Computer won't POST most of the time, but will occasionally POST

Toggle sidebar Toggle sidebar
2

20sean.walker

Sep 27, 2018
13
0
10
Computer has been a nightmare this last 3 weeks when I replaced the motherboard and the CPU. Moved from a mATX to an ATX motherboard, AM3+ GA-78LMT-USB3 R2 with an FX-8350 CPU. Lets start from the start of my experience.

When I replaced my mobo and CPU, I tried to turn on my computer. Fans and lights turned on, but no POST. Was frustrated, dealt with this for about a week. If I moved the RAM around, it sometimes booted. From there I had the issue of grey screens. This is when I would load up any game, wait for about 10 minutes, and the screen would turn grey. Audio would play, cut out, screen would then turn off, mouse and keyboard LEDs would freeze, then it would reboot.

Luckily for me, during this phase, my PC would actually boot normally. It was just crashing every time I would open a game, kinda what I do on it. Replaced the PSU to see if that was the issue, but to my luck, no good.

Now I have a new issue. I one day thought maybe I could fix this issue by altering the BIOS settings on my CPU by undervolting it and seeing if it would not crash. No good. Tried again my altering the MHz of my ram down by one (1600 to 1333) and computer reboot. But, instead of a BIOS, I was met with no POST. The light on my motherboard flashes orange for about a millisecond, then off, wait for about 1 seconds, repeat. Fans and lights turn on like before, and you can kind of hear the PC reboot itself very quickly.

With my luck, I can occasionally get it to boot. When I had both of my RAM sticks in (4GBx2), it would never boot. Taking one out would sometimes POST and get to BIOS. Able to play CSGO by changing my global shadows to medium, but anything higher will crash the PC. Playing any game with graphics that meet up with my GPU and CPU usually crash it.

I have been able to avoid all out loss of computer usage by pressing the power button on the windows key and setting it to "Sleep". Everything turns off, but able to reboot it and get back to everything.

Just a note, possibly the problem, as I said before I had a mATX mobo before this change. This means my standoffs only were able to cover these sections. When I installed my mobo, I only had half of the necessary standoffs for an ATX mobo. I am really hoping this is the main issue, because otherwise this is going to be an expensive fix. And no, my case does not have any more standoffs as I bought this PC prebuilt.

Now, I just mostly need a diagnosis for this issue. I resat almost everything in the PC, including GPU, CPU, CPU Heat sink, all the cables, and finally the MOBO. I reapplied thermal paste on the CPU. I replaced all my GPU drivers by using DDU.

Now, for my view on this, I hope it is the standoffs. My gut says that it is my RAM due to it being about 7 years of age, used almost the whole 7 years on gaming. Another part of me wants to believe that it is the motherboard, which I hope it isn't. You can't really find cheap AM3+ motherboard these days that can support a FX-8350. Worse case scenario says that it is the GPU or the CPU. GPU is about 2 years old and the CPU is about 3 weeks old.

Finally, to avoid the repetition, I have ran all the other programs to test what the problem is, forgot the names. Only one I have not tested yet is memtest. Should test this ASAP.

PC specs:

MOBO: GA-78LMT-USB3 R2 (rev 1.0)
GPU: MSI XFX RX 470 4GB Armor Edition
CPU: AMD FX-8350
PSU: CX 550W Corsair
RAM: Crucial 8GB (4GBx2)
HDD: 1TB (as old as the original build below from 2014)

Original build is here:
https://www.amazon.com/CybertronPC-Patriot-Gaming-Desktop-Discontinued/dp/B00D1KX2M4

Replaced literally everything except the power buttons, the 1 case fan, and the case itself.

Any feedback would be great, just need to go in the right direction before I spend more money on something that doesn't fix it. The sooner the better.
 
Sort by date Sort by votes
Darkbreeze

Darkbreeze

Retired Mod
Jun 24, 2014
80,411
7,631
176,790
Is the CPU overclocked?

Is the bios the MOST recent version, or an older version?

Missing standoffs won't cause these issues. Standoffs in the WRONG places, most certainly could cause just about ANY kind of issue, including magic smoke. Let me clarify that though with a caveat. If the board is flexing, because there are not standoffs where they need to be and you have a heavy CPU cooler or GPU card installed, then it could certainly create issues up to and including broken traces on the motherboard. Put standoffs where they need to be. Go buy them, take it apart and install them. Even if it is not THE issue, it is AN issue.

Did you buy the motherboard you are using now, new or used? Are the problems EXACTLY the same as with the old motherboard?

That motherboard has integrated video. Have you tried removing the GPU card and using only the integrated motherboard graphics?

IF and when it does POST, if you do not go into windows, are there problems, or will it continue to run in the non-windows environment like the bios or a DOS command prompt, or Ubuntu instance ok?

Memtest isn't going to be helpful if the system isn't stable even outside the Windows environment. If it is, then running Memtest86 from Passmark, free version, is pretty much a no brainer. Four passes of the full eleven tests.
 
Upvote 0 Downvote
rgd1101

rgd1101

Don't
Moderator
Nov 7, 2011
73,311
3,855
176,290
MERGED QUESTION
Question from 20sean.walker : "Computer won't POST"





 
Upvote 0 Downvote
cherry blossoms

cherry blossoms

Reputable
Apr 13, 2016
890
57
5,490
Double checking all your connections and seating of components a good thing to check. As darkbreee has indicated, you need to have the proper number and placement of standoffs to avoid board flexing while doing so, and to avoid any flexing with thermal cycling.

Though on paper and in theory, the 760 chipset will run that CPU, in practice, a 4+1 phase for 125W is probably not going to end well. The 970 and 990 series chipsets are more commonly associated with stability with that cpu, provided you have the power phases available on the board, AND adequate airflow to cool the CPU and said VRMs.
 
Upvote 0 Downvote
Darkbreeze

Darkbreeze

Retired Mod
Jun 24, 2014
80,411
7,631
176,790
That board is probably the best of the worst choices you could use, as far as power phase and chipset are concerned. However, the problems that would be LIKELY from using it are not the same as the problems he is experiencing, UNLESS it is a used board and somebody ELSE was using a 125w or higher FX8 CPU on it while overclocking or otherwise punishing it and then sold it to him after it started having problems.

It WOULD be helpful to know what the model number of the micro-ATX board he was using was, for reference, and to determine whether that too was a board without VRMs capable of supporting that CPU.

My very first inclination would be to almost unilaterally say that problems of this nature tend to be related to bent pins on the motherboard. If you have not done so, I would pull the CPU and using a magnifying glass or high powered reading glasses, check the pin structure on the CPU very closely. If there are any pins that look out of place or "off", anything that does not look uniform and like the rest of the pattern of pins, then it's likely there is a bent pin. One bent pin could cause exactly this type of problem, among about a hundred others. All of which could closely resemble problems that are commonly found with failures on OTHER types of hardware.

Also, DID you have this problem BEFORE you switched motherboards, or did this specific problem begin after that?
 
Upvote 0 Downvote
2

20sean.walker

Sep 27, 2018
13
0
10


No, computer would always POST on the old motherboard. Something more weird I figured out yesterday. It doesn't crash with high-graphic games anymore. If I open any game after having the computer on sleep, it crashes. No grey screen, just a black screen, a couple of seconds go by, and the PC stops. I go through the process of trying to get it to start, and when it does start, I can just play whatever game I want to on whatever graphics. Assuming this is a RAM issue. When I took out my CPU during the process to replace the thermal paste, I checked for bent pins. To my luck, nothing was bent. Will definitely just look for standoffs, I think the place down the road sells them cheap.

And I have no idea what type of mATX motherboard I had. All I know it was old and barely worked. I can try and find it on a shelf at home, probably have it somewhere.

Board was bought new.
 
Upvote 0 Downvote
2

20sean.walker

Sep 27, 2018
13
0
10


Have tested it before, got it up and going for the past 3 days. Had some weird crashes and then after that they work fine. It's weird. If I leave my computer up for a bit before I actually do anything, it works fine. Otherwise it just crashes to a randomly colored screen. POSTs all the time though now. There was a standoff in a weird place not connected to anything, so I moved it to a place where it should be. Worked just fine until like a couple days ago, and continuously crashes sometimes. Othertimes it works fine. I replaced all the ram to 16GB (8x2) and adjusted the settings to downvolt it to 1333 (originally 2133). It worked, for a while. and had some issues yesterday. I reset the CMOS by taking out the battery and putting it back in 30 mins later. Got it finally running when I left it on after start up and it went into sleep mode. Turned it back on and ran smoothly for the rest of the night. Left it in sleep mode last night to see if it was just something that was causing it to crash during startup. (I looked at startup settings and everything I had was the same as it always has been.) Will update sooner than later to see if I find a solution.

(Also, the PSU is about 2 weeks old. It would be very unlikely if it was the problem now due to it actually starting up.)

If there is a possible cause to what I just said, leave me a response, would be appreciated.
 
Upvote 0 Downvote
Darkbreeze

Darkbreeze

Retired Mod
Jun 24, 2014
80,411
7,631
176,790
Even brand new Seasonic Prime Ultra Titanium power supplies, right out of the box, can be bad. Don't ever think a piece of hardware can't be the problem just because it's new, or because sometimes it works. That is a good way to defeat all troubleshooting methods right out of the gate.

Does that mean yours is bad? No. Should you still test it? Yeah, you should. It costs you nothing in terms of wasted time to test it. Maybe three or four bucks for a really cheap multimeter from somewhere like Harbor freight, which can probably be used for a lot of other things afterwards.

As far as having a standoff in the wrong spot, this is the most likely culprit and historically, in almost every situation I've seen where somebody had done that, it resulted in damage to the motherboard by touching a trace and either damaging it or shorting something out. The damage doesn't have to be right where the standoff was located, it could have caused the short to occur elsewhere on the board.

I'd remove the board and check every inch of it, very closely, with a magnifying glass and look for "off looking" or burnt spots. Bulging or leaking capacitors. Anything that seems unusual and not like the surrounding area.

Very important to NEVER allow a standoff to not be where it is supposed to be. I've seen standoffs short entire systems out before. Everything. PSU, motherboard, CPU, memory, drives. Anything is possible once you start routing power where it doesn't belong or shorting a current path directly to ground.
 
Upvote 0 Downvote
2

20sean.walker

Sep 27, 2018
13
0
10


Alright, have done some messing with the settings and such and have done some testing (PSU works, thank god)

Option 1: As I have said before, but now that I'm looking at this it may be another issue. Standoffs. I tried to place the standoffs near important components to see if it works. Seems to somewhat work. Open games like CSGO/SWTOR on max graphics, it will crash in about 5 minutes of gameplay. However, I can play games like GMOD on full/maxed out graphics with no crashes, while OBS is running. Now, graphics on GMOD aren't the best, so that leads me to another possible cause.

Option 2: Max GPU/CPU temp is too low.

There is a possibility one of my components (GPU/CPU) max temps are set a bit too low. Now, I want to just make it so it will not shut down if my temps on a GPU are 50. Same with CPU. I personally have read that a low max temp can cause what I've been experiencing. This would explain games like GMOD not crashing, but CSGO and SWTOR causing a crash. My PC does start right back up, now. That seems to be solved. My computer only crashes during the use of these applications. Have yet to experience a crash on anything else. So, when I get home, I will do this. Check everything, raise some things and lower some things, then go from there.

Also, if I can tonight, I'm going to go get a full set of standoffs and screws to replace my other parts. If none of these work, and I can't find a definitive answer, I will just bring it to GeekSquad or my local Computer Repair place.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom