Weird cold boot issue on Crosshair Hero X670E

[h4x0rm1k3]

Hi all, i've got a weird cold boot issue with my system which consists of a Crosshair Hero X670E, 64Gb 6000Mhz RGB RAM, RTX 4090 & R9 7900X. Basically whenever I shut down my system on a night and go to turn it on the next day most of the time it refuses to power on at all, even when pressing the power button on the motherboard itself! All the board RGB is on etc but it just refuses to power on until I remove the power lead from the PSU and leave it for 5 minutes or so, then replacing it and then it'll usually boot, sometimes it'll take a 2nd try. The other day though it was like that for about an hour but what I did notice was that the board RGB's had turned off apart from the power & flex key ones and what got it running eventually was to remove the 24 pin motherboard cable (I did that to bridge the 2 pins to check if the PSU powered up which it did) and plugged that back in, then the RGB's lit up again and it responded to the power button on the board as i'd removed the other ones while checking things over and was using a screwdriver to bridge the 2 pins to act like the power button.
The system itself isn't that old, I only brought the motherbord, CPU, GPU & RAM in April of this year, the rest was from my older system which was itself only 2-3 years old so the parts were fine still and to start with all was perfect, it's only been the past maybe 1-2 months i've had these issues and would ideally like to get to the bottom of it so wondered if any of you guys/girls had any ideas? I've got a 2nd hand motherboard on order and am going to switch that in the next time it happens and see if it continues so I can rule out the board and if it isn't the board then i'd be leaning towards the PSU even though it was a top end one of the time (ROG THOR 1200P) and has no other issues when it's running and playing games on the RTX4090 so at the moment i'm leaning towards it being the board itself which again, has no other issues apart from this cold boot problem!

 

[ir_cow]

Hmmm it kinda seems like your PSU is preventing a short. It is shorting on something and unplugging it clear the protection.

Does any lights come on the Motherboard or is it dead inside until you unplug completely?

Last time this happened to me was I used a old SATA power cable with a new PSU. They leads are different at the PSU causing a big short. Didn't figure it out until the drive was dead.

 

[eidairaman1]

Breadboard your system and see if it improves, switch to a different power strip or a different circuit in you domicile, heck use your restart switch as your power switch. Try above and even a different power supply. Try system in a different domicile with above techniques.

Make sure your motherboard bios is up to latest.

 

[Redwoodz]

Agree with the above posts, something is shorting. Whether it is an external lead from the mobo, the motherboard itself or the PSU. Make sure all connectorrs are seating properly( look at each pin inside the connectors) especially to the gpu.

 

[h4x0rm1k3]

See, I would think it'd be more the PSU if it wasn't for the fact that in normal every day use, restarts, reboots etc are all perfect, no issues at all, it's only when I shut down the PC at night to go to bed and then go to turn it back on about 8 hours later that the issue presents itself, it's not every day either and the PC doesn't move 'cos it's too damn heavy as it's fully kitted out. I moved about 2 weeks ago and the problem was still present before then so it's nothing to do with household wiring, in fact the wiring in the new place is better than the old 'cos I don't get audio drop outs from my AVR when something gets switched on/off or unplugged etc from a socket anywhere in the house. The motherboard itself has signs of life as the RGB plate lights up when it's in the off state, it just refuses to respond to power button, restart button, or motherboard power button presses, it won't even start by bridging the 2 pins in the lower right of the board. When I disconnected the 24 pin connector from the motherboard and bridged the 2 pins on that the PSU powers up fine too, it was when I plugged that connector back in to the board last time that the PC fired up on the press of the power button. I've got a spare motherboard on the way to me so that's what i'll be trying next if I don't narrow it down before then and then i'll move on to the PSU next if I need to.

 

[eidairaman1]

See, I would think it'd be more the PSU if it wasn't for the fact that in normal every day use, restarts, reboots etc are all perfect, no issues at all, it's only when I shut down the PC at night to go to bed and then go to turn it back on about 8 hours later that the issue presents itself, it's not every day either and the PC doesn't move 'cos it's too damn heavy as it's fully kitted out. I moved about 2 weeks ago and the problem was still present before then so it's nothing to do with household wiring, in fact the wiring in the new place is better than the old 'cos I don't get audio drop outs from my AVR when something gets switched on/off or unplugged etc from a socket anywhere in the house. The motherboard itself has signs of life as the RGB plate lights up when it's in the off state, it just refuses to respond to power button, restart button, or motherboard power button presses, it won't even start by bridging the 2 pins in the lower right of the board. When I disconnected the 24 pin connector from the motherboard and bridged the 2 pins on that the PSU powers up fine too, it was when I plugged that connector back in to the board last time that the PC fired up on the press of the power button. I've got a spare motherboard on the way to me so that's what i'll be trying next if I don't narrow it down before then and then i'll move on to the PSU next if I need to.

Id suggest the psu from what you just wrote.

Also if the system is off the restart switch isn't going to do anything. What i was getting at is disconnect the power switch from the mobo front panel header and use your restart switch on the power up pins on the mobo front panel header.

 

[Bill_Bright]

See, I would think it'd be more the PSU if it wasn't for the fact that in normal every day use, restarts, reboots etc are all perfect, no issues at all

Since EVERYTHING inside the computer case depends on good, clean, stable power, you still need to eliminate the PSU as the possible cause. Therefore, you need to swap in a spare (borrowed if necessary) PSU and see what happens. I sure would try that BEFORE swapping motherboards.

Remember, the front panel power button, as well as any motherboard power buttons are simply remote switches that short two pins on that motherboard. Shorting those pins causes the +5Vsb standby voltage to signal the PSU and motherboard to start. The fact the system runs fine after it starts does NOT eliminate the PSU from the equation because during normal operation, the computer is using the standard +12V, +5V and +3.3V outputs of the supply - not the +5Vsb standby voltage. That is a different +5V.

For the record, while I understand by "cold", you mean temperature, that is not what a "cold" boot really is. A "cold" boot is when you start from a state that had no voltage applied. That is when the computer is totally "off" as in unplugged from the wall or the master power switch on the back of the PSU is set to "Off" or "0".

The ATX Form Factor standard requires all ATX PC PSUs to supply +5Vsb standby voltage to multiple points on the motherboard including USB ports, RAM and the case's front panel power button - unless you also unplug from the wall or flip that master power switch. When you shutdown your system at night using the Start Menu Shutdown button or the case's front panel switch, you really are just putting the computer into "standby". That +5Vsb voltage is still present at multiple points on the motherboard including USB ports, Ethernet port, RAM and the case's front panel power button. This +5Vsb is keeping several circuits "alive", but in a low power, stand by state.

When you unplug the power lead from the PSU, you are removing that +5Vsb. When you reconnect it, you restore that +5Vsb. It sounds like that action is resetting a missing or corrupt setting and that is why it then works. Where that setting is or how it is being set is the issue and once again, swapping in a spare PSU will tell you if that is in the PSU, or something connected to the PSU.

 

[ShrimpBrime]

If you power the system on and don't hear the relay inside the PSU, the relay is going bad. RMA the PSU.

 

[Bill_Bright]

If you power the system on and don't hear the relay inside the PSU

Do you have actual experience with the OP's specific power supply? Do note that many, if not most PSUs do NOT use mechanical relays anymore - haven't for years. So you wouldn't hear the relay click.

 

[ShrimpBrime]

Do you have actual experience with the OP's specific power supply? Do note that many, if not most PSUs do NOT use mechanical relays anymore - haven't for years. So you wouldn't hear the relay click.

I do not have that PSU. I do know Corsair PSU's mostly use mechanical relays and they even say so. Maybe this link is really old or something.

Why does my PSU make a *click* sound when I turn on my PC? | CORSAIR

www.corsair.com

Also mechanical relays have very low resistance. So are mostly used on PC PSUs that I am aware of.
And every PSU I've ever purchased has had a mechanical relay.

Now do you have any evidence to support your claim "most PSU's do NOT use mechanical relays anymore - haven't for years. ??
Not saying THOR 1200W does or does not have a relay, because I didn't do that research. I'll let you do it instead.

 

[h4x0rm1k3]

I wondered if it was the PSU at 1st but I went with the motherboard 1st purely because when I remove the 24 pin motherboard connector and bridge the 2 pins the PSU has no issue firing up straight away and then when I connect it all back up it usually starts the PC 1st time so it's made me wonder if there was something on the motherboard that was causing the initial issue and removing the 24 pin from the system reset whatever the fault was. As for the relay, i'm not 100% sure but I 'think' I hear a click when I power up the system but that 'might' be the fans kicking in too but I will listen out for it next time and see if I can trace that back further. As far as i'm aware these Thor1200P units should use 1 of the Seasonic PSU's of equivalent spec, i'm sure I read that somewhere. As for returning it under RMA, have you ever tried returning an Asus product in the UK? Most of the time they don't want to know and say that they don't deal with certain products and to send it back to the seller which if in like most cases it was purchased from Amazon you also know what they're like for dealing with products that are over a certain age, even though the PSU supposedly has a 10 year warranty! The only thing i've managed to get an RMA for from Asus was an old ROG monitor but when I tried a motherboard in the past they straight out refused it so I can only guess the PSU will likely have the same outcome so i'm not hopeful there, as I'm also not for the CH Hero X670E if that is at fault at all as that would be a big loss to take being it's still fairly new too!

 

[Bill_Bright]

As I said, some do. Many, or even most, do not. That said, regardless if mechanical or electrical, most (especially high wattage) PSUs still implement some sort of delay to prevent issues, such as inrush current problems.

Also mechanical relays have very low resistance.

A moot point. Electrical, time delay circuits don't have any resistance. They also don't wear out, they take up very little physical space, and there there is no risk of arcing across the contacts that may result in carbon buildup on the contacts - which is never good.

The advantage of mechanical relays is they are cheap to make. I am not saying they are of lessor reliability - that depends on the specific design and build of the relay. I am just saying they are inexpensive. You can come to your own conclusions about Corsair and why they may use them.

I'll let you do it instead.

No need to (don't know the users PSU anyway). 50+ years as a certified electronics technician and nearly as long as a computer tech is all I need. You can follow the link in my sig to see for yourself if that qualifies me as having just a little expertise in this area.

The type delay used is immaterial. What is important for this discussion and for the OP specifically is that "mechanical" relays (the type that produce audible clicks) are not always used in PSUs. Since we don't know specifically what PSU the OP is using, or if it uses a mechanical or electronic relay, I don't want the OP to jump to unfounded conclusions if no click is heard.

What needs to happen is the OP needs to try another PSU and see if the problem persists, or goes away.

 

[ShrimpBrime]

As I said, some do. Many, or even most, do not. That said, regardless if mechanical or electrical, most (especially high wattage) PSUs still implement some sort of delay to prevent issues, such as inrush current problems.

A moot point. Electrical, time delay circuits don't have any resistance. They also don't wear out, they take up very little physical space, and there there is no risk of arcing across the contacts that may result in carbon buildup on the contacts - which is never good.

The advantage of mechanical relays is they are cheap to make. I am not saying they are of lessor reliability - that depends on the specific design and build of the relay. I am just saying they are inexpensive. You can come to your own conclusions about Corsair and why they may use them.


No need to (don't know the users PSU anyway). 50+ years as a certified electronics technician and nearly as long as a computer tech is all I need. You can follow the link in my sig to see for yourself if that qualifies me as having just a little expertise in this area.

The type delay used is immaterial. What is important for this discussion and for the OP specifically is that "mechanical" relays (the type that produce audible clicks) are not always used in PSUs. Since we don't know specifically what PSU the OP is using, or if it uses a mechanical or electronic relay, I don't want the OP to jump to unfounded conclusions if no click is heard.

What needs to happen is the OP needs to try another PSU and see if the problem persists, or goes away.

OP is using the ROG THOR 1200W Power Supply. It is likely a Seasonic Unit, which makes PSU's for Asus. I believe Seasonic also uses inrush current bypass relays.

""most PSU's do NOT use mechanical relays anymore - haven't for years.""

I'll ask nicer?? Please can you supply something to support this claim?

 

[h4x0rm1k3]

OP is using the ROG THOR 1200W Power Supply. It is likely a Seasonic Unit, which makes PSU's for Asus. I believe Seasonic also uses inrush current bypass relays.


I'll ask nicer?? Please can you supply something to support this claim?

Yes that's correct, ROG Thor 1200P V1, RTX 4090, Crosshair Hero X670E-E, 2x32Gb Corsair vengeance 6000Mhz DDR5 + R9 7900X is the full specs along with a couple of NVMe's & SSD's and a full water cooling loop cooling both the CPU & GPU. The PSU will definitely be next on the list to try if the motherboard isn't the issue, I don't know anyone to borrow a PSU powerful enough to run my system so I'd need to buy a new 1 if that's the case.

 

[eidairaman1]

Yes that's correct, ROG Thor 1200P V1, RTX 4090, Crosshair Hero X670E-E, 2x32Gb Corsair vengeance 6000Mhz DDR5 + R9 7900X is the full specs along with a couple of NVMe's & SSD's and a full water cooling loop cooling both the CPU & GPU. The PSU will definitely be next on the list to try if the motherboard isn't the issue, I don't know anyone to borrow a PSU powerful enough to run my system so I'd need to buy a new 1 if that's the case.

Not even a shop?

 

[ShrimpBrime]

Yes that's correct, ROG Thor 1200P V1, RTX 4090, Crosshair Hero X670E-E, 2x32Gb Corsair vengeance 6000Mhz DDR5 + R9 7900X is the full specs along with a couple of NVMe's & SSD's and a full water cooling loop cooling both the CPU & GPU. The PSU will definitely be next on the list to try if the motherboard isn't the issue, I don't know anyone to borrow a PSU powerful enough to run my system so I'd need to buy a new 1 if that's the case.

Relays are funny. They just go bad, it happens. Had an M2 Freightliner box truck with ABS module that wasn't communicating. The kid was telling me the fuse was good. Yeah well, check the relay bud, sure enough bad relay.

The switch on the board is easy to diagnose cause you have 2 of them. This signal is across 1 wire on the 24 pin connector to the PSU. To test the PSU outside the PC, you would jump the green wire to a black wire. Here's a picture of the pinout. Green is PS_on. Can use a paper clip or small piece of wire.

 

[eidairaman1]

Relays are funny. They just go bad, it happens. Had an M2 Freightliner box truck with ABS module that wasn't communicating. The kid was telling me the fuse was good. Yeah well, check the relay bud, sure enough bad relay.

The switch on the board is easy to diagnose cause you have 2 of them. This signal is across 1 wire on the 24 pin connector to the PSU. To test the PSU outside the PC, you would jump the green wire to a black wire. Here's a picture of the pinout. Green is PS_on. Can use a paper clip or small piece of wire.

View attachment 369542

Either the internal contactor welds itself in closed position, corrosion, contactor arm fails, or inductor winding gets an open or short

 

[h4x0rm1k3]

Relays are funny. They just go bad, it happens. Had an M2 Freightliner box truck with ABS module that wasn't communicating. The kid was telling me the fuse was good. Yeah well, check the relay bud, sure enough bad relay.

The switch on the board is easy to diagnose cause you have 2 of them. This signal is across 1 wire on the 24 pin connector to the PSU. To test the PSU outside the PC, you would jump the green wire to a black wire. Here's a picture of the pinout. Green is PS_on. Can use a paper clip or small piece of wire.

View attachment 369542

That's exactly what I did, I disconnected the 24 pin connector from the board and bridged the 2 pins and the PSU fired straight up which is why it's leading me to think it had/has something to do with the motherboard because as soon as I reconnected the PC fired straight up too so my thoughts are somethings causing a short or something similar to cause the board not to respond to any kind of power on which is cleared as soon as the PSU is removed and then reconnected again, even though there was still 'some' power going to the system in the form of 12V+ for the CPU via the 2x8 pin connectors. Like I said, it doesn't happen when the system is just rebooted for an update or whatever, only when the system has been powered down for 8+ hours.

If it doesn't happen before this weekend again i'll have more time to swich the board over and then see how it goes, I should find out within the 1st couple of days if the problem has gone away or not.

Not even a shop?

No, no local shops that sell any computer gear, Cex don't sell them and that's the only PC parts shop in my area really. I've only just moved to the area but still kind of local to where I used to live-ish and know there's nothing like that out this way. Amazon will be my best bet really unless I find a decent offer somewhere else.

 

[eidairaman1]

That's exactly what I did, I disconnected the 24 pin connector from the board and bridged the 2 pins and the PSU fired straight up which is why it's leading me to think it had/has something to do with the motherboard because as soon as I reconnected the PC fired straight up too so my thoughts are somethings causing a short or something similar to cause the board not to respond to any kind of power on which is cleared as soon as the PSU is removed and then reconnected again, even though there was still 'some' power going to the system in the form of 12V+ for the CPU via the 2x8 pin connectors. Like I said, it doesn't happen when the system is just rebooted for an update or whatever, only when the system has been powered down for 8+ hours.

If it doesn't happen before this weekend again i'll have more time to swich the board over and then see how it goes, I should find out within the 1st couple of days if the problem has gone away or not.


No, no local shops that sell any computer gear, Cex don't sell them and that's the only PC parts shop in my area really. I've only just moved to the area but still kind of local to where I used to live-ish and know there's nothing like that out this way. Amazon will be my best bet really unless I find a decent offer somewhere else.

I was .entioning to test hardware at a shop so you werent having to buy and return hardware

 

[PenguinBelly]

Does S3 sleep function work well?

 

[ShrimpBrime]

That's exactly what I did, I disconnected the 24 pin connector from the board and bridged the 2 pins and the PSU fired straight up which is why it's leading me to think it had/has something to do with the motherboard because as soon as I reconnected the PC fired straight up too so my thoughts are somethings causing a short or something similar to cause the board not to respond to any kind of power on which is cleared as soon as the PSU is removed and then reconnected again, even though there was still 'some' power going to the system in the form of 12V+ for the CPU via the 2x8 pin connectors. Like I said, it doesn't happen when the system is just rebooted for an update or whatever, only when the system has been powered down for 8+ hours.

If it doesn't happen before this weekend again i'll have more time to swich the board over and then see how it goes, I should find out within the 1st couple of days if the problem has gone away or not.


No, no local shops that sell any computer gear, Cex don't sell them and that's the only PC parts shop in my area really. I've only just moved to the area but still kind of local to where I used to live-ish and know there's nothing like that out this way. Amazon will be my best bet really unless I find a decent offer somewhere else.

Man that's an interesting symptom. I missed one of your posts, sorry for that.

OK, so did you try with the case switch unplugged, then next morning (after 8 hours or more) just use the motherboard button? I wonder if the case button is janky, would utilize the same start circuit as the button on the board. I didn't read you tried this, and is the reason why I'm asking is all.

 

[h4x0rm1k3]

I was .entioning to test hardware at a shop so you werent having to buy and return hardware

Ah I get you, no definitely not then because the PC itself weighs about 30KG+ all in total so definitely not something I can just cart around as I don't drive etc.

Does S3 sleep function work well?

You mean sleep in general? It did once and then after that I couldn't get it to work as I kept waking up to a restarted PC so had the same issues I mentioned. The time it did work though it fired straight up into desktop with all apps I had open before.

Man that's an interesting symptom. I missed one of your posts, sorry for that.

OK, so did you try with the case switch unplugged, then next morning (after 8 hours or more) just use the motherboard button? I wonder if the case button is janky, would utilize the same start circuit as the button on the board. I didn't read you tried this, and is the reason why I'm asking is all.

Yes, it's currently unplugged still as I haven't turned the PC off in a couple of days now while i've been waiting for the new board incase it got to the point where it wouldn't turn on at all. With the button connected or not it has that issue, as does using the power button on the motherboard itself or shorting the 2 pins on the motherboard together to act like the power button had been pressed did nothing either. It was only when I disconnected the 24 pin and shorted those 2 pins that the PSU fired straight up so I removed the wire I used to short and plugged it in to the PB again and pressed the power button on the board and it fired straight up. Before that point i'd been messing for about an hour trying to get it working by removing the power lead from the PSU altogether for 10 mins which usually works but this time it didn't, I even left it disconnected from the mains for about half an hour at one point before I decided to test the PSU by shorting the 2 pins and then that's when things started going right again.

 

[AusWolf]

6000 MHz with 64 GB RAM seems a bit high. Have you tried lowering your RAM speed or running it at stock (without EXPO)?

I'm suspecting that your system is stuck in a memory training loop.

 

[h4x0rm1k3]

6000 MHz with 64 GB RAM seems a bit high. Have you tried lowering your RAM speed or running it at stock (without EXPO)?

I'm suspecting that your system is stuck in a memory training loop.

Yes, tried that, the RAM can run at up to 6400Mhz on this system, i've had it running stable at those speeds in the past. I lowered them a while back as I was having a couple of crashing issues to the EXPO rated speeds but yes, since this has been happening more regular I have tried reducing the speeds. In fact, i've updated the BIOS about 3 times so far, twice this month as there was a new BIOS a week or so back, if that and the issue still presents after the updates. During those updates the BIOS is completely wiped and back to stock settings which I then have to set up again for my system which at the moment is only a basic setup with settings I know run perfectly fine and should have no problem at all running. The PDE & MCR are both enabled in the RAM's BIOS settings also, I have to enable MCR after initial memory training else it takes 2-3 minutes sometimes to even display the BIOS screen when powered on due in part I think to the slow memory training each and every boot.

EDIT:- The board I got in replacement is the Asrock PG lightning X670E which I managed to grab for £80 because the PCIe 5 slot 'apparently' has some damage and it doesn't run at full speed according to the seller. When I inspected the slot in the pic & in person now though it doesn't look damaged apart from the shielding is very slightly bent, no signs of lifted pins on the board or damage in the socket so i'm wondering if he thought it was damaged because GPU's run at varying PCIe speeds depending on usage plus there is no PCIe 5.0 GPUs out yet either so i'm questioning his testing methods as I can't see how he could tell plus he said it works perfectly fine otherwise! I'm not bothered about that slot anyway if it is damaged, there's 2 other perfectly fine PCIe 4.0 slots below it I can use as I have my GPU horizontal on a riser cable anyway

 

[eidairaman1]

Ah I get you, no definitely not then because the PC itself weighs about 30KG+ all in total so definitely not something I can just cart around as I don't drive etc.

You mean sleep in general? It did once and then after that I couldn't get it to work as I kept waking up to a restarted PC so had the same issues I mentioned. The time it did work though it fired straight up into desktop with all apps I had open before.

Yes, it's currently unplugged still as I haven't turned the PC off in a couple of days now while i've been waiting for the new board incase it got to the point where it wouldn't turn on at all. With the button connected or not it has that issue, as does using the power button on the motherboard itself or shorting the 2 pins on the motherboard together to act like the power button had been pressed did nothing either. It was only when I disconnected the 24 pin and shorted those 2 pins that the PSU fired straight up so I removed the wire I used to short and plugged it in to the PB again and pressed the power button on the board and it fired straight up. Before that point i'd been messing for about an hour trying to get it working by removing the power lead from the PSU altogether for 10 mins which usually works but this time it didn't, I even left it disconnected from the mains for about half an hour at one point before I decided to test the PSU by shorting the 2 pins and then that's when things started going right again.

So about 66 Pounds then, a rucksack basically

Something worth trying, cmos battery