FX 9590 overheating on idle

wozzal

Commendable
Dec 13, 2016
10
0
1,510
Hello

I have a serious problem with my FX-9590 overheating almost as soon as I start up. I have never overclocked but have had to underclock (to 3.8-4ghz) just to stop it switching off in idle.

I have the H100i to cool it and have tied multiple times to reapply the thermal past and make sure it is connected properly.
My motherboard is the asus crosshair v formula z
My case is the obsidian 650 and has one large intake fan at the front and one exhaust fan at the back plus the two fans on the h100i

I have been able to use it underclocked for some time but now it has gotten worse and I can just about login before it switches its self off. I use AMD Onedrive to do this - I have tried in the bios but have run into problems as I am not an experience over/under-clocker

I am thinking of replacing the H100i with a KRAKEN X61 280 but my thinking is that it is not a problem with the cooler but the processor, as it is overheating so quickly.

I know that the 9590 runs very hot but this seems too much. I think I may have to give up and get the 8350 but can't really afford to if I don't have to.

I brought the pc second hand about 2 years ago so cannot return any of the parts. It was working fine until about 6 months ago. Not too long after I upgraded to windows 10 but I am not sure if that has anything to do with it.

I am pretty desperate now as I need my pc for work. Any advice will be very appreciated.

Thanks
Adam
 
Very few motherboards officially support the 9590. Lots are socket-compatible, but a 220W CPU requires enhanced power delivery/cooling/protection vs simply matching the socket.

The Formula Z supports it, and the H100i would be one of the recommended coolers. I assume it's definitely a Formula Z and not a "Formula" as that isn't sufficient.

You say it's overheating, but haven't confirmed what temps you're seeing? What are you monitoring with?

Can you post your full specs?

What are/were you actually experiencing?

I know it switches itself off now, but historically, what was it doing before it got "worse"? If it's been lockups/crashes, it may not be temp related - it may actually be the PSU, RAM, HDD etc.
 
here are my specs

Motherboard: asus crosshair v formula z
CPU: AMD FX 9590
GPU: Gigabyte ATI Radeon 7950 HD 800MHz 3GB PCI-E HDMI Windforce 3X OC
RAM: 2 x 8GB corsair vengeance pro series 2133MHz (set at
PSU: Aerocool Strike-X 1100W Modular 80+ Gold
Cooler: H100i
SSD (windows): Samsung SSD Pro 256Gb
HDD: 3TB WD Black

About 6 months ago my pc started freezing or turning off when I put it under pressure (played games, too many apps) the temps would be aroung 70c. After some time I managed to get this to stop by underclocking it using AMD Onedrive (only a little this time) and maged to run prime95 for 1 hour and not reach 70 (high 60s). Since then it has got progressively worse with me under clocking it more and more and it shutting off quicker while idle.

At the moment when I read the temps in the bios the temp is around 48 then rises to 50+ and when it reaches 60 it shuts off (in the bios)
 


So you're saying not every AM3+ socket motherboard supports the 9590 ? what does official mean ? how about the asus aURa motherboard , or that new gigabyte motherboard ? just wanted to know
 
BIOS temps are not always correct - especially true for AMD processors.

That being said, the max safe temp for a 9590 is in the low 60'C range (62 IRRC), so hitting 70'C/high 60's is absolutely problem territory. What's the ambient temp in the room you have this in? That'll impact the cooling ability & CPU temps.

Honestly, this is a little out of my area of expertise, so I'll see if I can get some assistance in here.

As an FYI, your PSU is pretty poor too - especially to be running a 9590.
 


Ok, don't want to get into hijacking the thread territory but a quick answer;
the 9590 is a 220W CPU, with the vast majority of AM3/+ CPU's being 125W or less. Most of the power delivery and cooling on AM3 motherboards are geared towards those 'mainstream' 125W or less CPUs - with very few being capable to adequately power a 220W CPU, nor keep itself cool enough to run stable.

I don't know which boards you're asking about specifically (I'd need full model numbers), but check the "CPU Support" page of the manufacturer's website for those boards specifically. Usually there will be a statement "supports 125W CPUs only", or shows the 9590 in a support list with some comments beside like "requires additional cooling" etc.
 
No there are only about a dozen AM3+ boards that officially support the FX9590 and FX9370. Officially meaning that the boards can actually support it up to full clock speed.


@ wozzel, don't spend more money trying to get the FX9590 to work. Just get rid of it. The 9590 is a pretty terrible cpu and it's near impossible to keep stable. There is a good chance that you have burned out the motherboard's vrms if it used to work, but now doesn't. Either way, sell the 9590 or trade it for a FX8320 or FX8320E and hopefully your problem will disappear. You probably want to look into getting a better psu.
 
I was thinking that I would have to just get rid of it. Just wondering why do you suggest the FX8320 not the FX8350?

Also what can I do if the vrms are burned?

And finally why do you not rate the PSU? I though 1100w would be more than enough?
 
It's not all about the wattage. 1100W is more than enough, but the PSU itself is pretty poor quality. I can;'t remember if that unit is lacking adequate protection, or is fairly unstable as far as power delivery/ripple etc, but it's not very good.

If the VRM's are damaged, the board may be a write-off too. It may just be that they're not capable of running the 220W chip, but a lower power chip wouldn't see the impact of (relatively) minimal damage. The 9590 is tough to be stable anyway, with no headroom whatsoever for an underperforming component (like VRMs).

IIRC, the 8350 is just an OC'd 8320.....and the 9590 is just a factory OC from that (in a simplistic sense).
 


Unfortunately, other than testing with a know working, stable CPU, it's near impossible to tell (unless they're majorly damaged then there should be visual clues)

 
Just wondering,do you have Asus ai suite running? If this is somehow set to overclock the cpu might that give issues.Get rid of it if you have it.
Did you try a clear cmos? Try this,it will set the cpu to normal speed so it might quickly get too hot again though. Especially if having used Asus ai suite do this.


I'm wondering since the pc did fine for 1.5 years,only after that it suddenly gives problems. Might just be that the others are right,but the above are easy things to check/do,maybe they help,maybe they don't,but imo worth the effort.
 
Hi, thanks for you help so far guys.

I purchased a new FX8350 (as it was the same price as the 8320). I can now get my PC to start and work without underclocking BUT it is still running very hot. On idle is is running at 60c. see - http://i1376.photobucket.com/albums/ah6/Wozzal/HWmonitor_zpssyfw6wn5.png

I assume this means either the cooler or as you suggested the VRM's.

Do you think I should try the stock cooler that came with the FX8350 to test this or is it not worth it?

I am just wondering what you guys think I should do. I don't really want to buy a new motherboard and Intel processor now but need my pc.

Thanks
 


I'd say to just use your 8350, but alas, I do not know for sure

 
Give the stock cooler a try. It will blow air down on the vrms. Disable turbo and run some tests monitoring clock speed to see if it is throttling. It should not be throttling at all. If it is, your motherboard needs to be replaced. I assume your case has adequate airflow?
 
OK. Thanks for the advice.

I have managed to get it down. I have not tried the stock cooler yet but I will give it a go to see if the H100i is working properly.

I have disabled turbo core. I have also disabled SVM and set cool n quite. Is this a good idea?

I am pretty sure it is not throttling. cool n quite is disabled and the clock speed on every core is sitting at 4013mHz constantly. Is this correct?

Thanks
 
I will say that on idle the temp will sit around 25 but then might suddenly go up to 35 and then move down again quickly. Not sure why this is. <mod edit> it not be at a relatively constant temp?
 
@mjslakeridge - cable management is very good. I might give the fan idea a go.

@CTurbo - Using CUPID HWMonitor - when i am doing nothing and when i test with prime95 the clock speed for all cores stays at 4013/4014MHz only occasionally dropping to 3952MHz.

I used prime 95 for 4 mins. The temps almost immediately shot up 70 after the 4 mins they were on 88 and I stopped it. This is too hot isn't it?

BTW I am taking the temp reading on HWMonitor from the Package temperature under the CPU not the CPU temperature under the motherboard. Is this correct? The CPU temp on motherboard was ~10 lower throughout. Normally it is ~10 higher.
 
Some folks here on Tom's have suggested that for AMD cpus, the only software that gives accurate readings is AMD Overdrive. (Overdrive gives Thermal Margin, which is the number of degrees you are away from the max temp for that cpu, so a larger thermal margin is better).
 
OK. I ran the prime test again. On HWMonitor similar results. On one drive it went to 0 very quickly. Surpassed this and went up to -20 then the pc shut off. This was after 5 mins using prime95 on Blend test.

The clock speed on onedrive was ~4GHz for the whole time.
 
OK that is good news about the motherboard. May I ask why you think that? Is it because of the clock speeds?

The temperature does return to the 20-30 range relatively quickly but I agree that the H100i might be the problem.

I suppose I need to test the stock cooler that came with the FX8350 and see how that works. From reading I can see that the stock cooler is not that great but should it be able to do a job better than what I am getting at the moment?

Thanks