Solved Fresh Build - Constant BSOD - Multiple Codes - 3b, 50, a

[LP81]

We also know that you are having problems with w7 64-bit, but can you try the 32-bit version instead?

Note that when carrying out testing such as this, you should set your BIOS to its defaults, and should stop over/under clocking. You should also ensure that you deselect the option to automatically activate when online.

I only have a 64-bit version DVD/License, no access to a 32-bit version.

I guess it's time to RMA the CPU and hope for the best on the next one that comes in - this makes me mad because I've had to now send back the CPA, Hard Drive, PSU and Case that I ordered (all from Amazon) they are going to think I'm nuts that everything they sent was broken


Regarding the "automatically Activate when online" are you referencing Win-7 Activation? What does activating it change?

 

[Dwarf]

Sorry, I didn't realise that you only had a 64-bit version of W7. I only suggested what I would have done in your situation.

As regards activation, it is only to save you having potential problems with it should you activate repeatedly within a short time scale. By unchecking that option, you have upto 30 days to ascertain whether everything is working as it should.

Hopefully, you won't have any problem with the RMA of your CPU, but what was the reason for the other returns? Were they connected with this problem?

 

[James7679]

but what was the reason for the other returns? Were they connected with this problem?

I'm just going to take a stab here and say that he either got a crap PSU or a defective one and it cooked everything. That's just a guess though.

 

[LP81]

Hopefully, you won't have any problem with the RMA of your CPU, but what was the reason for the other returns? Were they connected with this problem?

Little bit of A, little bit of B...

The Case I bought just seemed super cramped and loud compared to my old Cooler Master so I returned it and put the new components into the old case

PSU - I thought I ordered the 650 Watt version, but the Amazon Page defaulted to the 550W version and I didn't notice til it arrived so I sent it back and bought the X4 750W Modular one in place

Hard Drive - Hitachi 1TB HD - would throw up a "Power Error" on the Motherboard and wouldn't even POST. Determined it was the HD based on testing multiple cables and multiple HDs on the same Cables to the PSU/MOBO. RMAed it and bought a Western Digital in place.

And now sending back the CPU.... on account of because...

 

[Dave76]

Try a couple quick tests, take out two RAM cards, be sure the remaining two cards are in the correct motherboard slots, check your motherboard manual.

Boot the system, see if it is better or if it runs without crashing.
If it crashes again, then remove another RAM card and see if it will run with one card installed.
Swap out RAM cards if it is still throwing BSoDs.

You have tested the RAM cards with memtest86+, which indicates they are working. When running in Win7 they may still be throwing an error. This can usually be fixed by adjusting the settings, usually voltage.
With 4GB RAM or more you may need to bump the IMC (Integrated Memory Controller) voltage and possibly the DRAM voltage. This depends on how several of the hardware components are interacting.
Your VTT is low at 1.05v with 8GB RAM.
The quick test mentioned above should verify this, to get the 8GB RAM stable enough to run bump the VTT to 1.15v.
This may allow the system to run but, may need further adjustments.

Let us know if this helps.

Here is some info on your PSU:
PSU REVIEW DATABASE Ultra Products

I'm not familiar with the manufacturer Andyson, this may need to be looked into if the RAM tests above don't resolve the problem.

 

[LP81]

Hi Dave,

I just sent the CPU back for an RMA, I won't have a new one in hand until Wednesday. Once I get it, I will see if it helps and if not I'll move onto your RAM suggestion above. Thanks

 

[Dave76]

Hope the RMA solves the issue, either way let us know.

 

[LP81]

So the new CPU showed up while I was at work and I was all exited to put it into the computer when I noticed something that makes me think it's really the motherboard that is bad.

I took multiple photos of the CPU slot with the light coming from different angles and to me, it look like one or two of the pins are misaligned...

When I put the original CPU in, I thought to myself that it didn't sound right going in, like something was scraping, but didn't really think much of it.

Any thoughts on if this looks bad and if I should RMA the mobo to? I assume it's not something I should try to realign myself obviously

 

[James7679]

Yeah, that's not good. You can try and straighten them yourself with a pair of good tweezers(with the board out obviously). Or you can RMA it. Bent pins will do all kinds of bad stuff. Good job on catching that before installing the new processor.

 

[Vir Gnarus]

I can add that those do look like bent pins. Understand that bending them back may not resolve the issue, as there is metal fatigue that will occur through bending them in the first place and bending them back into place. Pin and hole placement is very meticulous and sensitive for ZIF sockets, so having it manipulated even slightly can pose fatal consequences.

I'm curious as to the purpose of those specific pins as they relate to your symptoms involving the corrupted bytes in your CPU registers. The corrupted bytes always appear to be in the exact same placement every time, which is very consistent. I'm curious if those pins are directly related to these bytes. However, that will involve me having to read up land listings and all that jazz, which is beyond me (I think a mere peon like myself doesn't even have access to much of it). It's still very interesting, though.

Anyways, hope you get things fixed up. Replacing the motherboard can be quite a chore, so I'd hate to see this not work for some reason.

 

[LP81]

Well...

New CPU...
New Mobo...

Loaded up OS.... dumped again...

Updated zip attached with 29 dumps ... Went to run an updated perfmon and it dumped again... but was able to get it run the next time - dumped again before getting it off the computer... minimum 31 dumps now.

 

[Vir Gnarus]

Curious. I do see the problem continue on the first new crashdump (I'm assuming you installed the new CPU/Mobo today). However all of them relate to threads dealing with disk I/O, such as paging file activity (Readyboost/Superfetch) and other misc. disk I/O (NTFS/Volume activity).

I find it extremely odd the same problem popped up again on the first crash but not really on the others (at least not obviously discernable). Unless you somehow got a bad CPU/Mobo again (it's happened), I can't really figure anything else out other than blame something involving your drives, like the drive controller. I'll have to look into it further. In the meantime, repeat the same diagnostic testing as previously recommended by me and others.

Btw, did you get the same exact mobo/CPU? We may very well be dealing with a mobo incompatibility with your RAM unless you managed to change that with a completely new brand/model as well.

 

[LP81]

Same exact Mobo & CPU model numbers as before. Just installed them this evening, booted and there's the last few dumps in that zip.

CPU was from Amazon
Mobo was another retailer (because amazon was sold out and only gave a refund)

RAM is the ONLY thing I haven't RMAed yet. It's the only thing left of the original items I purchased to build the computer.

It seems to be the last option that I might want to explore different RAM...

I'll see if I can get any of the tests running since the computer will BSOD after about 5 min of up time regardless of what I am doing on it.

It's weird that XP 32-bit runs perfectly stable but 7 64-bit just crashes (both fresh installation and and older one on a different HD)

 

[Dave76]

Try this below, it's not unusual for a new system to not boot due to some default BIOS settings, especially RAM settings.

Try a couple quick tests, take out two RAM cards, be sure the remaining two cards are in the correct motherboard slots, check your motherboard manual.

Boot the system, see if it is better or if it runs without crashing.
If it crashes again, then remove another RAM card and see if it will run with one card installed.
Swap out RAM cards if it is still throwing BSoDs.

You have tested the RAM cards with memtest86+, which indicates they are working. When running in Win7 they may still be throwing an error. This can usually be fixed by adjusting the settings, usually voltage.
With 4GB RAM or more you may need to bump the IMC (Integrated Memory Controller) voltage and possibly the DRAM voltage. This depends on how several of the hardware components are interacting.
Your VTT is low at 1.05v with 8GB RAM.
The quick test mentioned above should verify this, to get the 8GB RAM stable enough to run bump the VTT to 1.15v.
This may allow the system to run but, may need further adjustments.

Let us know if this helps.

Here is some info on your PSU:
PSU REVIEW DATABASE Ultra Products

I'm not familiar with the manufacturer Andyson, this may need to be looked into if the RAM tests above don't resolve the problem.

Just reviewed your thread.

Up to 5 successful passes on Memtest86+

here's a small snippet from the Motherboards manual: Gigabyte GA-Z68AP-D3 | Manual - Page 37

Extreme Memory Profile (X.M.P.), System Memory Multiplier (SPD), Memory Frequency(Mhz)
The settings under the three items above are synchronous to those under the same items on the Advanced Frequency Settings menu.

Performance Enhance
Allows the system to operate at three different performance levels.
Standard -- Lets the system operate at its basic performance level.
Turbo -- Lets the system operate at its good performance level. (Default)
Extreme -- Lets the system operate at its best performance level.

DRAM Timing Selectable (SPD)
Quick and Expert allows the Channel Interleaving, Rank Interleaving, Channel A Timing Settings, and Channel B Timing Settings items to be configurable. Options are: Auto (default), Quick, Expert.

Profile DDR Voltage
When using a non-XMP memory module or Extreme Memory Profile (X.M.P.) is set to Disabled, this item will display as 1.5V. When Extreme Memory Profile (X.M.P.) is set to Profle1 or Profle2, this item will display the value based on the SPD data on the XMP memory.

So it sounds like the XMP is on by default but I will verify setting after the 7th RAM testing pass

If the system boots with two RAM cards installed or not, leave the RAM frequency at 1333MHz and manually set the first four RAM timings to 9-9-9-24 and the Command Rate to 2T, leave the rest on auto.
Manually set the DRAM voltage to 1.55v and the QPI/VTT voltage to 1.15v.
QPI/VTT is the CPU Integrated Memory Controller (IMC).

Testing with these settings should tell you if the RAM is causing issues, at these settings it should boot.
Sometimes the RAM Auto settings can undervolt the RAM enough to cause it not to boot.

If you end up RMA'ing the RAM at some point you should get RAM that runs at 1.50v, this is the recommended RAM voltage for Intel i5/i7 CPUs.
The RAM you have now is Corsair CMX4GX3M2A1600C9

4GB XMS Memory kit for dual channel systems, 1600MHz, 9-9-9-24, 1.65V

 

[LP81]

Good news so far

Removed 2 sticks, booted up and started running Prime95 around 6 PM on the low setting.

Went out to dinner and forgot it was running and went to bed. Work up this morning to a login screen, BUT it's because Windows decided to deploy updates at 3AM. So sound like it was stable for atleast 9 hours even running Prime95.

I am about to swap out for the other 2 sticks and run it again just to ensure that there's not just 1 bad stick somehow.

I'll follow up after I do that as well as testing the above voltage settings (I'd prefer to run this with the 8GB I bought instead of just 4GB)

 

[LP81]

2nd run with other memmory in Prime95 didn't blow up.

so...
2 sticks = stable
4 sticks = not so much

Before I hit "Save and Exit" in the BIOS

For DRAM Voltage, I do not have the option for 1.55V
Closest are 1.48, 1.50 (normal), 1.56, 1.58 1.60

For QPI/Vtt Voltage, I do not have 1.15V as an option
Closest are 1.10, 1.12, 1.14 (normal), 1.16 & 1.18


For the RAM settings, should I manually set it for both Channel A and Channel B Timing Settings?

Also, the Command Rate only has "2" not "2T"

 

[LP81]

Small note, that in the DRAM settings, once I get to 1.66V as an option, they are all pink text which make me think the mobo isn't built for 1.65V RAM so best bet it sto return it for 1.5V RAM

I think I'll swap out for this: CORSAIR Vengeance 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Desktop Memory Model CMZ8GX3M2X1600C7R

Capacity: 8GB (2 x 4GB)

Speed: DDR3 1600 (PC3 12800)

Cas Latency: 7Timing: 7-8-8-24

Voltage: 1.5V

Multi-channel Kit: Dual Channel Kit

Heat Spreader: Yes

 

[Vir Gnarus]

Btw, you are setting Prime95 to Blend for the Torture Test, yes? Large FFTs setting will use extremely little RAM and focus solely on internal CPU caches.

 

[LP81]

When I did the tests, I did the "Small" test which I was told was RAM heavy. I never did the Large or Blend.

I can run those for you, if you want, but I think that I am going to just swap out the Corsair CMX4GX3M2A1600C9 with Corsair CMZ8GX3M2X1600C7R

 

[Vir Gnarus]

Ok that works. Blend uses somewhat less RAM but it still uses a lot and works the CPU more.