PC Graphics

[lambenttyto]

When I did the OCCT test on the PSU and GPU all the power readings seem fine, including temperatures, but when I do the test on the PSU mode, it turns my computer off completely. I'm guessing I have a bad power supply? I can't RMA because I'm currently living in a third world country, so before I go out and drop a buncha dough, would like to confirm a little more. Anything else I can do?

Here's exactly what I'm running if you wanna know: https://pcpartpicker.com/list/gGsCCb

Thanks

[elstaci]

When running OCCT PSU TEST:

Find a spare 4 pin MOLEX plug inside your computer case connected to the PSU and put a Mutlimeter or Voltmeter to the 12v and Grd pins on the molex. Than run the PSU Test or run a game to see if the 12 vdc reads below 11.4 vdc. If it reads below 11 vdc your PSU is probably weak or bad.

When running the GPU test, what are the PSU voltages showing for 3.3vdc, 5.0vdc, and 12vdc which should be indicated during the test. If not, then probably OCCT is not compatible with your motherboard's sensors completely. Or upload the voltages concerning the PSU from OCCT Folder showing the  3.3/.5.0./ 12.0 vdc.

This website give instructions on checking your PSU with a multimeter:

How to Check the Power Supply With a Multimeter | Chron.com

How to Manually Test a Power Supply With a Multimeter

Here is Youtube checking the PSU through a Molex plug with a multimeter:

How to check voltage on your computer using multimeter - YouTube

Make sure your PSU is clean around the FAN because a overheating PSU would do the same thing.

OCCT - PSU TEST basically runs both the CPU and GPU Tests at the same time putting maximum load on the PSU.

Are there any OCCT Folder with Graphic Data from the PSU Test in "Documents" Folder?

[lambenttyto]

All right, I'll take this advice. I'll get a multimeter when I go into town so I can check the voltage of my PSU. OCCT did gather up some CPU and GPU readings.

[elstaci]

Your PSU voltages are fine. In fact very good if it is accurate.

Run the PSU Test and see it you get the same results before buying a Multimeter. You can get one for about 10 USD.

What were the Temperatures of the GPU and CPU when you ran the GPU Test?

Run the CPU TEST and see what results you get.

If it shuts down during the PSU test, it could indicate that the PSU can't cope under heavy current usages or something is overheating including the PSU.  The OCCT  PSU TEST is basically running both CPU and GPU TESTS at the same time putting Maximum load on the PSU.

I run mine, which is an 8 years old Corsair 850 Watt PSU,  using OCCT without any problems. Except my 3.3VDC is around 3.1VDC which if it gets any lower I would need to replace my PSU.

If your PSU is high quality and in good shape it should also run without shutting down. my opinion only.

[lambenttyto]

Temperatures on the CPU and GPU were just fun. 51 C on the CPU and only 66 C on the GPU. When I run the PSU test, my system shuts down immediately, though. I feel like it's a PSU issue, though when I put my GPU in the second PCIeX16 slot, the computer barely turns on. It turns on, blinks like it's shutting down, and then boots up and everything is fine, so maybe motherboard potentially, I don't know.

Any way to measure PSU temps?

[elstaci]

I don't believe there are any temperature sensors that can be read by software that I am aware of in PSU units. Maybe the newer models may have that feature which includes it own software.

The best way is to make sure the PSU Fan is clean and cleared of any dust and particles that may obstruct air flow through the unit. Also if possible be sure the PSU Fan is running when power is turned on. Not sure if PSU fans are like CPU fans that turn on or off depending on temperature sensors.

As far as I know, The computer shouldn't shut off as soon as the PSU test is initiated. I hate to have you purchase a new PSU and find out it was good and it was some other hardware problem.

As to your GPU in another PCIe slot, you need to read your motherboard's manual to find out what the PCIe slot speed is and if it can be used for GPU cards. It could also be power related by your PSU not giving proper voltages to the Motherboard.

Those links I gave you might help you determine,by yourself, if the PSU is good or not.

Try downloading and installing a free Monitoring software called HWMonitor from here: HWMONITOR | Softwares | CPUID  . Have it running when the computer is under heavy loading. It constantly shows all the Temperatures, Voltages, and Fan speeds of your hardware. See if you see anything unusual with Temperatures, Voltages or Fan speeds.

That may help you determine what may be causing your PC to shutdown when the PSU is stressed. News, Tips, and Advice for Technology Professionals - TechRepublic

Copied from above link of a bad or failing PSU Symptoms:

The symptoms

Just about any intermittent problem can be caused by a faulty power supply. I normally look there first when the symptoms are a little weird. Common power-related symptoms include:

• Any power-on or system startup failures or lockups
• Spontaneous rebooting or intermittent lockups during normal operation
• Intermittent parity check or other memory-type errors
• HDD and fan simultaneously failing to spin (no +12V)
• Overheating due to fan failure
• Small brownouts that cause the system to restart
• Electric shocks that are felt when the case is touched

If you have any of these common PSU power symptoms than I would replace it. If not I would have a wait and see unless you are able to get another PSU to temporarily install and test with OCCT.

Also if your PSU is underpowered for your computer it will also cause similar symptoms. You never mentioned what you have installed for the PSU wattage. Take a look at this website and see if your PSU has the minimum Wattage for your GPU card: PSU REQUIREMENTS - RealHardTechX ​

[lambenttyto]

My PSU is very clean. This is a newly built system less than three weeks old. I'm running a 600 watt EVGA B1 PSU. Also, my PCI slots are PCIEx16. Is that not the speed?

Anyway, I think I might have figured out the problem. I think my card was overheating. And I know you told me to check that, and I did, so I'm a bit confused. Heat only went up to 65-66 on the OCCT one hour stress test. But before that I was using MSI Afterburner and I never saw the temps go up past 60 C. After uninstalling MSI and putting the AMD performance overlay on during gaming, because I heard AMD and MSI readings can conflict, I saw the temps climb up to 70 C pretty fast. I went into the settings to find out the default target temperature is 70 C with it set to a max of 90 C. It shut down shortly after hitting 70 C. Anyway, I decided to lower that target down to 50 C and kept the max up at 90 C and cranked the fan speeds way up. I was gaming last night for about an hour straight at +50 power, which usually crashes my system within minutes. I held at stable temps between 45-50 C and no crashes. What I'm still confused about is why the card would crash after rising just over 70 C? That's not very hot. Although I admit, my eye was off the temps for about 2-3 minutes before crashing, so who knows, maybe it rose far above 70 C, or maybe my AMD performance overlay is lying, but it seems to be in tune with HWMonitor that you told me to download.

I'm going to run some titles that are more demanding this afternoon while keeping an eye on the heat. If I still have problems, I'll let you know, and if not, I'll report back here. Up to this point I appreciate your help.

[elstaci]

AMD Wattman is the best way to control the GPU Fans and Power on the RX GPU Card. Sounds like its a configuration problem. kingfish​ can show you the proper way to configure your RX GPU using AMD Wattman.

Yes it is advisable to uninstall MSI Afterburner because it MAY cause conflict with AMD Wattman, even if it isn't installed or activated.

What you said I tend to agree. GPU cards should be able to reach fairly high temperatures without crashing.

So it seems it was an overheated GPU card causing the problem all the time. Your PSU is strong enough for the GPU card you have installed. Yes, it should work on the other PCIe x 16 slot then.

What threw me off was that your computer would crash with OCCT PSU TEST and not the GPU TEST.  But as mentioned before, OCCT uses both CPU and GPU tests at the same time to put max load on the PSU. 

After you get your GPU card to stay fairly cool and configured properly under AMD Wattman and it doesn't crash under heavy gaming. Just out of curiosity, run OCCT PSU TEST again and see if it runs without crashing. This is useful information for future troubleshooting using OCCT.

Also doesn't HWMonitor give show you the Maximum and Minimum Temperatures on the GPU Card while it was active?

anyway I am relieved that you may have found the reason for your computer crashing under heavy gaming loads.

[elstaci]

Follow kingfish​ image and advice on how to configure your RX to run without crashing during games. You can uninstall HWMonitor and install the one kingfish​ recommends. Both are very similar.

You can activate AMD Wattman through Radeon Settings, by the way.

[kingfish]

Third party controllers conflict with AMD/Wattman controls and settings. All of them. Delete them from your computer.

An old post..but relevant > AMD graphics performance

The Power Limit and Max and 'Target' temperature should be set to manual and the settings adjusted as shown. Don't forget to click "Apply". Do not change any other setting. Leave them at default (automatic).

OR

This setting will engage the fans at 45c (Target) and gradually/automatically increase the fans to keep the temperature at/under the "Max" temperature you have set (80c). (I use 75c as Max now).

Run Hardware Monitor in the background while you are gaming and it will show you the temps and fan speeds. Open Hardware Monitor - Core temp, fan speed and voltages in a free software gadget

[lambenttyto]

Thanks for the post. Elstacy was already directing me to your previous posts on the subject, too. So what about fan speeds? Leave them at default?  Would it be bad to ramp them up?

[kingfish]

Yes it would be bad. Leave them alone. Default setting is automatic...

[lambenttyto]

So I did some more testing today. Kept the target temperature at the lowest bar and I've been playing my most graphically intensive game for about two hours straight today. I can't run this game at ultra or even on high settings at 60 frames, so it's pretty taxing to my GPU and it's running good. I also did another PSU test with OCCT that lasted for about 40 minutes before I turned it off.

I can't remember who it was, but somebody said to uninstall all other monitoring programs because they interfere with AMD Wattman somewhere else on this forum, so I did that and that's when I realized I might have an overheating problem. Setting wattman back to default, the target temperature is 73C with a max of 90C. Ludicrous! I think that's why OCCT originally shut my system down when I tried it before. I was probably idling near 70C on a hot day and when I turned that thing on it shot my temperature up to whatever zero tolerance zone that was shutting my system down.

So it looks like everything is working fine now. I appreciate the help, kingfish and elstacy. Now I just need to rearrange the fans in my case so they aren't pushing a bunch of hot air from the top down toward my GPU, lol.

[lambenttyto]

Well, turns out I'm still having the problem, but it's much less frequent. I think this is coming down to two factors, for one, the GPU isn't spitting out so much hot air, and secondly, I moved it down to the second PCIeX16 slot. I moved it back yesterday and the problem got a lot worse, which tells me it's my NVME SSD drive. I did some research, and there's a lot of people complaining about their 960 EVOs shutting down their systems. I don't find it surprising that my second slot happens to be a lot farther away from my SSD chip. The first slot puts the GPU almost right on top of it. I'm going to move my Windows installation to another drive to see if that clears this up.

[kingfish]

Sounds like a good plan..

[black_zion]

Sounds odd to me. I've a 960 Pro mounted to my Crosshair VI which serves as my Windows drive, and after transferring 100GB straight it never heats up over 54*C or causes problems, but all Samsung NVMe drives contain what they call Dynamic Thermal Guard, which will throttle performance should it exceed its threshold, which is higher than 75*C.

[lambenttyto]

Yeah, I know. I've been looking at the temperatures of mine as well, and it doesn't seem to be heating up all that much. Actually around the same temp as you've got there. Some other people have reported their systems shutting down or restarting because of this little guy though.

[black_zion]

Have you tried WhoCrashed to analyze crash logs?

[lambenttyto]

No, I haven't. Is that a program? It's hard to get crash logs because the event viewer shows a kernel-power 41 task 61 report, which is a sudden loss of power without any discernible leading up to that point.

[black_zion]

https://www.resplendence.com/whocrashed