Major overheating issues with 990FX UD3 VER 1.1

[cliveneo]

Hey guys,

 I need some help with my set up. I have 10 years of PC building experience bout half that in overclocking. Here is the issue. I have set the board up and got my overclock stable. Only way to hit stability with my FX 8320 chip @ 4.5, was to set the voltage @ 1.5 and LLC (load line control) at high. AMD recommends for best OC results a voltage of 1.55 , however others would disagree but to each his own. With LLC enable and CPU voltage set to 1.5 the actual voltage never goes over 1.526 which is not bad. Ran Prime for 9 hours. 20 runs of intel burn at maximum OCCT for 12 hours. Max load temps 55C which is awesome in my book. Max  gaming temps were around 35-40c. Idle temps were iffy around 19c which is lower than my room ambient temp (Approx 73f)

All ran perfect for a month or more. last nite, after a hour of gaming Sniper Elite Zombie Army, I noticed the temps were way too high for me around 55c. So I loaded a synthetic OCCT temps shot straight to 80c at which I abruptly discontinued the test and began to scratch my head. I reset the PC and went into the BIOs checking for any changes from my original settings. Nothing the only change is the CPU temp in the BIOs is 40c from its usual 30c. That is when I stopped scratching and began to get pissed thinking I have defective hardware. I go back into windows and say" NAH cant be that hot, its got to be a sensor issue."  I load OCCT again and temps hit 82C on the CPU when the PC froze. Hard reset the PC went into the BIOs set all to default, restart and  ran Prime for 15 minutes load temps where still higher than they used to be,  around 50c. But all ran perfect no locks or freezes. 

I have a Corsair H100i in push pull with 4 corsair SP120 static 2500RPM fans in a Cooler master HAF X case. So now I am thinking the H1000i is failing, maybe the water flow is obstructed.  I check the fan speeds and pump speeds everything is normal. Pump is 2300 all 4 RAD fan  RPMS are 2500. All set to max. I stop Prime reset and reapply my OC, boot and run Prime, again hits 83c then the PC shuts off this time. Mind you for a month the CPU was MAX Load 55C running intel burn on MAX 30 to 40c running games. Nothing changed, all my settings as far as POWER OPTIONS in the BIOs and Boost are all disabled. Every single option is set the same as it was for a month of COOL operation. Right before the PC  shut down I noticed one issue. the TMPIN 2 temps for the motherboard were almost 100c. That was NEVER EVER that high, most it ever hit was 79 to 80c no more than that .

So now not only is my CPU reading ridiculous temps but so is my motherboard. I am so clueless. Any issue, PC related, I ever had was miniscule compared to this. Question. Will a OVERHEATING north-bridge cause the CPU to overheat? How can both of these be overheating at the same time?  Only other side issue u had since I built the OC was the -12vlt rail reads -16+ volts under load. And the XFX pro 850 WATT modular PSU I have, has always produced a "squeal or wine" since day one even with the PC powered off. Can the PSU have been defective since day 1 and finally its giving away? HELP PLEASE.  Here area my full specs.

AMD  FX 8320 @ 4.5 1.526vlts H100I Push/Pull 4 120mm
Gigabyte 990FX UD3 Ver 1.1
G.Skill Ripjaw DDR3 1600 @ stock
Sapphire 7970 GHZ Edition W/ boost
XFX Fatility1 Pro
XFX Pro Modular PSU 850watt
Cooler Master HAF X 4 200mm red LED fans

  

[cliveneo]

Also forgot to mention I can actively play games without any issues, for hours. Played CPU intensive games like Serious Sam 2-3, Tomb raider 2013 temps around 45-50 which is 5 to 7c higher. Played Sniper Elite V2 Nazi Army for more than 2 hours 45c. Temps are mostly around 45c. never went over that. When I boot into windows temps start at 53c and go down to normal idle 21c. and stay there. Motherboard is normal around 65c load. Man I am stumped. If i can game for hours, then I know the CPU is not dying. It just all of the sudden load temps are KILLER high. Ambient room temps are also the same. I am really starting to point my finger at the motherboard as it is my weakest link and the darn NB runs hot already so when it hit 100c probably heated the whole area around it. Soon as I load a synthetic bench, that MOBO tmpin 2 temp skyrockets first then the CPU temp follows it. 

[Vezina]

The LLC may be to high and the VRM get heat up.
You have a lot of voltage going up and down letting the room so high.Helps overcliocking ,but stresses the VRM

Use medium for LLC  and see what happens.
Add a little to the VID to compensate if instability arrives so the lowest voltage gets a little higher.

Anyways this should be posted under overclocking category , where most probable will remain buried 

Another trick would be to use the Turbo Boost and add there the highest clock leaving for all cores a lower target.In this way the CPU will up only the cores needed that can do it with complete stability.
There is no need to use high clocks on all cores when gaming ,as most games use 4 cores maximum ,except Crysis 3 that seems to be "knowing" 6 cores.
Turbo core will boost only a part of the cores and this should be enough.

[cliveneo]

Thx for the fast reply. I got u venzina however everything worked fine with those settings for a month. The only stress that heated the cpu n vrms was when i ran synthetic benchs. Running games the cpu barely hit 40c. Why all the sudden now is both NB n cpu temps skyrocketing? Even with no OC temps r now literally 10c hotter.

 

[Vezina]

Has the room temperature increased ,is there any dust in the fans or case air intakes ?!
Just clean the cooling stuff and see what happens.

[cliveneo]

I have a huge issue. Even @ stock volts stock speeds temps r through the roof...

TMPIN1 60C
TMPIN2 78C
FX8320 57C

Max load Prime95  

that is with a Corsair H100i Push/pull  4 corsair static SP 120 fans MAXED RPMS NOOOOO WAY. at stock. these are the temps I had when i was overclocked. NOTHING HAS CHANGED room temp is 72f.

Dude my temps were twice as low at stock you can fry an egg on my motherboard now. I even cleaned the Radiator with AIR applied arctic silver to the NB reapplied my pump with arctic silver. Right now I am running prime for 12 hours as I want to rule the CPU out. This is deff not a overclocking issue. Something is going on with this motherboard.

The CPU is the only item i DO NOT have a warranty on. Someone reassure me that my CPU is not going bad. CPUs generally cool down upon defect as the cores start failing first right? I am thinking the VRM or NB is just getting to hot n taking everything with. Or can it be the H100i? everything appears to be running normal with it pump speeds are normal all the fans are normal. Can the motherboard be going out? when components like cpus mobo r dying the cause side effects bsods hard locking resetting. Im getting non of that at stock other than the abnormally high temps. All my PSU reading are normal at stock to voltage fluctuations. How the heck would i know if my h100i is failing? the pump reads 2300 rpms which is where it should be. I have a PII x6 1090t stock amd cooler I can try to see if the stock temps r cooler. Can a failing cpu start to overhead even when good cooling? primes been going for 2 hours so far no issues   

[Vezina]

The CPU temps are fine in load at stock.
The CPU can fail if kept overvolted ,which you did though.
Signs of a CPU going bad are throttling when the load is high or cores going out.

The CPU VRM should take 100 Celsius just fine and the NB 70 Celsius as well.
The CPU is ok at under 70 Celsius.

Overclocking can cause damage to components though and  manufacturers always advertise this.
Just RMA the suspects and see what happens.

If you think the motherboard is a  problem RMA it , but if the back of the VRM area is brownish don t expect a replacement

If the CPU has no warranty you have a problem indeed.

[cliveneo]

yea been running prime for three hours no failure on anything. yea the temps are within reason under threshold but were way way way lower only a week ago. I Know OCing can kill things and I do agree cores start going out throttling etc. I am experiencing none of that. Prime is running on blend which just about burns mobo/mem/cpu. I dont think if I had issues with the Board or the CPU id know it. with the above posted identifiers. I thonk ima RMA this cooler first then go from there. Just not  sure that if a  CPU is failing its gonna run hotter run 25c hotter. As for the VRM area no brown whatsoever. Putting the arctic on the NB cooled things down a bit for the tmpin2. Ill try the stock amd cooler and post up later

[cliveneo]

ran the system with the stock amd heat sink fan for a 1090t. still produced similar results a the h100i. max temps in prime on tmpin 1 67c max package temp 62c. that is crazy that would mean test like OCCT and intel burn r out of the question and would be 10c hotter. there is deff something up here. i also completed 6 hours of prime on stock under the blend. was gonna leave it run while i was working but did want n e thing to happen so i chose not to.

a stock cooler produces the same cooling as a 125 closed loop water system in push/pull? Is it the h100i or the mother board? we know its not the cpu. bu i gotta say i doubt that if prime ca run for 6 hours n the mobo was failing in theory errors or system lockup should be a result. Just want to know what is the culprit as i dont really feel like RMA multiple things at once. Any suggestions?

[cliveneo]

anyone? even in the bios at stock it sits at 40c 45 to 50c when i apply my OC. that insane its gotta be the mobo? CPU is barely used during bios operations

[timmib]

This is absolutely a shot in the dark ...

As I understand it, Win8 has access to and can interact with the system bios.  I do know that with xp pro, windows can take up to 36 hours to go thru its internal routines before all the major changes of a build will be stable.  A couple of boots along the way help some.  So, I'm wondering that by making changes in rapid succession win8 is not getting an opportunity to react.  Your system was fine for a couple of months and then you noticed the temperature spikes.  You have worked at this and stressed the sytem with the diagnostics.  Given the built in fault correction of ms systems, especially one which is meant to be the simplest, etc. could it be that the system needs to run with no changes for at least a day or two.  What particularly has my attention is that win8 probably has the capacity to change bios settings as well as its own machine control parameters to ms' definitions of a proper setup.  For all I know, win8 may be death on oc-ing.

Just some thoughts.

[Vezina]

Just lower the CPU voltage until it BSOD-s and see what temps you get with the lowest value .
See if it s voltage related.BIOS bugs can be triggered.

Maybe something has happened in the mobo VRM department.

But 65 Celsius is not a lot in full load for a top AMD CPU.

RMA the board and the CPU cooler and see what they discover.