Question pc randomly posts very slowly after being switched off at the power supply for several hours. Doesnt boot windows

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
pc randomly posts very slowly after being switched off at the power supply for several hours. doesnt even get to booting windows.
it posts so slowly it doesnt even make it to the windows booting stage.
it just shows bios splash for about 10+ seconds then a blinking cursor icon.
and it takes like 10+ second to reach the bios splash screen alone

Normally it takes about 2-3 seconds to reach bios splash screen. then almost immediately within a second windows starts booting.
If I just switch off the pc again at the power supply. and turn it back on. then press the power button it'll boot fast and normally like nothing ever happened. No slow post.
Why the inconsistent behavior? And whys it random? sometimes it doesnt happen at all. and the pc boots completely normally and fast.
Power supply is pretty new. EVGA GQ 750W 80+ gold only couple months old.
Motherboard is old but worked fine for months without issues.
CPU is also old.
but ram is new. a month or so old.
graphics card RTX 3060 is old but its the most recent thing I installed.
it was all working fine until this morning.
 
Last edited:
pc randomly posts very slowly after being switched off at the power supply for several hours. doesnt even get to booting windows.
it posts so slowly it doesnt even make it to the windows booting stage.
it just shows bios splash for about 10+ seconds then a blinking cursor icon.
and it takes like 10+ second to reach the bios splash screen alone

Normally it takes about 2-3 seconds to reach bios splash screen. then almost immediately within a second windows starts booting.
If I just switch off the pc again at the power supply. and turn it back on. then press the power button it'll boot fast and normally like nothing ever happened. No slow post.
Why the inconsistent behavior? And whys it random? sometimes it doesnt happen at all. and the pc boots completely normally and fast.
Power supply is pretty new. EVGA GQ 750W 80+ gold only couple months old.
Motherboard is old but worked fine for months without issues.
CPU is also old.
but ram is new. a month or so old.
graphics card RTX 3060 is old but its the most recent thing I installed.
it was all working fine until this morning.
Cheap test.

Replace the bios bat.
 
  • Like
Reactions: dudio

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
bios battery reading fine. tested with a meter and it still outputs juice at 3.248v close to a brand new one.

Problem is a one-off occurrence so hard to reproduce. but its still a concern that it will happen again or there will be some other kind of issues soon in the future.
And when I switched my pc off. by that I mean after windows was shut down PROPERLY with shutting down windows normally through start menu.
then with the switch on the back of the power supply to turn off power to the power supply at night.
 
Last edited:

DavidM012

Distinguished
Please review the PSU tier list

(and the older version from the power supplies section of the forum)

Evga gq <=750 (750 or less) Tier C low end.

See The brewing problem with gpu power design from Gamer's Nexus.

3000 series can generate transient power spikes of 550w for 100ms.

The gpu is heavily taxing your PSU. Should be 850w or greater to handle the transients. And more quality so higher up the list. However on this forum, the Corsair tx650m has been cited as a psu which can handle the transients well.

Evga have a platinum p2 on offer at the moment if you're in a position to take advantage of it.

So the question? Did the problem start before after you added the rtx 3060? 'It was all working fine until this morning'. In any case the documentation suggests your PSU isn't well recommended to run with a rtx 3000.

You are also overclocking your CPU. Set back to default for diagnostic purposes: does the problem settle any? The overclock is a bit more power draw and contributing to stressing the PSU along with the GPU.

Does that sound like a reasonable hypothesis to you?
 
bios battery reading fine. tested with a meter and it still outputs juice at 3.248v close to a brand new one.

Problem is a one-off occurrence so hard to reproduce. but its still a concern that it will happen again or there will be some other kind of issues soon in the future.
And when I switched my pc off. by that I mean after windows was shut down PROPERLY with shutting down windows normally through start menu.
then with the switch on the back of the power supply to turn off power to the power supply at night.
Try just doing a normal shutdown and leave the psu on.

See if the symptoms change.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
Please review the PSU tier list

(and the older version from the power supplies section of the forum)

Evga gq <=750 (750 or less) Tier C low end.

See The brewing problem with gpu power design from Gamer's Nexus.

3000 series can generate transient power spikes of 550w for 100ms.

The gpu is heavily taxing your PSU. Should be 850w or greater to handle the transients. And more quality so higher up the list. However on this forum, the Corsair tx650m has been cited as a psu which can handle the transients well.

Evga have a platinum p2 on offer at the moment if you're in a position to take advantage of it.

So the question? Did the problem start before after you added the rtx 3060? 'It was all working fine until this morning'. In any case the documentation suggests your PSU isn't well recommended to run with a rtx 3000.

You are also overclocking your CPU. Set back to default for diagnostic purposes: does the problem settle any? The overclock is a bit more power draw and contributing to stressing the PSU along with the GPU.

Does that sound like a reasonable hypothesis to you?

Ok it was just a one-time thing. It didn't happen again so far.
It was fine for over a week with the RTX 3060 added in.
it only happened after I changed ram power down mode to disabled. but it was fine for 2 days afterward. until this morning.
I have no clue whats going on.
It doesnt happen every time either. it was just a one-off event not replicable reliably.

my gpu under load normally only draws up to 190w maximum with furmark running. and cpu only 83w with prime95 running.
No problems so far with a normal shut down and leaving psu on. I can try that next time I go to bed and see if it boots properly next time. could just be psu doesnt like being cold-started with my pc setup.
I thought 750w psu would be overkill for my setup so far.

Maybe the 13 total fans I have in my pc are making too much of a current spike on startup? Most of them are hooked up to dedicated molex plugs or individual separate motherboard headers. Not using splitters.
thats including cpu case psu and gpu fans

My psu is the semi-modular one with eco mode. but I'm not using the eco mode.

No issues while gaming/using pc normally. no sudden shutdowns or anything.
 
Last edited:

DavidM012

Distinguished
Fans don't draw a lot of power particularly. If you have a lot of wires with them maybe the wiring is a bit dodgy, Figuring that out is a pain. Which out of a dozen might have a loose wire? Molexs and fan headers are generally crud. Are some of them RGB fans? I don't have any RGB fans I don't like it. 10 fans of 2.2 watts approx on molexes and fan headers is only 22 watts. So that by itself isn't pulling enough current to cause any problems.

No sudden shutdowns ok but the transients could be stressing your PSU or triggering some of it's safeguard circuits. See what happens if you power off completely, switch off the power switch of the psu if it has one, switch off the power at the wall, and press the power button a few times to drain the residual power. Leave it for 5 minutes and then switch on.

Also want to check the power outputs are normal if you can load hwmonitor. There's a bug somewhere clearly.

It's not necessarily the quantity of watts the psu is delivering, it's the quality of the components of the PSU. Hence the reason a decent psu with excellent capacitors and build quality, but only a 650w capacity would cope better with rough power conditions like transients.

A transient spike seems consistent with your story since it might not be occurring frequently but there is no way to detect exactly what's going on without specialized equipment.

Well if your troubles started after the GPU was installed what else could it be?
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
No rgb. i stripped all the rgb out.
HWmonitor is showing everything as completely fine no spikes anywhere at any time over many hours of uptime.

I triple checked all my fan cabling and its all secure and firmly hooked up. nothing loose or incorrectly wired.
Double checked all the wiring to the motherboard also. all firmly plugged in.
Bare essentials. no extra rgb or anything. Just power switch and hdd led lights both which work fine also.

my pc would post but it was abysmally slow and didnt make it to the windows logo screen before I gave up after about 15 seconds and powered it back off at the psu switch again. then switched it back on and it booted flawlessly in about 3-4 seconds to windows logo and windows booted quickly without any issues.
 

DavidM012

Distinguished
Yeah hwmonitor won't detect transient power spikes they're like invisible gremlins that torment you and rob you of your peace. In the Gamer's Nexus review they used some specialist diagnostic equipment to detect the transients.

All we have is our powers of deduction. The PSU isn't highly rated not top of the line you have an rtx 3060 and after you installed it the problems started.

So the struggle to start up. Well how about putting your old GPU back and trying it instead of the rtx 3060? Process of elimination. Proofs of puddings being in the eating? Science. Hypothesize and test.

Cables all securely hooked up. Only you can see that. Sure that the cable fasteners whatever they may be aren't squishing or pulling any wires? You can't always see bad connections or maybe wires that aren't really a strong gauge for what you're doing with them.

I have about 7 fans attached to my system via mobo headers or psu and it all works just fine and serenely but I'm not doing anything like 13 fans and extravagantly pushing the boat out of vague obscure specifications that nobody really thinks much about. Should be ok.

Who knows how many fans a molex connector can support? Nobody really thinks about it much because they simply don't often do things like that. It's getting into kind of grey areas a bit.

Something is clearly going on in the power delivery system and co-incidentally the symptoms converge with some documented meridians of the problem.

You simply can't rule out or dismiss a potential problem with the PSU+GPU combination without a reason.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
My pc booted completely fine this morning like nothing happened nothing out of the ordinary.
Problem is not repeatable. did not come back.
All I can think now is just like a cosmic ray strike or something? I'm super confused
 
Last edited:

DavidM012

Distinguished
Well it turns out I have lots of thoughts. I've read on this forum that a Corsair tx650m or rm series psu can cope with rtx 3060 transients. And a few others like Superflower.

Otherwise you have to 'overbuild' your PSU so that the transient Spike doesn't exceed it's capacity. Even then if it isn't a quality power supply with good capacitors it might not hold up for long. So if your CPU, Mobo and mem (core system) are using 200w (approx plus a bit of room for peripheral drives) and your gpu is using 170w and a transient power spike goes up to 550w

The mafs is 370w plus 550w-170w is 380w so if your GPU is pulling 550w for 100ms instead of 170w then your PSU is hitting it's max capacity for 100ms (750w). Since it's only estimating what the power draw could be and maybe guesstimating that the psu isn't quite delivering 750w and could be going over.

If this is occurring frequently though they say the transients occur under load well at least, it's bothering your power supply like an itch. It hasn't tripped it's overcurrent protection (yet) but, maybe the overcurrent protection isn't great and doesn't work quite as well as it's supposed to. So your symptoms aren't a fatal crash it's more like, veering close to the safety barrier so you get these slight symptoms.

The trouble with transient gremlins and power issues in general is you can't really see anything. You see symptoms but it's all about the quality and specifications of the power supply and electrical stuff that you need to know.

The simple version is just always buy a quality power supply and you usually don't have to worry and it just works. Or maybe you worry a bit to figure out that the Power Supply is modern to fit into modern systems. And then you stop worrying. Usually most of the time.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
I dont know it's been fine for a couple days and the past like 10+ pc restarts ive been doing while converting my system to GPT/NVME in preparation of upgrading to an i3 12100f system with a new board and ram ect.
No issues since that one time.
I hope its just a fluke random event or something.
One time a few weeks ago I was having issues with my bios settings not applying when "saving and exiting" one time. which required a bios reset to fix. so it may even just be an old crummy buggy bios issue.
 

DavidM012

Distinguished
Well if nothing else that's a very power efficient CPU so perhaps it'll take off some of the strain and you're getting away with a tier c power supply.

On the flip side the cpu will be able to work the gpu harder. You are only going close to the capacity of the power supply when it spikes for 100ms. You've unwittingly stepped into the role of stress testing your power supply since the GQ is of marginal quality.

The only way to really know is to try and find out. Time will tell if it holds up or not. If a PSU fizzles it can cause damage, or sometimes the pc components are fine and survive.

If you're worried about it then, err on the side of caution. Your system isn't highly demanding of power besides the gpu. If you have a low rated psu you don't have much room for upgrades even if it's a quality one. Overbuilding too much with a high rated quality power supply is a waste of money. The idea is to find the goldilocks zone of a tad of overbuild with a quality power supply so you might have capacity for 1 more cpu or gpu upgrade further down the line.

You don't have the luxury of testing various grades of power supply in your system. Next time you buy a psu, get one from further up the PSU tier list . If you plan to use a chunky gpu like the rtx 3000 series.
 
Last edited:

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
Ive tried all sorts of power stress tests like running furmark and prime95 small fft at the same time. cycling them on and off. and no issues. including having the vram being 90% used too. and nothing. no issues at all no instability.
even while my cpu is overclocked to 4.6ghz and my rams running at its full 2400mhz and even GPU power limited cranked up to +12% and everything.
in games my cpu is fine and only gets to about 40-60C usually around the 45-55C range
gpu a bit warmer at 65-68C
 

DavidM012

Distinguished
I won't ask how for long. :) The facts are that the GQ is tier c. The gpu can generate transients.

Either way, to move out of the margins and stop worrying, or if it fizzles, you'd have to buy a new power supply. You're stuck with your GQ for the moment. See how it goes, since the rest of your system isn't highly demanding of power.

No need to stress it more if you've done it already & if it's working for gaming, OK. If it pops or wears out sooner than expected, you might know why. Stress tests are only doing what others have done before already.

It would be nice to have a data pool like the psu tier list, of combinations of hardware. Since there's no reference data the only point of a stress test is to prove your hardware is working 'at the moment'. If in 10 years or so the capacitors on the mobo had worn out or something, it would be unable to run a stress test. Well if there was a lightning strike and a power surge or a natural disaster it wouldn't be able to complete a stress test. The point is that stress tests are useful if done scientifically and the data is available for use.


If all it's going to do is push your system 'til you find out it's failure point- this does not seem like something to aim for. You don't want to buy a new power supply. OK - so don't attempt to crash it or drive it into the ground with a stress test, then!

The 3060 uses less power than a 3070, 80 or 90. The 90 produces the biggest transient spike. It might've been the source of the bug. Might've been a disk error, sometimes happens, occasional glitches, ok. Could've been several things. Science can't know exactly right now.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
Well i'm not using that old motherboard anymore now.
Got a 12100F CPU and H610 motherboard for it along with some DDR4 32GB ram 3200mhz. CL16
2 sticks of 16gb
so far pc running smoothly. though I did have 1 freeze/lockup in the bios when going through the bios settings where my cursor vanished and i had to shut off the pc with psu switch
but after that no problems so far. very weird.
also pc refuse to boot off of sata ssd anymore. winload.efi error corrupt or missing.
but my cloned NVME drive works fine no problems. boots just fine. which is what I kind of wanted. but its super strange that the sata SSD doesnt wanna boot anymore. but the nvme boots fine.
 

DavidM012

Distinguished
Only glitches. winload.efi missing is generally attributed to disk corruption. However, you could simply blat the partition on the ssd and should be able to use it normally as a secondary disk (at least 'til the warranty expires, or usage conditions are exceeded), now you've cloned it.

Probably something to do with having converted to gpt to prepare for the upgrade and then switching from an older board with a legacy bios to a modern board with a uefi bios. Article about (legacy) compatibility support mode.

Technically the bios communicates with the drives so, maybe it had a momentary brain fart. You couldn't strictly rule out that something could possibly have happened in the depths of the system.

Perhaps, had you opted for a clean install of Windows instead of cloning, you may not have noticed anything amiss.

A clean install of windows is generally recommended when upgrading, on this forum. It's basically a question of methodology. The entire reason for the procedure is to avoid odd glitches and bugs.

Well you seem ok for the moment. If you get any more glitches consider a clean install.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
Oops secure boot was disabled in bios. thats probably why it derped on my sata drive

My nvme drive is fine. which is what I cloned my sata ssd onto. and thats fine.
I didn't clone the nvme to the sata drive.
I cloned my sata drive to the nvme.
the old sata ssd which i was using with the previous pc to boot from doesnt want to work for unknown reasons. could be boot order. sata port order. or from secure boot not being enabled properly in the bios. and so it freaked out not knowing what to do.
but i know 100% it was not corrupted because it was working 100% fine just an hour before i took it out and put it into the new motherboard/pc
it still works all the data is there.

My old bios from my 3770k pc WAS in uefi secure boot mode. so not legacy or csm.
so it may of got the error BECAUSE my new system wasn't initially in secure boot mode for whatever strange reason.
both drives are gpt and uefi. cause thats how I cloned it in the first place.
the nvme drive for some reason didnt care if secure boot was on or off. and booted anyways regardless. which made me even more confused.
and it took a while to find out that secure boot wasnt even enabled at first.
I also had to go dig through and set secure boot mode to enabled and STANDARD cause it was set to custom before and who knows.

I can't do any clean installs right now its complicated.
I can't use my phone. so I can't log in to 90% of my stuff. microsoft account included. without the 2FA code. and I can't get the code cause my phone has no service anymore. its too old.

My phones dead/outdated unsupported. no more service anymore. I need to upgrade to a new phone on the new service. and keep the same phone number. then I wont have to worry about that.
 

realflow100

Distinguished
Feb 22, 2013
108
3
18,585
I guess it was just a random one-off cosmic ray or something. who knows. My pc has been working fine so far.
along with the new motherboard cpu and ram. no issues so far. all working totally fine.
though any bios updates resets/clears makes me have to screw with the boot order a lot to find out which one is my nvme boot partition in order to boot windows properly from the nvme drive.
bios is updated to version 1.8