looked at several links explaining ways of fixing this problem, but so far none of them have helped. Any thoughts? If this is the wrong place to post this, please let me know. I don't know if sevenforums covers RAM issues or not....

Thanks

EDIT: Further information--

The memory diagnostic turned up nothing, so I was probably wrong about the RAM after all. I am looking through the event viewer and have found a couple of errors, both of which saying "\SystemRoot\SysWow64\Drivers\driverx.sys has been blocked from loading due to incompatibility with the system" or simply "This driver has been blocked from loading". This shows up several times. sfc /scannow also turns up dirty, but I'm not sure what to look for in the cbs.log file.

 

[Britton30]

You probably should start a thread in the BSOD forum, as per, http://www.sevenforums.com/bsod-help-support/96879-blue-screen-death-bsod-posting-instructions.html


RAM (memory) test:

This will require memtest on a bootable CD or USB stick.
RAM - Test with Memtest86+​

Hello Matt, you might want to start a thread in the BSOD forum, following these instructions.
http://www.sevenforums.com/bsod-help-support/96879-blue-screen-death-bsod-posting-instructions.html


This a more comprehensive RAM test, the w7 can be wrong.

 

[derstrom8]

Thank you very much Britton30. I am downloading the SF debug tool now and will post to the BSOD forum.

Regards

EDIT: For future reference, here is the thread I submitted to the BSOD forum:

http://www.sevenforums.com/bsod-hel...mes-per-week-stop-0x0000001a.html#post2669481

 

[Britton30]

You're welcome, you'll get more BSOD specific help there but they are busy and may take time. If you run the RAM test as I suggested let them know that.

 

[derstrom8]

You're welcome, you'll get more BSOD specific help there but they are busy and may take time. If you run the RAM test as I suggested let them know that.

Hello Britton30,

I am running Memtest86 and oh my, 1000+ Errors in the first 15 minutes! Looks like I'll be at this a while....

Am I correct in the assumption that I should test it with all 4 sticks installed, then test it one stick at a time?

Thanks,
Matt

 

[Britton30]

Yes, stop the test if even one error. Then as told, test each stick in each slot to find either a bad slot or RAM module.

 

[derstrom8]

Yes, stop the test if even one error. Then as told, test each stick in each slot to find either a bad slot or RAM module.

Well I ran a few (fairly brief) tests and found one card that caused the computer to not boot at all. I expect it was the cause of most of the problems (if not all of them).

I installed the three working ones in the proper slots and am running memtest once more, and so far I see no errors. It certainly appears that one card was the problem.

Thanks for the link, it was VERY helpful. I thought I'd run memtest on this computer before, but perhaps that was my laptop. Anyway, many, many thanks! I believe you solved my problem. I will let you know what happens from here.

Regards,
Matt

 

[Britton30]

A sure way is to use only the suspect RAM in each slot and see if it fails.