Question Power Supply for this particular config ( Frequent PC crashes and auto Restarts)

[Rudranil]

Hi guys, i have the processor, graphics card and the PSU as shown in the attached images. The PSU is 450 WATT, processor is i7 8700, and graphics card is GeForce GT 730. Do you think the PSU is good enough to take the load?

https://photos.app.goo.gl/39gaMBPJTt6vy4ax5

New item by rudranil personal

photos.app.goo.gl

I am using this pc for about 1.5 years, till date it ran smoothly, I often use animation software and Adobe tools. What is happening now, it has started automatically shutting down and restarting all by itself. It happens randomly. Not a window crash, BTW, the cpu shuts down and restarts, sometimes after a few mins, sometimes after an hour.

Here is a recording of it: https://photos.app.goo.gl/Q61BkfQQdXJteGxF8

After going through a lot of websites, I realized this is caused mostly due to faulty PSU. I took it to technicians but the problem didnt happen at their store, they returned it giving some useless testing steps. I have tested the pc after removing all devices including the monitor, all usbs free, and it still crashes. I have tried a different power source, with and without ups, under all conditions I have reproduced the crash.

Do you guys have any suggestions on how to go about troubleshooting this ? I understand software very well, but I am a dud when it comes to hardware.

Any help is appreciated.

 

[maxamillionfeettall]

Well, at the minimum, it supposedly is dc-dc on the secondary side, so cross load shouldn't be a big problem. But, the unit is only rated between 200-240v which is a sign of low quality units. A decent psu will always have a voltage range of 100-240v. So if your voltage in the mains are a bit low or fluctuate a lot, the psu will shut off.

You can try plugging the pc into a different socket. After that try to re-seat the ram, reset the cmos, start in safe mode and ddu the gpu drivers. Does it happen in safe mode as well?

Where in the worl are you based?

 

[Rudranil]

I am from India. I use my computer with a UPS, always. So I hope it takes care of the fluctuations. Either way, I have tried a different socket and seen the crash, with and without UPS. So, that is tested.
Will try the rest of them as you have suggested, then get back to you in the results.

Also, from a long term perspective, do you think I should replace my PSU with a better one?

 

[Rudranil]

Did the following tests today.

1. Reset the cmos.
2. Cleaned the ram pins.
3. Checked the voltage readings showing up in the bios. (I have a msi Z370 Motherboard) The bios voltages for cpu look good. I guess the PSU is working fine.
4. Pulled out a few connections and fit them in again.

Didn't help. The system fails much faster now, just when the windows login screen is about to appear.

Next, I opened up the heat sink (the fan over the processor) of the processor, seems that the thermal paste has dried up, I can barely make a smudge without putting pressure. The cpu Idle state temperature in the bios averages around 49 deg Celsius. Goes up to 52. This is not with windows running, just inside the bios. Is that causing an issue?

Moving forward with the story. I had pulled out the SSD connection before removing the heat sync. I placed the Hest sink back, I connected the 2 SSDs to the motherboard and PSU again.

Now, I started the system, it booted and goes beyond the login screen of Windows. I felt hopeful. I immediately downloaded the open hardware monitor softwars and started noting down the cpu temperature. In Idle state the temperature hovers around 40-45 deg Celsius. I tried opening some heavy duty Adobe tools at once, this is where the cpu temperature momentarily shot up to 87-90 deg celsius, but then came back to 50 60 degrees, remaining at 75 degrees for a couple of minutes.

The system is running fine for over an hour now. I have kept the open hardware monitor running and it is logging the temperature in a csv file.

What do you think? It could be a loose SSD wire, or it could be the cpu heating. Or may be something else. I am going to get a thermal paste from store tomorrow either way.

 

[maxamillionfeettall]

I tried opening some heavy duty Adobe tools at once, this is where the cpu temperature momentarily shot up to 87-90 deg celsius, but then came back to 50 60 degrees, remaining at 75 degrees for a couple of minutes.

Those are far too high temps, usually when it gets that high the cpu will throttle and slow down to cool down. The cpu shouldn't exceed 80C at any time, 75 is also too high but if you're using the stock cooler that's about normal. The pc usually won't shutoff , but if the cpu stays hot despite throttling down, it'll shut off. Check the thermals again after you apply new grease, it should drop the temp ~5-10C hopefully.

If the sata connection was loose on the boot drive, that certainly could cause problems, though random restarts isn't a normal response, usually it won't be detected or won't load at all.

I'm not too concerned with the psu at this point, especially since you have a ups and a low power gpu. If you were to ever upgrade the gpu, I'd get a new psu as well.

 

[Rudranil]

Those are far too high temps, usually when it gets that high the cpu will throttle and slow down to cool down. The cpu shouldn't exceed 80C at any time, 75 is also too high but if you're using the stock cooler that's about normal. The pc usually won't shutoff , but if the cpu stays hot despite throttling down, it'll shut off. Check the thermals again after you apply new grease, it should drop the temp ~5-10C hopefully.

If the sata connection was loose on the boot drive, that certainly could cause problems, though random restarts isn't a normal response, usually it won't be detected or won't load at all.

I'm not too concerned with the psu at this point, especially since you have a ups and a low power gpu. If you were to ever upgrade the gpu, I'd get a new psu as well.

I will try greasing tomorrow. If the temperature doesn't go down. What would be the next step? BTW there is something called OC Genie on my motherboard. It is off right now. Is it useful in this case to turn on?

 

[maxamillionfeettall]

I will try greasing tomorrow. If the temperature doesn't go down. What would be the next step? BTW there is something called OC Genie on my motherboard. It is off right now. Is it useful in this case to turn on?

Keep it off, your temps are far too high to bother overclocking(OC genie is a simple overclocking function). Overclocking will also increase power usage so I would avoid that.

 

[Rudranil]

The temperature surely has seen some improvement after the thermal paste application. For the time being, I used a cheap silicon paste. Someone has asked me to get an expensive paste and be safe for a few years.

The idle temperatures go to about 38-39 deg. The moderate use temperature hits about 47-50 deg. I only have diablo 3 on my computer. Ran the game at CPU temperature of 50 deg. Only when I fire up 3-4 adobe applications within 2-3 seconds, I see the temperature hit around 85, but comes back to 70 in about 5 seconds and then to less than 50 in another 15-20 seconds.

https://photos.app.goo.gl/XhhJnvcKdzx5QE7S6

The image shows the cpu heat chart. Those peaks are when I fired several adobe apps.

The system has been running for about 6-7 hours without a crash. I have done a windows shut down 3-4 times in these 6-7 hours and the shutdowns have been clean.

I have kept the CPU temperature logging ON, so if I see another crash I can refer the logs. Need to check with the adobe guys if the temperature I see is expected. The first time I had seen this crash was when I was updating photoshop. Not sure if all this is linked. What else? Any other suggestions?

(Edit) About 2 weeks back, I had updated 2 Nvidia settings edge enhancement and noise reduction to "Use the NVIDIA setting" (See the image) and pushed the slider to 100%.

https://photos.app.goo.gl/AjNT9UgQmbGSt62C7

I changed it to default a few hours back. Could this be causing the CPU to get worked up?

 

[Rudranil]

This issue is back after about 2 weeks of running absolutely without any problem. The PC just switches off and restarts at random times and sometimes frequently while booting.

Tested even with unplugged hard disks and ram monitor, the PC is just on cpu, motherboard, psu and graphics card. I am able to reproduce this issue.

Memtest runs well.

I believe it is not a heating problem because the sensors read temperatures at 40 deg on idle and even at that moment the pc crashes. Also, I have used a good thermal paste.

Now, what I found interesting was that there is one way I am able to replicate the random shutdown-restarts with 100% success. The Intel processor diagnostic tool. As soon as the prime number test starts the system shutdown almost instantly and restarts again.

Is this sign of a faulty cpu? Or is it some other faulty part causing the cpu to misbehave?

[edit] : The prime number test passes now, and the system shutdown-restart happens on cpu load. During the prime number, floating point and math tests, the cpu temperature rises to 99 degrees, and just after that the cpu load test starts and then the crash.

It certainly seems like overheating issue, but I have seen the crashes happen even during a relatively less stressed cpu, like the HWinfo reading of temperature is at 45 or 55. I am confused with all these observations.

 

[Rudranil]

Just to help others, I have finally got this problem fixed. The motherboard and the Power Supply both were damaged. Luckily, both were in the warranty period, so I got it fixed for no extra cost. So, that was the complete issue.

Regarding the CPU heating, I got the cpu back and it still kept going close to 100 degree if I ran heavy software like premiere pro, and after effects. I was worried that something will break down again. So, I started looking into the BIOS again. Turns out the CPU heating had to do with the "Intel Turbo Boost Technology" which by default was switched ON. I switched it off, and now with premiere pro, android studio, and 2 more adobe software, the temperature doesn't cross 65.

So, if someone is seeing a CPU heating issue, this could be one of the reasons. Make sure to check it as well.

 

[maxamillionfeettall]

Turns out the CPU heating had to do with the "Intel Turbo Boost Technology" which by default was switched ON. I switched it off,

Turbo boost is what makes intel cpu's boost to higher clock speeds. Having it off means you can only run at the base clock speeds. In this case, you have a base clock of 3.2Ghz, so that's the max you'll ever run with it off. Turbo boost allows the cpu to boost up to 4.6Ghz as long as the cpu is cooled properly. I'm assuming you're just using the stock cooler, which is utter garbage. If you want to get the turbo boost back, you'll need a aftermarket cooler.