BSOD booting from Win7 installation CD/USB

"Tried with SATA configured as IDE in the BIOS, no effect"... Is the SSD normally configured on a SATA 3 port and set as AHCI?

 

Yes. All 8 ports on this motherboard are SATA3 and all were set as AHCI by default. This was the setting I used for all of my tests, except for that one.

 

Can you temporarily install a standard Hard Drive set as AHCI and try the Widows install and Windows Updates again? How big is the SSD?

 

I disconnected the SSD and attached an ordinary Seagate 500GB 5400rpm drive. I verified that the SATA ports are set as AHCI in the BIOS. Same results: I can't boot the Windows 7 installer, 4 bluescreens after 4 attempts (2 attempts with the installation CD, 2 attempts with the installation USB stick). As a second test I disconnected the Seagate and left the build without any hard drive installed, 4 bluescreens after 4 attempts.

The SSD is a SanDisk 128 GB drive, the one recommended by Steve Walton in his "Budget Box" buying guide here: https://www.techspot.com/guides/buying/page2.html.

 

Can you post the minidumps after you re-install the SSD by going to C:\windows\minidumps. Select all the .dmp files and right-click sending the files to a compressed(zipped)folder. Upload this folder as a file here. The blue screens you posted are showing a possible hardware fault, but it may be false

 

I would do this, but the 9 .dmp files in that directory are all empty. Zero bytes each. For comparison, I looked at the Minidump directory of my 2 working PCs and they had a couple of old .dmp files each and those are not empty. I'm guessing you won't find these empty .dmps to be of any value. If I do get one with data I'll post it. But for now, what's next? I greatly appreciate your help so far.

 

At this point, I would replace that motherboard with something around the $130 range...

 

I gotta say, that's the best idea I've heard all month. I'll post an update after I do so. Thanks.

 

Yes, the most important part of a computer is the motherboard... Even in a budget build. I'm building a system for a friend... $650 total

 

Update: Replaced the Gigabyte motherboard with an Asus A88X-Pro FM2+ ($128 when ordered). Beautiful ATX board, fits snug in the case. BIOS was too old for the CPU, flashed to the latest version. The Windows 7 installation CD booted but the OS never finished installation onto the SSD; it crashed after the first restart. Attempts to boot from the SSD after this point result in the Windows Error Recovery screen reporting either "Windows failed to start" or "Windows did not shut down successfully." From here it either loops around (I.e. crashes again) or comes to rest at the "Windows installation cannot proceed" dialog box. OK at this box restarts and around and around we go. Bluescreens occasionally appear, but they are very brief before the next restart is underway. The Minidump directory remains empty.

Observed the same behavior when I replaced the SSD with the Seagate HD.

One difference is that the Asus has Q-code LEDs. During the first stage of the Windows 7 installation (prior to the first reboot), the code is AA. While the Windows Error Recovery screen is up, the code is 66. When resting at the "Windows installation cannot proceed" dialog, the code is AA. When I boot from the Hirens USB stick, which always succeeds, the code is 00.

Before I installed this motherboard I used a multimeter to test the voltages at each pin of the PSU. All voltages are within tolerance.

Any idea where this build keeps going so wrong? It is becoming the bane of my existence.

 

Could it be Windows 7 with the latest kit? You can download a 3 months free trial of Windows 8.1 and Windows 10 development ISOs were recently made available. If they install OK you know it's not a problem with your build.

 

Thanks. I created a Windows 10 Tech Preview installation disk. When I try to boot from it, it displays the cyan Windows logo for a bit, then briefly displays the "Your PC ran into a problem and needs to restart" message. I also created a Windows 8.1 installation disk and the result is the same. (I verified these disks on a working PC and was able to boot to the drive selection screen.) So it has to be the build, right?

I've now swapped every component except the CPU and the case, and I see essentially the same behavior: the system crashes either booting from the OS installation disk or prior to completion of the OS installation. Does anyone think the CPU (A8-7600) is to blame? I've never seen the BIOS report a temp over 41C, and if I just leave it idling at the BIOS, it settles at 36C with the tiny stock cooler. Is there a test I can run to check the CPU? I have subjected it to prime95 a few times, but it never finds errors. (FYI I can always boot from the Hiren's CD and run any tests I want for as long as I want, no crashses.)

 

You've been very thorough and yes the cpu does seem to be the leading supect. I can't see any posts about a problem similar to yours though.

 

Thanks. One of the reviews of this CPU at Newegg suggests a similar experience, though he at least was able to get his OS installed. Come to think of it, I did too initially, though my build never got past the hundreds of Windows updates.

I've ordered a replacement CPU and will post an update after I install it. It's safe to install the replacement CPU onto this motherboard, right? Bad CPUs don't readily destroy motherboards, right? Just trying to avoid an endless cycle of component replacement.

 

Update: I replaced the CPU with an A10-7800 and added a D92 cooler (the earlier motherboard replacement nudged me beyond a purely "budget level" build anyway so I figured why not). The Windows 7 installation completed successfully, including all updates. All programs purchased by the client were installed and are working. A half day of prime95, 4-worker testing yielded a thermal margin that was no less than 48C as reported by AMD OverDrive.

So it was a faulty CPU that was my problem all along, yet it took me much longer than the 30-day CPU replacement window to determine this, mostly because I assumed that prime95's error-free reports were telling me that the CPU was fine. Someone is working on a program that can detect computationally fine, though otherwise unusable CPUs in the future, right?

Thanks bazz2004 and TMagic650 for your assistance. Happy New Year!