Solved 2 BSODs 2 Days

[reventon]

I decided to do a memtest over night
Both sticks are in. 0 pass 33 fails.
ANOTHER RMA.. Great!

I would hold off on that RMA - continued RAM failures are not co-incidences.

Please wait for Dave76 to confirm that your RAM voltages etc. are correct.

If they are - then I would definitely RMA the CPU this time, not the RAM.

 

[oli356]

Ok thanks, though the problem about warranty and overclocking might come in? To be honest only OC in the first place to see if there's any noticeable difference. Didn't 'feel' any so I just left it... all for show I guess

The dates of my last RMA for RAM:
19/07/2010
and the time before
08/03/2010

Picture attached of todays memtest. The last RAM which I RMA took 2 seconds in memtest to show way over 100 results it was like 9999

 

[kcmartz]

Drivers or hardware is the main fault I would imagine

Correct. I have actually seen VIPRE cause a few BSODs though...

Given that they appeared again after months of absence - I do have to suspect the latest bit of hardware which of course is your sound card.

VIPRE can? What antivirus would you then suggest? My dad is a reseller for VIPRE.

 

[oli356]

Drivers or hardware is the main fault I would imagine

Correct. I have actually seen VIPRE cause a few BSODs though...

Given that they appeared again after months of absence - I do have to suspect the latest bit of hardware which of course is your sound card.

VIPRE can? What antivirus would you then suggest? My dad is a reseller for VIPRE.

I see a lot suggesting MSE in this part of the forum. Also you can have a look at this, its a computer forum link. **********The All New Official OcUK Anti-Virus Program Thread - KEEP ALL AV QUESTIONS IN HERE********** - Overclockers UK Forums

http://www.microsoft.com/security_essentials/

 

[CarlTR6]

On some systems, Vipre seems to do fine; on others, it causes problems. This is true of most 3rd party antivirus programs. Avast seems to run well with Win7 and, of course, so does MSE. There are valid reasons why third party security programs conflict with Win 7. It due to changes in Win 7's own internal security.

 

[oli356]

Your next step is to remove one stick and repeat the test on the remaining stick. Then remove it, put the other stick in and test it. If you get no errors on either stick, put a stick in the second slot and test it. Try running your computer with just one stick and see how it does.

Can't actually see that post, only from email.

Why do I actually have to do this?

 

[CarlTR6]

This will tell you which stick is bad and will test your slots in case a slot is bad. When Memtest shows errors, all you really know is that you have a RAM problem. Further testing narrows it down.

 

[oli356]

Ok fair enough about the mobo slot. Though if it needs RMA again then they go back together of course. I'll do a test later in slot 1.

Though:

1) If i get errors in Slot 1, Stick 1. Do i then test stick 2 in Slot 1
2) If both sticks in Slot 1 get errors. I guess I still test atleast 1 stick in slot 2

 

[oli356]

Forgot about this

My TRUE CPU cooler covers the RAM slots, so can't remove sticks. If I even can find the stock cooler i have no paste to keep applying it a few times :/

 

[CarlTR6]

The answer is yes. I suggest that after the holidays, you get some paste and a stock cooler. You are going need it anyway if you have to RMA the RAM.

 

[Dave76]

I agree with Carl and reventon, your CPU has an integrated memory controller.
If it was running hot before, it could be damaged.

Get some thermal paste and use the stock cooler for the tests, then run the memtest86+ as per the tutorial.

Go to your BIOS and reset to Optimized Defaults.

Change the 'CPU NB VID Control' to manual, it should tell you a voltage, let us know what it is.



In Core Temp, TJ Max 79 Deg C, is the max temp for the CPU where it will start throttling down your CPU to prevent damage.
AMD lists your max temp at 62 Deg C, you shouldn't exceed this temp unless you don't mind replacing it.

 

[oli356]

Ok thanks, will look for my stock cooler later. Will need to buy a lot of paste if I will be reapplying it multiple times /sigh
And the memory tests take hours.. so could be doing tests for somewhat 24 hours

The max temperature is rather low then... considering it was going very hot while playing games before I upgraded. While I'm taking the coolers off I'll have another go at taking the cpu out. But when I lift up the latch it just doesn't seem to move anything. So not sure how I could get it out.

Well in the BIOS it doesn't actually give me a number as shown in picture. Set everything back to normal anyway. Meaning the RAM is also normal so now +0.200v. And the OC has been removed
Thanks

Merry xmas.

 

[Dave76]

Can you list your system specs in the 'My System Specs' in the lower left in any of your posts.
This will let everyone find your specs without searching through the thread.

What CPU cooler do you have?
Have you checked that it is tight?

If you apply not enough thermal paste your CPU will run hot.
If you apply too much thermal paste your CPU will run hot, it's possible that if there is too much paste and the CPU gets hot the paste can run out the sides and cause the CPU to be difficult to remove.

Guide for applying thermal paste

Most AMD CPUs temps I've seen have been very low, even OC'ed, a couple deg C above ambient temp at idle is common.
There are threads in various sites that mention heat issues due to too high CPU voltage, we might need to address this also.

What case do you have?
How many case fans?
Have you cleaned it?
Are there any cables obstructing the air flow?
How many fans on the CPU cooler?
Which direction is the fan blowing?

We need to findout what the 'CPU NB VID Control' voltage is set at, this will likely need to be bumped up by a slight increase to help RAM stability.

Bump it by the smallest amount, +0.025v, it should give you the total voltage reading, post it here.

 

[oli356]

Can you list your system specs in the 'My System Specs' in the lower left in any of your posts.
This will let everyone find your specs without searching through the thread.Done

What CPU cooler do you have? Thermalright TRUE Black 120
Have you checked that it is tight? Checked when installed it’s tight and massive =/

If you apply not enough thermal paste your CPU will run hot.
If you apply too much thermal paste your CPU will run hot, it's possible that if there is too much paste and the CPU gets hot the paste can run out the sides and cause the CPU to be difficult to remove. Will have a look again when I get the paste when i take off the TRUE. Will try to remove the cpu and just sit it back down I guess? Will also look at the guide for thermalpaste.


Most AMD CPUs temps I've seen have been very low, even OC'ed, a couple deg C above ambient temp at idle is common.
There are threads in various sites that mention heat issues due to too high CPU voltage, we might need to address this also. Currently, just idle the cpu is at 26c – room is rather cold though. CPU-Z reporting at 0.960V. I guess because it's only at 803.6MHz currently

What case do you have? Coolmaster storm scout
How many case fans? 4
Have you cleaned it? Yes,
Are there any cables obstructing the air flow? I guess so yes, all the ones by the PSU (mainly like graphics cards one being so big, and needing a selection of cables just to power one HDD...So blocking off front fan
How many fans on the CPU cooler? 2 can’t remember what they do, I think one was intake and one was outtake.
Which direction is the fan blowing? Case fans: Front=intake, Back=outtake Side=Intake, Top=outtake

We need to findout what the 'CPU NB VID Control' voltage is set at, this will likely need to be bumped up by a slight increase to help RAM stability.
Bump it by the smallest amount, +0.025v, it should give you the total voltage reading, post it here. 1.1250V with the small increase

 

[Dave76]

Thanks for filling in the specs.

That's a nice case, same as mine

Try to clear the wires from infront of the front fan. It's important for the air flow.

Here are some pics of my rig.

The fans on the CPU cooler should be in a push/pull configuration.
First one blowing into the cooler and second one blowing out towards the rear exhaust case fan.

Bump the 'CPU NB VID Control' voltage to 1.2v.
Run memtest86+ again.
If you get any errors you can stop the test.
Then you will need to follow the memtest86+ tutorial, swapping out the RAM cards and testing.
After you get more TIM and find your old cooler.
This should find the source of the RAM errors.

Let us know the results.

 

[oli356]

Ok will do the memtest now as long as I can keep both sticks in Nearest PC shop is open tomorrow so can get some 'cheap' thermal paste and then a better one for when I put the TRUE back on. Hate installing that thing!

Its an ok case. First I had a cheap £30 Coolermaster 330. Thought this would be better. Just wish it had cable management holes as mine is a mess! Though I also need to look into if you can buy things so instead of me having like 4 SATA power connectors in the case, I wonder if I can just buy one. If that makes sense

Thanks

What do you recommend to clean TIM off with? Anything special?

Also, not had a BSOD for two days now.. how weird

Edit: Should I also put DDR2 voltage control up, as at default it's 1.8V?

 

[Dave76]

My CM Storm Scout has cable management holes

A modular PSU has cables that can be taken out.

Check the thermal paste site I linked above, it explains how to remove the old paste.
Use Isopropyl Alcohol to clean off the old paste.

You can leave the DRAM voltage at 1.8v for now.

 

[oli356]

Well not side holes. Small one at the top.

Yes it's modular and came with a box of cables, but they aren't single cables. Example, 4 power on one cable, not just a single power. Plus a pain to fit the graphics card cables for the 4890. Not much space
Well running the memtest for around an hour now, left ram as 1.8V.

Will give the thread about paste a read, though just remembered I think I bought some Akaska paste cleaning liquid

 

[oli356]

This is stupid

9hours 42minutes: 5 passes errors 0

Its taking forever!! Normally it takes an hour and a bit per pass

I'll stop it at 7

 

[Dave76]

How did the test go?

Any errors?

Had any BSoDs?

Let us know the results.