[SOLVED] 1 week old PC shuts down randomly or doesn't boot at all.

[Swollengoat555]

Hello. I just recently built a fully brand new PC 1 week ago and have been having some very strange issues. My specs currently

Asus ROG B550-A mobo
XFX Rx580 pro GPU
64GB DDR4 RAM 2666HZ
AMD Ryzen 3900 CPU
EvGA 750 platinum/Cougar GX 1050 gold PsU

If any specific details are needed apart from what's given i will try and provide.

So first day build went great. Booted up just fine and installed all my drivers and OS. Did free 10 pro till i bought a key later. Upon the first few hours of it being up and running I tested 1 game that wasn't even stressful. Within 5 minutes ai got BSOD with a failed memory management error. I essentially troubleshot that 1 of my 4 sticks of ram died even though all 4 were registering in system info before. I replaced 2 of the sticks and bough a new pair of Hyper X DDR4 16gbx2 and replaced them. Worked just fine for a few days no issues at all. This past Wednesday i got home from work and played the PC for 6 hours straight without a single issue. Came back from some errands and it was completely shut down. When i tried to turn it on again it would not stay on! It kept cycling on then off. I unplugged and removed ram again though i didn't think that would have been the issue considering it was just working and I didnt get a BSOD. After not being able to get it started again we concluded that maybe the PSU already died. (Cougar gx 1050 gold). Next day it started up and ran for 4 hours as if nothing happened! I ordered an EVGA that came in yesterday 750 plat just to be safe if it was still the PSU and unplugged everything and connected the new PSU. No luck at all. Even with unplugging and trying to see if it was RAM or thr GPU but it still wouldn't post. I gave up and went to bed. This AM(saturday) it booted up just fine!! Like wewehhhhaat. Then after an hour it powered down on it's own again and continued to do this weird cycle of not booting or booting and powering down on its own. At this point all i can think of would be the CPU or Mobo. But the CPU was not overheating and also running just fine. Could it be the mobo and also the reason the ram stick got fried day 1? At this point I'm at my wits end.

 

[rgd1101]

test with another gpu?

 

[Turtle Rig]

Ok thanks for the detailed information my friend. That is a sexy Rig you got I see no faults or cheap components. Ok so as of now to make sure you have a 32GB Kit installed correct? The PC shuts down to prevent it from dying because of overheating. Each CPU has a TJ Max which is the max temp it will work with then once it reaches that max temp the computer shuts down so the CPU doesn't fry. I think you have a contact issue with your cooler which I would like to know which it is; If the PC is shutting off with the new EVGA solid CPU then its not the PSU. I need you to do 2 things. Take out the battery on the motherboard for 10 minutes then put it back in then go to BIOS and choose set to optimized defaults and then you boot up. Then I want you to use HWmonitor or the app you like to check the CPU temp. Please do this and tell me the temp in C. Seems like the cooler is not properly installed or you didn't put right ammount of thermal paste or put too much and also perhaps you didnt fully tighten the screws so tighten those but not too much to chip the motherboard. So after you do these two things lets see what happens then if it shuts down again and the CPU temp is not much we can take further troubleshooting actions. Well fix you up have no worries.🤷‍♀️👽💯🎗👶

 

[Swollengoat555]

test with another gpu?

I have not tested it with a different GPU but considering it ran just fine for multiple days and ran in my old PC I don't see why it would be an issue. And the power cycling was happening even without the GPU even connected and just using the onboard video. I have another GPU I can test but I have a weird feeling that isn't the issue.

 

[Swollengoat555]

Ok thanks for the detailed information my friend. That is a sexy Rig you got I see no faults or cheap components. Ok so as of now to make sure you have a 32GB Kit installed correct? The PC shuts down to prevent it from dying because of overheating. Each CPU has a TJ Max which is the max temp it will work with then once it reaches that max temp the computer shuts down so the CPU doesn't fry. I think you have a contact issue with your cooler which I would like to know which it is; If the PC is shutting off with the new EVGA solid CPU then its not the PSU. I need you to do 2 things. Take out the battery on the motherboard for 10 minutes then put it back in then go to BIOS and choose set to optimized defaults and then you boot up. Then I want you to use HWmonitor or the app you like to check the CPU temp. Please do this and tell me the temp in C. Seems like the cooler is not properly installed or you didn't put right ammount of thermal paste or put too much and also perhaps you didnt fully tighten the screws so tighten those but not too much to chip the motherboard. So after you do these two things lets see what happens then if it shuts down again and the CPU temp is not much we can take further troubleshooting actions. Well fix you up have no worries.🤷‍♀️👽💯🎗👶

The Ryzen 9 already came with thermal paste already on the top of the chip, and it came with the wraith cooler. I have the armour crate installed and have been monitoring the CPU and GPU temp doing different things. The CPU temp shows that it never goes above 45 Celcius, and ranges from 28-45. But if the CPU was overheating would that really cause it to not want to cycle on even after sitting? because it ran for 6 hours just fine on Wednesday and then got it's first shutdown while I was at the store. Lastly I don't think I understand what you mean by a 32GB kit installed correct. Are you talking about the RAM?

 

[Swollengoat555]

Even after removing CMOS and also removing the GPU and trying to use Onboard video to see if it would even boot and stay up and it still just cycles off after 2 seconds. With a new PSU being installed I can't believe that it would be it and all I can think now is that it's the Mobo or the CPU.

 

[Turtle Rig]

Ok I think I know whats going on. I think the CPU might be DOA. I have experience with this exact problem happening. Things turn on for a second or so then turns off. Then turns on again and turns off. It is not the PSU, your PSU is fine. I think IMHO something is up between the motherboard and the CPU. Also your RAM is fine as well. In my past experiences what your experiencing I would put my money on the fact that its a CPU that as I said is DOA or the MOBO is DOA. I put my money on the CPU being DOA. However theres one more thing, can you update the BIOS from behind the motherboard it should have a button. Stick the BIOS in a USB stick and flash it to latest version. After doing this I personally would send back the motherboard + the CPU and get new one. When did you purchase this box. You should have 30 days possibly depending on where you bought the stuff from to return it and say its DOA and then buy the same mobo and cpu again. This is all I can think of to tell you and Im sorry its bad news but no worries youll be back up in no time depending on where you bought it from and how there turn around time is.😢🙌☮🎗✝🚔🍩

 

[Swollengoat555]

Ok I think I know whats going on. I think the CPU might be DOA. I have experience with this exact problem happening. Things turn on for a second or so then turns off. Then turns on again and turns off. It is not the PSU, your PSU is fine. I think IMHO something is up between the motherboard and the CPU. Also your RAM is fine as well. In my past experiences what your experiencing I would put my money on the fact that its a CPU that as I said is DOA or the MOBO is DOA. I put my money on the CPU being DOA. However theres one more thing, can you update the BIOS from behind the motherboard it should have a button. Stick the BIOS in a USB stick and flash it to latest version. After doing this I personally would send back the motherboard + the CPU and get new one. When did you purchase this box. You should have 30 days possibly depending on where you bought the stuff from to return it and say its DOA and then buy the same mobo and cpu again. This is all I can think of to tell you and Im sorry its bad news but no worries youll be back up in no time depending on where you bought it from and how there turn around time is.😢🙌☮🎗✝🚔🍩

I was already leaning towards those two being the main culprit anyway. But the real question is why would they work so fine for 5 days straight and then crap out? is it just a failure on making and just can't handle the stress? I can easily just get a refund from Amazon because that is where I got them. I actually updated the BIOS to the latest version the day after I got it all set up. But I can do it again easily. At this point I just wanted to try new ones and hope for the best. If I did should I just go for the same ones again?, or try another Mobo. I have an MSI 550 in my cart just in case. It is essentially the same one instead of Asus.

 

[Turtle Rig]

I hear you man, EVGA makes good PSU's as I said your rig is rock solid sexy rig. The fact it worked then crapped out is sorta a mystery. BTW do you have a LCD screen on the motherboard which gives you a error code? Also one more question when you turn the darn thing on and it turns off, before it turns off if you can stare right where the RAM and CPU are there are lights there and it shows a red light if its RAM and above it shows a red light if the CPU is bad. See if you can notice that. As for the fix, send it back to amazon they are great with this and re buy the motherboard and CPU and install the cooler properly which Im sure you already did and you should be good to go. Also Im assuming the video card fans spin briefly in the time the PC posts for a second thens turns off. Just for the hell of it I would send the RAM back as well just for the hell of it and to ease your mind. Then re buy CPU and MOBO and RAM and install it and you should be good to go my friend. Stay on top of this also Amazon has chat if you google it you can talk to someone immedietaly and do it that way as it is simpler then calling them being put on hold and having some guy from India with a accent trying to help you. Good Luck my friend.👍☮👶

 

[rgd1101]

Even after removing CMOS and also removing the GPU and trying to use Onboard video to see if it would even boot and stay up and it still just cycles off after 2 seconds. With a new PSU being installed I can't believe that it would be it and all I can think now is that it's the Mobo or the CPU.

ryzen don't have igpu, check the ASUS Q-LED?

ROG STRIX B550-A GAMING | Motherboards | ROG Global

www.asus.com

 

[Swollengoat555]

I hear you man, EVGA makes good PSU's as I said your rig is rock solid sexy rig. The fact it worked then crapped out is sorta a mystery. BTW do you have a LCD screen on the motherboard which gives you a error code? Also one more question when you turn the darn thing on and it turns off, before it turns off if you can stare right where the RAM and CPU are there are lights there and it shows a red light if its RAM and above it shows a red light if the CPU is bad. See if you can notice that. As for the fix, send it back to amazon they are great with this and re buy the motherboard and CPU and install the cooler properly which Im sure you already did and you should be good to go. Also Im assuming the video card fans spin briefly in the time the PC posts for a second thens turns off. Just for the hell of it I would send the RAM back as well just for the hell of it and to ease your mind. Then re buy CPU and MOBO and RAM and install it and you should be good to go my friend. Stay on top of this also Amazon has chat if you google it you can talk to someone immedietaly and do it that way as it is simpler then calling them being put on hold and having some guy from India with a accent trying to help you. Good Luck my friend.👍☮👶

So I got it start boot sequence just now after reinserting the GPU again and making sure everything was connected, my light sequence is this; DRAM-yellow light, CPU-red light, VGA-white light, then green light for starting boot. Does that mean the CPU is bad? also what would the yellow lights mean for DRAM? I just got them and they are registering fine in the system info when it was on, so would I really need to get new RAM? or just get a replacement CPU. Another post on this website mentioned that the CPU debug LED means that the Mobo can't use the CPU, that the cables possibly aren't connected properly or something is bent mayhaps, but considering this happened days after it was already working and didn't get moved around, I still feel like this is a bad CPU.

 

[Swollengoat555]

Also I want to make it clear that none of the lights stay on once it boots(if it boots) I just retried booting it and it went yellow for DRAM, red for CPU, then white for VGA, then green for boot, started the ASUS startup screen, then cut off after about 7 seconds before it could even finish the screen where I can enter BIOS.

 

[wi5pa]

it will use the lights as a check through, then go green and boot,
If it went onto the yellow one a second time then you need to check the ram,
reseat (blow out any dust from slots etc) and try again, or try them one at a time to find the faulty ones

 

[Swollengoat555]

What's confusing is, sometimes if fails to start up even on the DRAM check. SO it goes yellow light for DRAM, then cuts off and flashes red on the GPU light next to the 8 pin connector. other times it goes all the way through the light sequence into green boot light and then cuts off. and some times it goes into windows and cuts off after booting fully up. So I have no idea how to even isolate things when any one of them can happen at any given time. At this point I am just ready to return the Mobo, the CPU, and the RAM and get all new ones sent while I still have the fully options for refunds/replacements from Amazon.

 

[jasonf2]

Specifically what is your RAM and are you running SPD timings and clock or XMP / custom timings and clocks? If running XMP or custom clocks set back to SPD and see if it stabilizes. Ryzen, RAM overclocks and I have not got along well. The intermittent stuff you are describing sounds alot like RAM running right on the edge of stable.

 

[Swollengoat555]

Specifically what is your RAM and are you running SPD timings and clock or XMP / custom timings and clocks? If running XMP or custom clocks set back to SPD and see if it stabilizes. Ryzen, RAM overclocks and I have not got along well. The intermittent stuff you are describing sounds alot like RAM running right on the edge of stable.

Ram was set to 2666hz in BIOS as per what it said the hz was on the sticks. I was not overclocking anything and didn't change any settings in BIOS besides setting the RAM speed to what it said it was on the sticks. I think naturally the BIOS had it running at the lowest hz which was 21-- something. I have 2 sticks of Corsair Vengeance LPX DDR4 2666hz 16GB a piece. (Note that these did not come in a pair, but were running the PC just fine up until I bought my second set of RAM sticks) The second set I DID buy in a pair and those are the HyperX DDR4 16gbx2 2666hz sticks. Both are 1.20V as well. Should I just return the Corsairs and buy another set of the HyperX? or is the 2666hz too slow for the CPU and Mobo I am trying to use, and would it be better to run a higher HZ like 3200?

 

[jasonf2]

Ram was set to 2666hz in BIOS as per what it said the hz was on the sticks. I was not overclocking anything and didn't change any settings in BIOS besides setting the RAM speed to what it said it was on the sticks. I think naturally the BIOS had it running at the lowest hz which was 21-- something. I have 2 sticks of Corsair Vengeance LPX DDR4 2666hz 16GB a piece. (Note that these did not come in a pair, but were running the PC just fine up until I bought my second set of RAM sticks) The second set I DID buy in a pair and those are the HyperX DDR4 16gbx2 2666hz sticks. Both are 1.20V as well. Should I just return the Corsairs and buy another set of the HyperX? or is the 2666hz too slow for the CPU and Mobo I am trying to use, and would it be better to run a higher HZ like 3200?

Set the speed back to auto and your lower frequency and see if it stabilizes. Anytime you change settings from SPD it is an overclock. That 21-something number is the "stable" setup for the RAM from the manufacturer. When you say Corsair I am also assuming that it is Vengeance in which I have had no luck taking any higher than SPD on Ryzen. You also have what is called a "mix and match" on your RAM. That is simply taking multiple RAM sets and throwing them in the same machine. It can be a recipe for disaster. Your board manufacturer will have released a QVL (Qualified Vendor List) for your motherboard. It is usually downloadable from their website. RAM on the high end is advertised to a "tested" frequency. What they don't tell you is that not every system can handle running at that speed and you often have to manually set timings and usually voltage to get there. Just bumping frequency will almost always destabilize the system. The QVL will have a list of verified RAM kits and the actual frequency they have been verified at. Depending on the vendor they will sometimes give you stable timings as well. I would find one kit marked at your target capacity from the QVL and stick with that. Others have also noted that GSKILL and Hyperx have been playing a little better with Ryzen than Corsair, but I cannot verify that.

Go to page 5-1 in your motherboard manual. Make sure that you have slotted your RAM correctly. Personally with the stability issues you are having I would only slot the hyperx kit in the short term, run it at SPD and see if it levels off. After you have a stable system then make some decisions about other kits if this works.

One other thing to mention. I never overclock or tweak with anything on a new build until the OS is up and going and stable. This is especially important on RAM. When RAM runs right up against the edge the machine posts but things get corrupted. On an established system this usually just means rebooting. If this is going on while installing your OS your operating system will be corrupted. If you bumped that frequency before installing your OS there is a very good chance that you need to reinstall your OS before the system will be stable. I made that mistake one time on a build I was working on and won't do that again.

Memtest86 is a utility that tests RAM. If you are planning to run anything over SPD I would verify it with this utility. It has to be installed on bootable media like USB and takes a while but anytime you do any overclock on RAM verification is a must. You are going for zero errors.

 

[Swollengoat555]

So I already started a refund on all 4 sticks of RAM and ordered a full 4 (2 paired sets) of 3600 DDR4 corsair vengeance pros. the 16gb x2 (x2) for 64gbs. It is one that is on the compatibility list and should be arriving tuesday. Though I am still going to set my RAM speed back to the lowest in BIOS once I get it to boot there just to see if it will stabilize and allow me to even run it. My good friend who knows building and repairing PCs also said he is betting that it is the RAM and that weird pairing issue. So fingers crossed. The one thing that still confuses me is the Timings part of RAM sticks. Does the system not automatically detect what timing they should run at or do I have to do that myself. I have never had to do that for DDR3 on any of my old systems. I still feel that this somehow goes back to manually setting the bios to have them run at the 2666. Last thing I would like to add is I did not adjust any speeds until after the OS was installed and I had some drivers updated, but I think I did adjust them before properly getting the updated BIOS.

 

[Swollengoat555]

Set the speed back to auto and your lower frequency and see if it stabilizes. Anytime you change settings from SPD it is an overclock. That 21-something number is the "stable" setup for the RAM from the manufacturer. When you say Corsair I am also assuming that it is Vengeance in which I have had no luck taking any higher than SPD on Ryzen. You also have what is called a "mix and match" on your RAM. That is simply taking multiple RAM sets and throwing them in the same machine. It can be a recipe for disaster. Your board manufacturer will have released a QVL (Qualified Vendor List) for your motherboard. It is usually downloadable from their website. RAM on the high end is advertised to a "tested" frequency. What they don't tell you is that not every system can handle running at that speed and you often have to manually set timings and usually voltage to get there. Just bumping frequency will almost always destabilize the system. The QVL will have a list of verified RAM kits and the actual frequency they have been verified at. Depending on the vendor they will sometimes give you stable timings as well. I would find one kit marked at your target capacity from the QVL and stick with that. Others have also noted that GSKILL and Hyperx have been playing a little better with Ryzen than Corsair, but I cannot verify that.

Go to page 5-1 in your motherboard manual. Make sure that you have slotted your RAM correctly. Personally with the stability issues you are having I would only slot the hyperx kit in the short term, run it at SPD and see if it levels off. After you have a stable system then make some decisions about other kits if this works.

One other thing to mention. I never overclock or tweak with anything on a new build until the OS is up and going and stable. This is especially important on RAM. When RAM runs right up against the edge the machine posts but things get corrupted. On an established system this usually just means rebooting. If this is going on while installing your OS your operating system will be corrupted. If you bumped that frequency before installing your OS there is a very good chance that you need to reinstall your OS before the system will be stable. I made that mistake one time on a build I was working on and won't do that again.

Memtest86 is a utility that tests RAM. If you are planning to run anything over SPD I would verify it with this utility. It has to be installed on bootable media like USB and takes a while but anytime you do any overclock on RAM verification is a must. You are going for zero errors.

I reset all my speeds in BIOS back to AUTO and am only running the x2 HyperX sticks. Currently at almost an hour with it running super smooth with no crashes. Will keep monitoring but that might have solved it. Like I said above though I do have two paired sets coming in of the same model and size on Tuesday to give me the 64Gbs I want and have them be more stable for this build.

EDIT: 1 hour in and WINDOWS initiated shutting down... not even just a full power drop. Booted back up just fine though afterwards.

 

[jasonf2]

I reset all my speeds in BIOS back to AUTO and am only running the x2 HyperX sticks. Currently at almost an hour with it running super smooth with no crashes. Will keep monitoring but that might have solved it. Like I said above though I do have two paired sets coming in of the same model and size on Tuesday to give me the 64Gbs I want and have them be more stable for this build.

EDIT: 1 hour in and WINDOWS initiated shutting down... not even just a full power drop. Booted back up just fine though afterwards.

If you installed windows with the increased clock you may need to reinstall it again at SPD. The last time I went through something like this I had bumped the clock when setting it up and windows appeared to install just fine. Even after bumping back clock BSOD kept coming until I reinstalled windows. After that it leveled out. I am not advocating any RAM but I would shy away from the corsair if you want to bump up the clock at all. Ryzen and the Vengeance don't seem to get along above SPD speeds.

 

[Swollengoat555]

If you installed windows with the increased clock you may need to reinstall it again at SPD. The last time I went through something like this I had bumped the clock when setting it up and windows appeared to install just fine. Even after bumping back clock BSOD kept coming until I reinstalled windows. After that it leveled out. I am not advocating any RAM but I would shy away from the corsair if you want to bump up the clock at all. Ryzen and the Vengeance don't seem to get along above SPD speeds.

No i installed windows at auto and then bumped afterwards thinking that it was a good thing. So far it's running stable and my buddy and I are doing stress tests for memory and cpu. Close to 1 hour and 10 minutes with no issues this time. I think windows shut down for some kind of update. It just went into a shutdown and not a crash.

 

[wi5pa]

run memtest 84, as benchmarking stresses wont pick up on memory errors.

 

[Turtle Rig]

So I got it start boot sequence just now after reinserting the GPU again and making sure everything was connected, my light sequence is this; DRAM-yellow light, CPU-red light, VGA-white light, then green light for starting boot. Does that mean the CPU is bad? also what would the yellow lights mean for DRAM? I just got them and they are registering fine in the system info when it was on, so would I really need to get new RAM? or just get a replacement CPU. Another post on this website mentioned that the CPU debug LED means that the Mobo can't use the CPU, that the cables possibly aren't connected properly or something is bent mayhaps, but considering this happened days after it was already working and didn't get moved around, I still feel like this is a bad CPU.

Sorry for the late reply Ive been busy. Ok that red light you see for CPU is your answer the CPU is DOA. But possibly the MOBO can be DOA so you must send both back and not have another headache but Im 100 percent telling you that red light for CPU should never come on unless something is wrong in which case there is something wrong. The CPU is the culprit but as I said send the mobo back as well just for peace of mind.

 

[Swollengoat555]

Sorry for the late reply Ive been busy. Ok that red light you see for CPU is your answer the CPU is DOA. But possibly the MOBO can be DOA so you must send both back and not have another headache but Im 100 percent telling you that red light for CPU should never come on unless something is wrong in which case there is something wrong. The CPU is the culprit but as I said send the mobo back as well just for peace of mind.

Well the red light is only during the boot LED sequence. There isn't a light anywhere else.

 

[Swollengoat555]

Sorry for the late reply Ive been busy. Ok that red light you see for CPU is your answer the CPU is DOA. But possibly the MOBO can be DOA so you must send both back and not have another headache but Im 100 percent telling you that red light for CPU should never come on unless something is wrong in which case there is something wrong. The CPU is the culprit but as I said send the mobo back as well just for peace of mind.

After testing I'm just going to take your advice and return the mobo and the psu and get whole new ones. I'll change the board out to the better model that is more known, the b550 F board versus the A. The board was originally set to not be in stock till end of OCT then somehow got two updates saying it was now ready. So maybe it was rushed in processing?