AMD FX 8370 Overheating HELP!

ng9000

Commendable
Sep 2, 2016
10
0
1,510
Help me please!
I bought this pc almost 2 months back and still I haven't been able to use it properly. I'll list my problems as detailed as possible.

My full set up is:
AMD FX8370 WRAITH COOLER
MB GIGABYTE 970-GAMING
RAM 8GB CORSAIR VENGEANCE
HDD INT WD 1TB SATA 7200 RPM
CABINET COOLER MASTER ELITE 311
GRAPHICS ASUS GTX 960 STRIX 4GB
SMPS CORSIAR Vs 650W
DELL MONITOR S2216H 21.5”LED
DVD WRITER LG GH24NS94

On the first day when I switched on my pc and it was running without load, there was no problems whatsoever. On running a 5-6 years old game it still did not show any problems. But on running newer titles like GTA V or COD Advanced Warfare, it shut itself down within 5 minutes. Openhardwaremonitor and Heavyload proved that the CPU was overheating to 72.9 C before it shut itself down. GPU did not show and problems when stressed. I updated the individual drivers but still no change. AMD replaced the full CPU with the cooler and the new one showed the same problems. I went to GIGABYTE and updated the BIOS. Now it peaks to 79 C before shutting down. GIGABYTE recommended purchasing a liquid cooler but the ones they are suggesting require a different Cabinate. My question is using a more powerful cooler is just suppressing the symptoms but what is causing the overheating? This forum is my last hope. Please help me in anyway you can.

 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
Firstly, what is your average room temperature?
Secondly, was this a fully assembled PC or did you buy the components & install it yourself?
Finally, have you done any overclocking at all?

Temps are a little higher than where they should be (under load, at least).

Maybe worth buying aftermarket cooler and apply some high quality Thermal Paste.

GIGABYTE recommend watercooling, fair enough, but an air cooler is easier to install/maintain, this all depends on your know-how, of course.
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510

My avg room temps are quite high, around 30 - 42 C.
It was installed by the shop technicians from where I purchased it. I have reinstalled the heatsink a number of times after that but all with similar results.
No overclocking at all. The situations is like this from the time I first booted up the PC.
I purchased the Wraith Cooler variant knowing my room temps are high and I had hoped it was sufficient for my setup. I tried a liquid cooler which I forgot to mention. It was a Cooler Master SEIDON 120V PLUS. It is a single fan setup. Support Socket Intel LGA 2011-v3/1151. Socket 1150 compatible. Socket FM2+ compatible. This was one of the few which are compatible with my cabinet. It showed the same results with temps a little higher than the Wraith Cooler, if not the same.
Is this whole thing a compatibility issue?
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
I think it is linked to your room temps, have you tried loading up your CPU % + Temps + CPU Fan speed monitor, then tried playing a game & monitoring your readings?

It is pretty normal for an FX chip to reach 30C-40C on its own under full load, add your room temps, and a few degrees for PC Case depending on airflow 35C + 35C = 70C almost already at max temps.

Does the CPU go to full load? Does it then drop (throttle) once you reach the temp limits?
The Wraith Cooler isn't bad, surprised you're getting a similar temp reading with a water cooler though, the advantage of the Wraith Cooler is that as it applies cooling straight downward, it also helps to cool around your CPU (VR lines).

Monitor the performance & temps as they rise, try to find a trigger point (any turbo-boost enabled? try disabling it).

Then come back & with results.


I do not have this FX chip so cannot say for certain what the issue is, most apparent is your room temps slightly above average.
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510

I don't think this is related to my room temps. I took it to the shop a few days back and it showed the same temps in an ac environment with the cabinet panel open. Although it did take more time to reach them. And yes, I monitored my CPU temps with OpenHardwareMonitor while putting it under full load with Heavyduty. If you are not familiar with it, it puts all the 8 cores under 100% load continuously and stresses it. It runs for about 3 mins before overheating and shutting down.I don't think any boost is enabled as I have not tampered with the default settings as this problem started from day one. As for throttling, I'm not really sure but right up to the point when the pc shuts itself down, the games have never stuttered or shown any frame drops.
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760


Fair enough - regarding room temps.

I am not familiar with Heavyduty but I've used Prime95's torture test on many occasions so I am aware of putting all cores under 100% load.

[For a PC that has been built, with the Thermal Paste removed/re-applied several times, with no overclock & differnet coolers, to fail/overheat after a short period of time and not owrked since day 1] <- I would be sending it back and at the very least swapping the CPU out.

Unless someone else has any better ideas?
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510

As I said.. AMD has already replaced the CPU along with the cooler. I'm really stuck here. Can't think of what is malfunctioning.
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
One quick & easy way to reduce temps is to reduce the voltage, what CPU Voltage are you at? Idle/Full Load?

Try to reduce it as much as possible, avoiding an unstable system.

Hope this helps, check you BIOS for Voltage settings control
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510

I guess I have to if nothing else remains. But, I am really looking for a cause for this problem so that I can use my PC normally. :/

 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
Why not try an aftermarket cooler like the CM 212 Evo, it isn't too expensive, provides decent cooling, may solve your problems?
Better than the Wraith Cooler I believe, you will see a temp-difference.
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510


I already used a liquid cooler but it did not help. So I am afraid of wasting money on another one and then seeing no results. I hope you can understand. But can you figure out the cause? Does my motherboard not support my CPU or Graphics Card or something?
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
I completely understand your reasons to not want to shell out anymore without guaranteed success.
The motherboard does support this CPU. Or else it simply wouldn't work.
I could be the motherboard revving up the CPU due to the 125w TDP demand. But the only way to be sure is change mono. Considering you position don't bother. Again I understand.

Check your voltage. The turbo options always use way more voltage than what is really required. And you can really make a difference in temps with the smallest of changes...
Set your clock to lowest speed. Disable turbo boost. Then using bios first or AMD Overdrive second, with reboots after every change, check the voltage needed for your lowest clock speed, next increase the multiplier one notch at the time without changing the voltage. You want to get the highest stable clock at the lowest voltage possible.
As soon as you've reached the max for that lowest voltage, check your temps while running a bemchmark-stability test.

For example my turbo boost for my A8 7650k will be at 3.8ghz with 1.475v.
But i can get it stable at 4.1ghz with 1.4v
The difference in degrees for that 0.075v change with my aftermarket cooler is about 6-8CC depending on load...
I'm afraid that (or speaking directly to Gigabyte about the inner workings of your board with that CPU) are the only options I can think of.
Hope this helps a little bit at least..
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510

Thank you so much for all your help. I have contacted GIGABYTE and they have asked to get my mobo replaced with a new one in case of this one being faulty. I'll update when the process is finished.
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760


Please do... hopefully that is the issue.

I wouldn't say no to a new CPU & now mobo :)
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510


The new motherboard did not help. I installed the Cooler Master Seidon 12v Plus with an extra fan in a push pull style. Temps still reaching 78.9 and shutting done. Only change is that previously the cpu fans and other components kept running and only the display was cut off. Now the whole system shuts down. The motherboard gets very hot too. Please help :/
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
Hello again.
What Thermal Paste are you using?
Is it an old tube or new one?

Those temps are pretty high and if you're getting that right off the bat let me know what Thermal Paste you got, how is it being applied, and (stupid question) did you clean it with Isopropyl Alcohol every single time you remove the Seidon?

The whole system shuts down as a 'failsafe' measure, when temps get too high - let me know about the above.
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510


Hi,
Currently, I am using a syringe of Arctic Silver 5. I got this just a couple of days ago through Amazon and I use the dot method where I put a pea size in the middle and let it spread itself. No, I don't have Iso Alcohol with me but I thoroughly clean it with a dry cloth. As this occurred from the first time I used it, I would guess that is not the issue. Another thing is that, no, temps are not high from the start. after using the dual fan Liq Cooler temps idle between 34 - 19 degrees. it can take the stress test for about 3:10 mins before giving out. Thank you for helping.
 

Decends

Respectable
Jul 3, 2016
685
0
2,060
I know what your problem is. Sadly, you will not like the answer to it. The problem is the Voltage Regulator Modules (VRMs) on the motherboard. That board likely comes with a 4+1 phase VRM power deliverly like most 970FX chipset motherboards for AMD processors do. The 125 watt TDP of the 8370 is to much for the motherboard to handle. Combined with playing a game like GTA V that will suddenly send the CPU to near full load makes to much power want to pass through the VRMs. The VRMs get to hot so the PC shuts it self down before the VRMs fry. You need a motherboard with a 990FX chipset that has a 8+2 Phase VRM design. Only those can sufficiently handle the TDP of the FX 8000 and 9000 series.
 

JCLDJB

Commendable
May 24, 2016
153
0
1,760
ng9000 I am unfamiliar with this motherboard; however what Decends says makes the most sense, overheating VRMs will result in the shut-down you experience.

I had a sneaky suspicion this was the case and did mention this on the my reply dated 2nd September.

As I know for you, buying another motherboard is not an option at the moment, why don't you try a simple solution.

Grab a fan, small or large, and place it a few centimetres above the VRMs to provide direct cooling/airflow to the VRM's heatsink.

This is the section between your CPU socket and the USB ports at the (back) of your motherboard, there is a heatsink and a row of square caps with cylindrical capacitors.
Underneath that heatsink are your VRMs...so try providing some direct cooling with a fan, and see if your benchmark runs for a few minutes longer before shutting down (if you can manage a steady cooling you might be able to keep it stable that way; however there is no absolute guarantee) OR (if your motherboard stays on for a few minutes longer, you know for a fact that that is your problem and can start working towards a solution).

Either saving up for another mother board, knowing your situation I know that's not an option so, try complaining to Gigabyte that their motherboard is not 'as described' or 'not designed correctly for the purpose intended' i.e. compatibility with that CPU. In the UK you could potentially have grounds to pursue the matter further under UK consumer legislation;however I don't know the rules/laws in your country & how much they differ.

Otherwise, go crazy with cooling, and see if a stable system can be met.

And finally, you can try to disable some cores OR turn off any Turbo Boost in the BIOS, to stop the CPU from going max out.


A few options to try...
 

ng9000

Commendable
Sep 2, 2016
10
0
1,510
Thank you guys for all your help. I think now, after such a long time, I have an idea about what the problem is. And yes, as you mentioned, I am not really comfortable in buying a new MB but, if I have to, I will. For that I need to be 100% sure that this is the source of the problems. I'll update as required and inform the solutions. And if anything else comes to mind, please post it. Again, thank you so much.
 

Decends

Respectable
Jul 3, 2016
685
0
2,060


Your 970 chipset board is of unusual circumstances. That board features a odd 7+1 VRM design. 7 for CPU, 1 for North Bridge. But in general 970 chipsets are terrible for FX 8000 series, but this board is 1 exception.