Problems when starting / restarting a PC

[Komshija]

My previous PSU some 1,5 years after the purchase started developing some serious problems, so I replaced it with new FSP Hydro G 650W.
Everything works fine, except one little annoying issue which was present even when I had old PSU (Lepa MaxBron B700M PSU). With old PSU it happened more frequently, but now it happens rarely.

The issue is present when starting / restarting a PC. Please note that this happens rarely, eg. once in a 20. Simply put, I press the start button, PC starts, immediately shuts down and starts again all by itself. Sometimes when restarting a PC, it shuts down and then immediately starts, opposed to a standard "uninterrupted" restart.
PSU voltages are OK, CPU works OK, RAM has no errors and SSD has no errors as well. Could it be motherboard causing these problems?

 

[IceScreamer]

My guess is the motherboard, check for bulging caps. Last week had to change cap on a 478 motherboard and it had similar issues as yours.

 

[_UV_]

From a motherboard perspective this looks like retraining due to OC or non JEDEC memory timings or both. But if it occurs with new PSU very rare (especialy compared to old one) when i can assume voltage ripple probably caused by bulged caps inside old PSU.

 

[Komshija]

My guess is the motherboard, check for bulging caps. Last week had to change cap on a 478 motherboard and it had similar issues as yours.

There are no bulging caps as I recall during the last cleaning, but I will inspect them closely once more. I had this problem from the day one since I bought components and built this PC. At first it happened very rarely (eg. once in 20) and then more frequently (eg. once in 10) - that was with old PSU. With new PSU it's about 1 in 20.

From a motherboard perspective this looks like retraining due to OC or non JEDEC memory timings or both. But if it occurs with new PSU very rare (especialy compared to old one) when i can assume voltage ripple probably caused by bulged caps inside old PSU.

Memory is overclocked via XMP from default 2133 to 3000 MHz. I haven't manually change timings. CPU is stock.
Surprisingly, old PSU didn't have any bulging or leaking capacitors.

 

[TheoneandonlyMrK]

First thing I would try is the on off switch, when worn they can do stuff like this.

 

[_UV_]

Memory is overclocked via XMP

Non JEDEC "defaults", so MB sometimes have troubles determining is it stable or not and trying to retrain (restart)
LN overclockers do same process: reseting MB multiple times and adjusting timings by small margins till they get desired results

Surprisingly, old PSU didn't have any bulging or leaking capacitors.

it was just assumption

If you have time you may try this:
discard all OC for both CPU and RAM
and try to power up/down system multiple times, not reset, full power cycle
if you meet same double-tripple starts - most possible cause small voltage ripples in PSU, not harmfull, just annoying, or this just a poor engineered/made board (like many ASUS from time of Pentium4/Athlon64 having double starts for no reason and telling people OC failed, despite there are was no OC at all)

if nothing occurs, then try overclock CPU and RAM independently till you meet this behavior again

and finally OC both

 

[Deleted member 178884]

I've got this issue too myself every time, I press the power button it shuts off then comes back on to boot fine.

 

[Komshija]

... if you meet same double-tripple starts - most possible cause small voltage ripples in PSU, not harmfull, just annoying, or this just a poor engineered/made board (like many ASUS from time of Pentium4/Athlon64 having double starts for no reason and telling people OC failed, despite there are was no OC at all) ...

It could be that this is the problem, since it's unlikely the PSU, RAM, CPU or SSD. I will see what happens after a few weeks. Right now the system is OC'ed at maximum "safe" voltage. The CPU (i7 6700K) is set to 4600 MHz @ 1,32 V adaptive (stable, 1h 20min Prime95 26.6 test, 10 times Cinebench R15 test) and RAM's are at XMP, as they are from the day one.

I've got this issue too myself every time, I press the power button it shuts off then comes back on to boot fine.

In my case it happens rarely, but I've noticed during restarts that it's more likely to happen after stress (eg. if I restart the system within one minute after heavy gaming, synthetic benchmarks etc.).

The other thing is when I enter BIOS and and exit (even without modifying anything), it shuts down and restarts every single time rather than having an uninterrupted exit and restart. So I suspect poorly engineered mobo.

 

[Deleted member 178884]

It could be that this is the problem, since it's unlikely the PSU, RAM, CPU or SSD. I will see what happens after a few weeks. Right now the system is OC'ed at maximum "safe" voltage. The CPU (i7 6700K) is set to 4600 MHz @ 1,32 V adaptive (stable, 1h 20min Prime95 26.6 test, 10 times Cinebench R15 test) and RAM's are at XMP, as they are from the day one.


In my case it happens rarely, but I've noticed during restarts that it's more likely to happen after stress (eg. if I restart the system within one minute after heavy gaming, synthetic benchmarks etc.).

The other thing is when I enter BIOS and and exit (even without modifying anything), it shuts down and restarts every single time rather than having an uninterrupted exit and restart. So I suspect poorly engineered mobo.

Doubt it - my z270x gaming 7 is amongst the best z270's available.

 

[Komshija]

Not your's, mine MSI Z170A Krait Gaming 3X. But, I also might be very wrong...

 

[Deleted member 178884]

Not your's, mine MSI Z170A Krait Gaming 3X. But, I also might be very wrong...

Doubt it's the motherboard still - It even happens on my x58 ud3r rev 1.6

Not your's, mine MSI Z170A Krait Gaming 3X. But, I also might be very wrong...

And I somehow doubt it's the psu - I've got a power pro 11 750w and a seasonic snow silent 750w, both bought new.

 

[Caring1]

Any error codes showing on the Motherboard just prior to the automatic reboot?
You might have to go in to the BIOS and select halt on ALL errors to catch it.
I remember reading a while back the Krait boards are rubbish, but can't confirm that, perhaps one of the reviewers here can. @cadaveca
I'd also suggest replacing the 3V battery, just in case.

 

[jboydgolfer]

I'm not going to get into the science of my idea, nor will i argue it, but based on my past experience, I have noticed similar issues when my power and reset header plugs were not plugged in properly.

I've had people claim that it doesn't matter which way they go in, and several other rebuttles to this, but I have to ignore them, because I've had computers do weird things at power up ,and shut down, which were resolved (after trying many other things) by properly plugging the header plugs according to polarity or guidlines in manual.

Maybe this will help your issue as well

 

[Deleted member 178884]

I'm not going to get into the science of my idea, nor will i argue it, but based on my past experience, I have noticed similar issues when my power and reset header plugs were not plugged in properly.

I've had people claim that it doesn't matter which way they go in, and several other rebuttles to this, but I have to ignore them, because I've had computers do weird things at power up ,and shut down, which were resolved (after trying many other things) by properly plugging the header plugs according to polarity or guidlines in manual.

Maybe this will help your issue as well

I will have to check this - thanks.

 

[eidairaman1]

I'm not going to get into the science of my idea, nor will i argue it, but based on my past experience, I have noticed similar issues when my power and reset header plugs were not plugged in properly.

I've had people claim that it doesn't matter which way they go in, and several other rebuttles to this, but I have to ignore them, because I've had computers do weird things at power up ,and shut down, which were resolved (after trying many other things) by properly plugging the header plugs according to polarity or guidlines in manual.

Maybe this will help your issue as well

It is true, some switches have polarization. Look at switch and header.

My only other things to check for, case shorting -ensure no other standoffs in case touching pcb, ensure all plugs are seated. Change bios/cmos/rtc battery. Reset bios to optimized defaults for troubleshooting.

That last one relates to OCs which you actively have one, if problem stops after last 1 it is an unstable oc.

 

[Komshija]

First thing I would try is the on off switch, when worn they can do stuff like this.

What about the on off switch ?

It happened from the day one with previous PSU. I bought all components brand new, unopened and unused.

Any error codes showing on the Motherboard just prior to the automatic reboot?
You might have to go in to the BIOS and select halt on ALL errors to catch it.
I remember reading a while back the Krait boards are rubbish, but can't confirm that, perhaps one of the reviewers here can. @cadaveca
I'd also suggest replacing the 3V battery, just in case.

Nothing besides standard system debug lights (very short CPU-DRAM-CPU-DRAM blinking) and then followed by the 1-2 second VGA blink which immediately dissapears after a splash screen appears. The shutdown-restart cycle always happens after prior to the mentioned VGA blink.
I will try with the new battery if nothing else works, but I somehow doubt it since MOBO was produced in the early 2016 and I built my system in mid July 2016. This problem was there from the day one. There were also no BIOS resets to defaults like it's known to happen with defective or empty battery. Could it really be the battery?

I'm not going to get into the science of my idea, nor will i argue it, but based on my past experience, I have noticed similar issues when my power and reset header plugs were not plugged in properly.

I've had people claim that it doesn't matter which way they go in, and several other rebuttles to this, but I have to ignore them, because I've had computers do weird things at power up ,and shut down, which were resolved (after trying many other things) by properly plugging the header plugs according to polarity or guidlines in manual.

Maybe this will help your issue as well

All connectors are seated properly. I'll check the polarity once more - maybe that was really the issue and I overlooked it when building this system.

It is true, some switches have polarization. Look at switch and header.

My only other things to check for, case shorting -ensure no other standoffs in case touching pcb, ensure all plugs are seated. Change bios/cmos/rtc battery. Reset bios to optimized defaults for troubleshooting.

That last one relates to OCs which you actively have one, if problem stops after last 1 it is an unstable oc.

So far CPU and GPU were stock - I only OC'ed them for some testing. Just very recently I applied CPU OC. Prior to that it was only RAM's that were OC'ed via XMP to their advertised speed.

I noticed that it's more likely to happen after a shutdown and then "immediate" start (eg. after 10 seconds) and after some heavy system load (eg. Cinebench) and then restart within less than one minute. But it also happened in other ordinary/normal situations.

 

[phanbuey]

You might be able to fix it if you disable quickboot. Most likely memory doesn't initialize properly on the quick-boot and it thinks the ram is unstable and loops.

 

[Komshija]

@phanbuey It's disabled.

I checked header polarities and they were properly installed.

 

[Hood]

I'm not going to get into the science of my idea, nor will i argue it, but based on my past experience, I have noticed similar issues when my power and reset header plugs were not plugged in properly.

I've had people claim that it doesn't matter which way they go in, and several other rebuttles to this, but I have to ignore them, because I've had computers do weird things at power up ,and shut down, which were resolved (after trying many other things) by properly plugging the header plugs according to polarity or guidlines in manual.

Maybe this will help your issue as well

Also, don't forget to walk around the room counter-clockwise 3 times before pushing the switch. I've been doing this for years, and my PC always starts normally. Sometimes I accidentally go clockwise, and it doesn't seem to like that as much.

 

[TheoneandonlyMrK]

It happened from the day one with previous PSU. I bought all components brand new, unopened and unused.


Nothing besides standard system debug lights (very short CPU-DRAM-CPU-DRAM blinking) and then followed by the 1-2 second VGA blink which immediately dissapears after a splash screen appears. The shutdown-restart cycle always happens after prior to the mentioned VGA blink.
I will try with the new battery if nothing else works, but I somehow doubt it since MOBO was produced in the early 2016 and I built my system in mid July 2016. This problem was there from the day one. There were also no BIOS resets to defaults like it's known to happen with defective or empty battery. Could it really be the battery?
Switches can be faulty from day one or poorly designed so they double bounce, it's a momentary switch.

All connectors are seated properly. I'll check the polarity once more - maybe that was really the issue and I overlooked it when building this system.


So far CPU and GPU were stock - I only OC'ed them for some testing. Just very recently I applied CPU OC. Prior to that it was only RAM's that were OC'ed via XMP to their advertised speed.

I noticed that it's more likely to happen after a shutdown and then "immediate" start (eg. after 10 seconds) and after some heavy system load (eg. Cinebench) and then restart within less than one minute. But it also happened in other ordinary/normal situations.

A on off switch is a momentary switch and if faulty or by design faulty they can double bounce ,or not press.
It's also been there every time and received no consideration ,you think I suggest something out of order , i suggested it because with those symptoms I've fixed it via on button ,reset swap.

 

[Caring1]

My train of thought keeps going back to the actual Motherboard as the culprit behind your issues, heat might be a factor if it happens sometimes during a restart, but I tend towards a poorly written BIOS or perhaps an electrical issue in the design.
I guess the only way to find out is swap all your components into another brand motherboard of the same socket, or update the existing boards BIOS and see if that helps.

 

[Komshija]

Also, don't forget to walk around the room counter-clockwise 3 times before pushing the switch. I've been doing this for years, and my PC always starts normally. Sometimes I accidentally go clockwise, and it doesn't seem to like that as much.

I jump 5 times, touch my nose and bury an axe into a honey bucket, but it still happens.

A on off switch is a momentary switch and if faulty or by design faulty they can double bounce ,or not press.
It's also been there every time and received no consideration ,you think I suggest something out of order , i suggested it because with those symptoms I've fixed it via on button ,reset swap.

True, I never suspected that it could be on/off switch just because it happens rarely. I will visually check damaged circuitry/wires, since I have no extra on/off PC switch and no spare PC case.

Still, I suspect poorly engineered mobo. Why? Every single time when I enter BIOS, with or without modifying something and then exit, the system shuts down and restarts by itself. On my previous PC's and on every other PC I've dealt with, BIOS entrance/exit and OS loading were smooth - uninterrupted.

My train of thought keeps going back to the actual Motherboard as the culprit behind your issues, heat might be a factor if it happens sometimes during a restart, but I tend towards a poorly written BIOS or perhaps an electrical issue in the design.
I guess the only way to find out is swap all your components into another brand motherboard of the same socket, or update the existing boards BIOS and see if that helps.

I don't have any spare mobo and cannot really borrow one except buying. Now if I open my wallet and go for the Asrock Z270 Extreme 4 (the candidate if my mobo fails), my wallet will be thinner for totally unplanned 160 €.

 

[ne6togadno]

try power up the pc from header with screw driver. this way you can check if power button is causing the troubles

 

[Komshija]

try power up the pc from header with screw driver. this way you can check if power button is causing the troubles

I just find out that this is actually possible (I've never done it), but it since it happens rarely I will restrain myself from using that method to avoid any damage to the components. My best luck right now is to visually inspect whole on/off circuitry for damages.

Yesterday I replaced CMOS battery and will see if a few days whether a faulty battery was the issue, although I doubt it.

EDIT: The battery replacement didn't solve the problem and on/off switch & circuitry seemed perfectly OK without any visible damages.

SOLUTION: on page two...

 

[ne6togadno]

it isnt that hard or dangerouse if you are careful.
you need simply to touch both pwr pins with tip of screw driver and pc should start. no need to apply force or anything. if you remove only pwr connector and leave all other connectors attached to the header you will have only 2 pwr pins available and you wont be able to mess up anything if you touch wrong pins.

ps you dont need to hold screw driver to the pins after pc start. just touch and put away.