[FireCaptain]

Hello RR,

My system seems a little more stable since Saturday, a very rough day. Yesterday I was looking for the cause of the Sidebar clock and cpu gauge freeze when I found that the problem was with the DPI being set too high, so I closed Sidebar rather than adjust the DPI back to its regular setting. I had the task manger open when I closed Sidebar and noticed that the CPU usage dropped rather dramatically. Although I’m still experiencing errors, now they don’t seem to occur randomly when I’m using Explorer or when the computer is in sleep mode. Very strange.

Yesterday I tried to make the computer crash, so I could try to narrow the possible error causes. The only things that made the computer crash was using the device manger, trying to download the AVG installation files, and using a video card stress tester. I had an error when the device manger was open, and I was trying to find the motherboard drivers. I’ve also had errors when using the “search online” function for the display drivers. As for the AVG download, it seems that every time I try to download the files, I get an error; I have no idea why. Saturday and yesterday the download would run for a few minutes before an error would occur. Last night the download got all the way to the half-way point before an error occurred. I got Video Card Stability Test from majorgeeks.com last night because I thought either the video card or its driver was causing the some of the errors. I first ran the stress test (three minutes) and then ran the benchmark test without any errors. A little while later I tried to run the stress test again, and I got an error. I tried it once more last night and once today, and I got errors both times.

Regarding the cdd.dll, although the problems that kb938194 corrects don’t exactly apply to my case, I tried to download the file anyway but had an error during the download (0x3b caused by ntoskrnl.exe and win32k.sys). I’ll try again later; I’m trying to download the avast! installation files right now. If I can’t get the file, could cdd.dll be fixed with the Windows disc and the repair function? As for SP2, I’m going to have to order the DVD from Microsoft.

I know that you have a lot to do and that reading these dump files takes a while, but if you want the latest minidump files, I’ll upload them. As for the warranty, I sent the computer to a Gateway repair depot just before my one year warranty expired. They looked at it for a day then shipped it back with a letter that says that the unit was tested and that the unit passed all tests. The letter says that “All repair work is covered under our standard 90 day parts and labor warranty, or under the original warranty, whichever is longer.” I don’t know if the ninety day warranty is just for the “repairs” they did, the problems (BSODs) that I reported to the Gateway technicians, or all repairs to my computer. Before I give them a call, I’d like to get a better idea of what the problem(s) is, so they don’t run me into a dead-end. If they won’t give me any support for my computer, I have a two year warranty with HSN, so I’ll contact them.

[joeten]

Hi you could try running superantispyware to check for nasties

[Rollin' Rog]

The main reason I am concerned about cdd.dll is that it appears to be corrupt -- no MS driver should be shown with file info "unavailable".

Try working in a "clean boot" environment when attempting to do anything that is currently likely to cause an error -- such as a scan or a download.

http://support.microsoft.com/kb/929135/en-us

Also, run SpeedFan again and this time, as well as monitoring the CPU temps under stress, click the S.M.A.R.T tab and do the "Online" analysis for your drives. Check the system drive temp against the max shown in the data given in the online analysis

I really do think this is a real hardware problem though -- but I just don't know how to isolate it short of systematically replacing things.

[FireCaptain]

RR,

Last night I clean booted and ran a thorough scan with avast!, ran the video card tester, and downloaded a few files from major geeks without any problems. Today while still in clean bootI tried a standard scan and got an error (Oxa). I also tried to download three files with Free Download Manager, succeeded with two and got an error (Ox3b) with the third. After those errors, I tried updating the display driver with the device manager and didn’t have an error. I left it open for thirty minutes and intermittently searched for updates of various drivers and didn’t receive an error. I ran the video test again (two minutes) without an error. I’ve attached the errors I’ve had since 10/24 to this post.

I did the analysis with S.M.A.R.T. and found that the hard disk had a temp of 45c and that “All of the attributes of your hard disk have normal values.” Also, the overall fitness is 94% and the overall performance is 95%. The cpu wasn’t under any stress at the time of the analysis.

I wanted to stress the cpu, so I decided to use OCCT. Before I ran the stress test, I took a look at the voltage numbers. Those were Vcore: 1.34, 3.3V: 2.99, 5V: 5.08, and 12V: 11.90. Speedfan confirms those numbers and adds: Vcore2: 1.15V, -12V: 6.64V, -5V: 8.43V, +5V: 5.59V, Vbat: 3.18. The current box next to “CPU frequency” fluctuates from 1199.9 to 1200, the original box is at 2600, and the overclock box is at -53.8%. The current and original boxes next to “bus frequency” are at 200 and the overclock box is at 0. Do any of these numbers seem strange?

I ran the cpu stress test for the minimum six minutes (I’m a little nervous about using this thing). It idled for the first minute, ran the test for the next minute, and compiled the data for the last four minutes. The core temps reached 47c for core one and 45c for core two. After the test was over, the cores temps fell to 34c for core one and 32c for core two. The other temps never budged except for temp one, which rose up a few degrees and then went back down.

Should I run OCCT for a while longer?

After the test was finished and OCCT had made its graphs, something strange happened. I clicked on the x button to close OCCT when suddenly the screen went black. At first I thought it was another error, but then the computer restarted. No error messages popped up. Event Viewer doesn’t show anything unusual--just events 10, 6008, and 15016 within the last hour. I’ve no idea as to what happened.

“I really do think this is a real hardware problem though -- but I just don't know how to isolate it short of systematically replacing things.”

I’d like to repair the computer myself because last time not only did Gateway get the diagnosis wrong but they also bent the side panels on both sides of the case. I'm apprehensive about further damages to my computer. However if you feel that we’re running out of options, I’ll contact Gateway and HSN and tell them that I’ve got hardware issues and that one of them must either fix the hardware issues or replace my computer.

[Rollin' Rog]

Seems strange that OCCT is only reporting on 1 Core. Speedfan's readings seems suspect -- particularly these:

Vcore2: 1.15V, -12V: 6.64V, -5V:

I'm not sure it is capable of reading and reporting voltage values reliably.

The current box next to “CPU frequency” fluctuates from 1199.9 to 1200, the original box is at 2600, and the overclock box is at -53.8%. The current and original boxes next to “bus frequency” are at 200 and the overclock box is at 0. Do any of these numbers seem strange?

Yes. I don't know if they are accurate, but they are certainly strange. It seems like you are running at half your nominal CPU speed.



See what you get with Everest >> http://www.softpedia.com/get/System/...-Edition.shtml

Look at Computer > Sensor; and also verify the "overclock" settings for the CPU speed.

Also, try entering your BIOS/Setup properties and look to see if the values are displayed there -- that would be the most reliable if present.

You might want to try removing the CMOS battery on the motherboard -- and in the process of removing it for a couple of minutes and replacing it you will reset the BIOS/CMOS settings

You will have to reset the time in the BIOS after doing that.

The minidumps are just more of the same.

[FireCaptain]

RR,

Sorry for the delay.

While I was downloading Everest, I checked the OCCT CPU frequency numbers and saw that the numbers would randomly change from 1199.9, 1200, 1200.7, 1599.9, 2599.8, 2600, 2600.1.

I downloaded Everest 2.2, but Windows requires a signed driver, so I couldn’t full access to the sensor data--all that it showed was the CPU temperature. In the overclock section, the clock speed was around 2600; sometimes it would go above 2600 to 2603.17 and sometimes below to 5996.72.

I downloaded CPUID and the trial version of Everest Ultimate to see if I could get any more information about the CPU frequency. CPUID and Everest Ultimate overclock had the CPU clock at 1200 and the multiplier at 6x, and sometimes CPUID would show the frequency jump to 2600.

“Also, try entering your BIOS/Setup properties and look to see if the values are displayed there -- that would be the most reliable if present.”


The only voltage numbers I could get were the in the BIOS setup were the Vcore: 1.34, 3.3V: 2.99, 5V: 5.08, and 12V: 11.90 numbers.

“You might want to try removing the CMOS battery on the motherboard”

I haven’t removed the motherboard battery because messing around with the hardware scares the heck out of me, and I also don’t know whether messing with the hardware in that way would void the warranties I have.

Just to give an update:

After the latest errors (including an 0x24 NTFS.sys error; I checked the hard drive and found zero errors), I finally decided Saturday to contact Gateway support about the errors. The tech said that the 90 warranty was still in effect. He also gave me a long list of questions to answer and asked for all of the errors codes I had. I completed that task last night and also told him about the CPU frequency. I’m now waiting for a reply.

Today I tried to cause an error while in safe mode. All of the devices and drivers were installed at the time. I got three errors. One when I tried to scan with Windows Defender and had AVAST! open but not running at the time. Avast! appears to have caused an error. It’s also caused a few other errors while in normal mode. After re-entering safe mode, I ran Defender again this time without an error. I then ran AVAST! and got 80% complete when a 0x3b error occurred. Sometime between those two errors a 0x1e error occurred. I don’t exactly remember what I was doing when that occurred. Blue screen viewer flags Ntfs.sys, nostril.exe, and timntr.sys.

I downloaded the latest NVIDIA video driver over three days and installed it this afternoon after the latest nvlddmkm.sys error. I hope that solves that error. But I need to ask a question. The exe. file was in the downloads/software folder that I created for the Free Download Manager. After entering safe mode, uninstalling the old video driver, restarting into safe mode, and installing the new video, the installer wanted to create a new folder to put the driver files in. It automatically created and offered a folder in the C: drive. I clicked ok, the driver installed, and now a NVIDIA folder is listed with the C: folders. Was I supposed to put it somewhere else? The device manager has the new driver listed, shows no conflicts, and says that the device is working properly.

I’ve attached the latest minidumps.

[Rollin' Rog]

Typically display driver installs will create their own folders -- where they are is generally irrelevant once the setup is complete as the drivers will be placed in the system32 directory.

I'm not going to look at the minidumps any more as I don't think they have anything further to offer.

The variations in CPU frequency are indeed odd -- and the only thing I can offer is that typically CPU's ramp down in frequency to compensate for overheating -- and that might be what is happening even though you are not seeing it in the temp monitors.

I hope Gateway comes through for you since other than an inherent motherboard problem -- I don't know what could cause this variety of symptoms ---- although your experience with Avast would make me want to uninstall it for testing.

[FireCaptain]

RR,

I was reading about how memory tests sometimes miss a problem when both sticks of RAM are tested together. I wanted to know whether memtest86+ missed something when I ran the test with both sticks of RAM, so I decided to test the memory sticks individually.

The first stick made two passes in both slots and didn't find any errors.

The second stick, however, made two passes in the first slot and found eight errors on test five. I then tested it in the second slot for two passes but didn't find any errors.

I re-tested the first stick in the first slot for two passes and didn't find any errors.

I re-tested the second stick in the first slot for one-and-half passes and found two errors on test five.

I don't know if the memory stick or if the motherboard is causing the problems, but I've alerted Gateway about the memory stick, and I really hope a new memory stick cures the BSODs.

I've tried to tell the techs about a possible motherboard issue, but I don't think they don't fully read the e-mails or understand the problem. If it turns out to be a motherboard problem, I don't have a lot of confidence that they'd take care of it, so I'll probably have to go with HSN if the memory stick isn't the answer.

Anyway, I'm going to mark this thread as solved, and I'd like to say thank you for your time and help. I greatly appreciate it.

[Rollin' Rog]

That's definitely a good find.

It does sound like the second stick has problems -- however it might be an issue with the contacts and may not necessarily require replacement.

These can be cleaned. The method I've used occasionally is to use a rubber type eraser to "polish" the contacts, then wipe with an alcohol swab and let dry.

Naturally you want to avoid electrostatic issues when handling ram, so do occasionally ground your hands on a metal surface -- and best to avoid working on a carpeted surface if possible.

[jspidey]

Originally Posted by Rollin' Rog

The variations in CPU frequency are indeed odd -- and the only thing I can offer is that typically CPU's ramp down in frequency to compensate for overheating -- and that might be what is happening even though you are not seeing it in the temp monitors.

This would most likely be Intel's Speedstep which will adjust the processor speed:
http://en.wikipedia.org/wiki/SpeedStep

Originally Posted by FireCaptain

RR,

I was reading about how memory tests sometimes miss a problem when both sticks of RAM are tested together. I wanted to know whether memtest86+ missed something when I ran the test with both sticks of RAM, so I decided to test the memory sticks individually

I should try this too, i never heard of the test not finding the issue with both in at the same time but it might work for me too.

I just tried a different vid card and was able to get it to crash still, although it did take much more effort. Still, SpeedFan showed the Core temps getting to about 45C right before it crashed.

Good luck with your issue. I know that I still have some work to do on mine. The one thing I wanted to ask you FireCaptain is if your CPU fan will automatically adjust when the temp get's hot? Mine will not adjust automatically. I'd recommend you set yours to full on in the BIOS and test. It's loud, but my guess would be you wont have issues with that fan on high -- I didnt...