cancel
Showing results for 
Search instead for 
Did you mean: 

Processors

verscotchi
Adept II

FX 9590 downclocked, Temps aren't looking so good, is it the PSU or could it be something else? (100% util in any game)

  • Issue Description: I've recently made the decision to downclock my FX 9590 due to insufficient power draw from my current 600W EVGA GD unit, and now I'm running my CPU at 100% about 2 min any game whether it be Minecraft, R6 Siege, etc. I want to know if upgrading my PSU will fix my issue, because my thermals don't seem that great either. (15-20c thermal margin under load)
  • Hardware: CPU: FX 9590, CPU Cooler: H100X AIO, MOBO: Asrock 990FX Extreme 9, GPU: GTX 970, Case: Thermaltake S71 Urban, Storage: PNY 120gb SSD, WD 500gb HDD.
  • Any help is appreciated on this topic, been sitting on it for 2 weeks now.
0 Likes
1 Solution

The GTX970 requires a minimum PSU wattage of 500 watts while your FX 9590 TDP is 220 Watts.

The maximum Operating Temperature of your Processor is 57C. So if the temperatures don't go above 57C you are within Operating temperatures.

The Power consumption of the FX 9590 at full load using OCCT is 350 Watts according to this Tech Site review: 2014 Test Setup and System Benchmarks - AMD’s 5 GHz Turbo CPU in Retail: The FX-9590 and ASRock 990F... 

So it is possible under very heavy CPU Loads and GPU Loads it may come near or pass your PSU ability to give proper power. That is where OCCT PSU Test will prove by running it.

I would suggest you restore your FX9590 back to its default settings and then download OCCT.

OCCT is a free diagnostic that checks PSU by stressing it.  This is the only Diagnostic software that checks and stresses the PSU.

Download the OCCT.  Run the CPU Stress test first and see if it crashes or passes after about 45 - 1 hour of testing. During the Testing you can keep an eye on the PSU outputs 3.3/5.0/12 Vdc. and the CPU's load and temperatures and other pertinent info.

If it crashes try to find out what the PSU outputs were. They should be within 5% tolerance of each reference voltages. Also check to see if it crashed due to overheating which might indicate something is wrong with your AIO CPU Cooler.

IF the CPU Stress test passes now run the PSU Stress test. This test basically runs both the CPU and GPU Stress tests at the same time putting maximum stress on the PSU.

While running keep an eye on the PSU's Outputs again and on temperatures and loads. If it crashes immediately and shuts down you need to look at the test results to see why it crashed. Did it crash due to Overheating or PSU voltages being out of tolerances?

If you don't want to use OCCT you can download one of the free Hardware Monitoring software that shows PSU outputs. Run a game and check the Monitoring software for any unusual temperatures or voltages.

View solution in original post

35 Replies

The GTX970 requires a minimum PSU wattage of 500 watts while your FX 9590 TDP is 220 Watts.

The maximum Operating Temperature of your Processor is 57C. So if the temperatures don't go above 57C you are within Operating temperatures.

The Power consumption of the FX 9590 at full load using OCCT is 350 Watts according to this Tech Site review: 2014 Test Setup and System Benchmarks - AMD’s 5 GHz Turbo CPU in Retail: The FX-9590 and ASRock 990F... 

So it is possible under very heavy CPU Loads and GPU Loads it may come near or pass your PSU ability to give proper power. That is where OCCT PSU Test will prove by running it.

I would suggest you restore your FX9590 back to its default settings and then download OCCT.

OCCT is a free diagnostic that checks PSU by stressing it.  This is the only Diagnostic software that checks and stresses the PSU.

Download the OCCT.  Run the CPU Stress test first and see if it crashes or passes after about 45 - 1 hour of testing. During the Testing you can keep an eye on the PSU outputs 3.3/5.0/12 Vdc. and the CPU's load and temperatures and other pertinent info.

If it crashes try to find out what the PSU outputs were. They should be within 5% tolerance of each reference voltages. Also check to see if it crashed due to overheating which might indicate something is wrong with your AIO CPU Cooler.

IF the CPU Stress test passes now run the PSU Stress test. This test basically runs both the CPU and GPU Stress tests at the same time putting maximum stress on the PSU.

While running keep an eye on the PSU's Outputs again and on temperatures and loads. If it crashes immediately and shuts down you need to look at the test results to see why it crashed. Did it crash due to Overheating or PSU voltages being out of tolerances?

If you don't want to use OCCT you can download one of the free Hardware Monitoring software that shows PSU outputs. Run a game and check the Monitoring software for any unusual temperatures or voltages.

First of all, you have been the most helpful out of anyone on any forum I've been on as of late with this issue, but to continue to clarify on what I need to do is to clear my CMOS and then download OCCT to monitor PSU voltage outputs along with CPU loads and temps? CPU stress test then PSU stress test if it passes the first, then report back with more info?

0 Likes

Yes, resetting your BIOS/UEFI back to default is a good idea. If that is how you under-clocked your Processor.  Then run the OCCT  CPU Stress test and see if it passes observing PSU outputs, CPU temps and voltages and loads.

IF the CPU passes then try the PSU test to maximum stress of the PSU. Again while testing observe the PSU Outputs and CPU/GPU Temps and loads and voltages.

Ok, so I passed both tests and my computer seems to running fine now, with decent temps and better performance. But, I'm still having issues with having a game open and another program running, is that the limitations of the CPU or still something possibly related to thermals and PSU?20191015-222538-CPUOCCT.png

Can upload the PSU Stress results?

Was the PSU Outputs normal (With 5% tolerance from reference voltage)?

0 Likes

Let me try to rerun that test did not save those results.

Ok, so my system did crash took it about 5-7 min to do so. I can't figure out to find out the temps or power usage before it crashed and also (foolishly) I decided to bump up my GPU clock speeds in the middle of the test. Man, I am not good at this. So, on top of not knowing my temps or voltage I think I added a possible variable that intervened. I did run the test last night and it went stable for a prolonged amount of time. Is there a way to check those before I crash?

0 Likes

What type of crash (BSOD , computer shutdown by itself, etc)  and any errors pop up?

0 Likes

Screen froze, no input or output. I was in a discord call with my friends after the crash there was no response or anything I said was going through. Other than that the computer continued running at the same fan speed, but again the screen froze and was completely unresponsive.

0 Likes

Found this old Tom's Hardware Thread with the same processor and symptoms that you are having: Can a PSU cause your computer to freeze? | Tom's Hardware Forum 

It is a long thread but you may find it useful.  One User does mention freezing could be due to Power supply issues.

Also from Tom's Hardware PSU 101 explains everything you need to know about PSU and testing them: How We Test Power Supply Units - Tom's Hardware | Tom's Hardware 

Thank you, will read this.

0 Likes

Why don't you open a EVGA Warranty ticket or a EVGA Support ticket and see if your PSU is the problem with your freezes. Your 600 EVGA GD has a 5 year warranty on it.

They might be able to determine if your PSU is capable enough to support the FX 9590 Processor under full load.

NOTE: From what little reviews there are on the EVGA GD PSU, these Models are considered to be Budget PSU (Inexpensive type).

0 Likes

I got this power supply literally like 2 weeks ago, and is still under the Amazon return policy, so that's why returning it and buying another PSU is still an option. Nevertheless, let me see what EVGA has to say.

0 Likes

I used FX9370 4.4Ghz 1.48V core voltage manually with my old system. I turned off turbo and cool'n'quiet and made stable my old system. Install MSI AfterBurner and set Power Limit to -50%. Maybe it won't crash.

0 Likes

I'll be sure to try that if the issue isn't solved, but I'm also concerned that that particular configuration will definitely tank my performances, unless the reason I'm having performance issues is due to stability. Can you confirm that this rig was decent enough? Say run a AAA title and stream a video?

0 Likes

I've written "reduce Power Limit" for OCCT PSU test not for performance. Because if you know that your system is stable, you can improve the situation by purchasing a power supply.

0 Likes

Oh, I see so to test stability I should try to tweak my system to find stability. Turbo and Cool'n'Quiet are extraneous then I'm assuming. One more thing, should I touch my LLC or no?

0 Likes

Sorry. I cannot use English language fully, I cannot write in detail. First of all, I want to share that your PSU is definitely not enough. However, you may have problems even if you purchase a good power supply. Because FX9000 series processors require very high voltages. It also has a very high TDP value. I had to buy a 750W power supply for a FX9000 series processor. And I had to change some settings with BIOS. I turned off Turbo and Cool'n'Quiet features with BIOS for stabilization. I made processor multiplier 44 and gave it 1.48V manual core voltage. And it worked stably. 

As I understand it, you can't do an OCCT PSU test. To make this test, try GPU Power Limit setting to -50% with using MSI AfterBurner software. If this test can be done with these settings, you will use your system with a new and good PSU. Your money won't be wasted. Sorry. I don't explain my article. Because I'm having a hard time. I'm trying to write short and clear. Your problem is already very clear so is PSU.

I hope you can solve your problem in a short time. I want to repeat what I wrote from the beginning. Get rid of the FX9000 series system as soon as possible.

0 Likes

Honestly, I didn't even notice your English impediment, but to continue on what you're saying is to try to find stability through the OCCT PSU test and then get a new power supply? Or that if I pass this test this PSU is good enough? Thank you again, and I understand now that it is a my PSU.

0 Likes

You're welcome. Try to find stability through the OCCT PSU test and then get a new power supply. I think that's the right thing. Because reducing the GPU power limit will result in loss of performance. However, it will provide your PSU to load less. I want to give an example. When I used FX9000 series CPU and performed OCCT test only for CPU, the processor was less heated. But when I tested it for OCCT PSU, I saw a significant increase in processor temperatures. That is, 9000 series systems have a northbridge and the processor temperature is affected by GPU temperature. So a good PSU may not solve the problem so I wanted to mention the possibilities that you will have. I'm writing as detailed as I can.

0 Likes

I also agree with Verscotchi about your English. It is very good. I thought you were an American User. You have a good command of the English language which is beneficial to communicate in a English forum like AMD.

Do mind if I ask what is your native language?  I am also bilingual whereas I speak English, main language, and Spanish, my parents language. but I only understand Spanish and can barely read or write in that language.

0 Likes

Ok guys, we got some good news. After downclocking to 4.2 ghz and 1.35v I established stability for about 43 min. That was a huge step over the 6-8 seconds of stability of the clocks of (4.4 ghz - 4.7 ghz) After the last freeze, I couldn't really pinpoint the issue for the freeze if it was temperature or voltage, but after rebooting to 4.1 ghz at 1.3375 v I think I have found stability. I did try dropping the gpu to 50% Power Limit, but it still freezed under that setting. So, it seems like the issue was with the clocks and 4.1 ghz and 1.3375 seems to be the sweet spot. Now, am I looking just to upgrade my PSU?

0 Likes

I'm from Turkey. 42 multiplier and 1.35V may not work stably. The voltage may be low. You can solve the problem by experiencing with the BIOS. This is just a guess I will write, 42 multiplier can work stably with 1.42V. So you can only find the correct multiplier and the correct voltage. Although you do not use OCCT, the system crashes at Windows startup probably due to low voltage. 

0 Likes

JonnyGURU.com – Home of the PSU GURUs.  is a good website to find out about all the new PSU and how good they are.  If your PSU is new, can you return it for a refund? 

8 years ago when I built my computer, I purchased at Tiger Direct Retail store a Corsair 850 watt PSU. At that time it was one of Corsair Top of the Line PSUs. I am still using it today and working like a charm. Even with a High Quality modern GPU and a FX 8350 Processor with a Asus motherboard that is completely filled with hardware and a USB Hub.

The PSU is probably one of the most important hardware for you to purchase. Try to purchase a PSU of at least 850 watts or more, so that you will have enough power in the future to upgrade your GPU or any other hardware connected to your motherboard.

Good Luck

I have the same opinion. I took a picture of my old PSU. 9370 CPU and 7870 GPU was ran properly by this PSU. But 850W is better.
72572507_752919741820170_2997722984598208512_n.jpg

Ok, am looking into a new power supply now, I think we came to a reasonable enough conclusion until I actually get the new power supply. Thank you for help guys, I will come back whenever I finally get a new one.

Ok, so I have finally got my hands on a EVGA 850 W GQ PSU, all installed and ready to go, until I end with the same problem I had last time, instability due to Vdroop

I’m gonna need some help on this one.

0 Likes

If you have not made any changes with the BIOS, that is, if you are using the system in factory settings, there is probably nothing to worry about. Try disabling AMD Turbo Core Technology with the BIOS. Or adjust the system with manual core voltage. It shouldn't be a problem. I couldn't find any documentation for your motherboard BIOS. If I had this document, I would clearly write how you should try to make changes but I can't.

0 Likes

Found stability at 4.7 ghz and 1.4 v with a 50% LLC, any time I would set my LLC to disabled my voltage would just spike to 1.5v+ and stay there until my computer freezes up.

0 Likes

1.4 core voltage is low. 50% LLC is very high. Try 1.48 vcore and 20% LLC. Try to achieve stabilization with values close to the standard. Also 4.7Ghz is very high. Disable the turbo and use a speed of 4.4 or 4.5.

0 Likes
mstfbsrn980
Grandmaster

Buy new CPU+mobo+DDR4 as soon as possible. You're using the most hot-running processor of all time, and many motherboards with 8PIN CPU power are not working properly with this CPU. 

0 Likes

Unfortunately, I won't have that kind of luxury since I obtained this rig fairly recently and don't plan to spend money on more parts at this time. But, I am aware that I can get way better components. Basically, I'm stuck with this thing and the only "upgrade" option I have is the PSU.

I used a similar system with HD 7870 (has 2x8PIN PCI power port). I had to buy 750W PSU, because 650W was not enough. In addition, the motherboard crashed with turbo CPU frequencies. I turned off turbo and used it stable. I wrote to you what I have done to use this system stably. I hope it helps. Should retire the system as soon as possible I think. The system heats up even when idle, you can't touch Northbridge. So I understand your troubles.

Thank you for the attempt, I can't seem to find these settings you sent is it in my email or somewhere on this website? Thank you again for being helpful.

Found this previous AMD Thread where a User mentions that all FX 9xxx Series NEEDS A MINIMUM OF 54 AMPS on the 12 VOLT RAIL. Your maximum 12 volt is 50 AMPS.What type of power supply unit is recommended for FX-9590?  

From EVGA Website about your PSU: EVGA - Products - EVGA 600 GD, 80+ GOLD 600W, 5 Year Warranty, Power Supply 100-GD-0600-V1 - 100-GD-...