Overclock.net banner

7101 - 7120 of 16635 Posts

·
Registered
Joined
·
5 Posts
Bought a X570 I Aorus Pro Wifi, 3700X and Corsair LPX Vegeance 32GB(2 x 16GB) 3200mhz (CMK32GX4M2D3200C16).



Getting a lot of BSODS during installation and booting into desktop with Windows.

BSODs include:
- System Service Exception
- Irql_not_less_or_equal
- APC Index Mismatch
- Kernel Security Check Failure
- Kernel Auto Boost Invalid Lock Release
- Resource Not Owned (most frequent)


Ran memtest86 on a USB stick and it completed overnight with no issues. F5 Bios and CMOS reset so far seems to be the stablest, but still crashing whilst doing nothing on the desktop after installing the latest AMD chipset driver. The latest BIOS seem to cause more unstability and crashing on booting.


Was going to go through the BIOS options and see if there's any stable with default settings. (F3, F4, F5n, F6bn, F10, F11, F12e). I'm running out of ideas of what to do. The store offered to take them back and test them, but I'm scared that they won't have the same issues I have or can't replicate them. It's also busy for them and would take some time.
 

·
Old crazy guy
Joined
·
3,157 Posts
Bought a X570 I Aorus Pro Wifi, 3700X and Corsair LPX Vegeance 32GB(2 x 16GB) 3200mhz (CMK32GX4M2D3200C16).



Getting a lot of BSODS during installation and booting into desktop with Windows.

BSODs include:
- System Service Exception
- Irql_not_less_or_equal
- APC Index Mismatch
- Kernel Security Check Failure
- Kernel Auto Boost Invalid Lock Release
- Resource Not Owned (most frequent)


Ran memtest86 on a USB stick and it completed overnight with no issues. F5 Bios and CMOS reset so far seems to be the stablest, but still crashing whilst doing nothing on the desktop after installing the latest AMD chipset driver. The latest BIOS seem to cause more unstability and crashing on booting.


Was going to go through the BIOS options and see if there's any stable with default settings. (F3, F4, F5n, F6bn, F10, F11, F12e). I'm running out of ideas of what to do. The store offered to take them back and test them, but I'm scared that they won't have the same issues I have or can't replicate them. It's also busy for them and would take some time.
This looks pretty bad... did you check what's the vcore voltage during all core workload like CB20 multi thread bench?
You could try a + or - offset of 0.05 volt.

Is Windows a fresh install?
Otherwise try a new install on a different disk or on a USB stick with:
https://www.easyuefi.com/wintousb/
 

·
Registered
Joined
·
276 Posts
Yep running them at 100% usually works but can also be that they are not actually running at full speed...

From the Noctua whitepaper:



"1.The trace from PWM output to the fan header must not have a pull up or pull down. The pull up is located in the fan hub. The presence of a pull up on the motherboard will alter the fan response to the PWM Duty Cycle. In some cases this may prevent the fan from achieving full speed even with the Hardware Monitor device issuing a 100% duty cycle."



Indeed another issue with the AMD drivers is many storage vendors utilities doesn't work, forgot about that.

I remember also some weirdness in random performances.

Just use the Microsoft drivers to avoid issues; except maybe some very specific edge cases they can only be better.
I don't run then at 100%, just when i find where they stop, lets say at 700rpm, i set minimum 800rpm. With this i never experience any kind of trouble, just don't stop them and everything will be fine because they suck to much power when they start and i suppose some motherboards don't like it.
 

·
Registered
Joined
·
286 Posts
What's the issue exactly?

My Noctua fans work fine
The problem is, for some reason, the fans randomly stop working. This is an issue with the custom PWM controller used by Noctua, and a long time ago Gigabyte supposedly was working with them to get it fixed.
 

·
Old crazy guy
Joined
·
3,157 Posts
The problem is, for some reason, the fans randomly stop working. This is an issue with the custom PWM controller used by Noctua, and a long time ago Gigabyte supposedly was working with them to get it fixed.
What do you mean randomly?
I thought it was just when going down to the lowest RPM for the profile.
Do they stop suddenly even when running at high RPM?

If so, doesn't look like the issue afflicting other people.

Did you try using Argus Monitor to see exactly at which value they stop?
It has a 30 days trial.
 

·
Registered
Joined
·
221 Posts
Hi all just had a quick check for my fans

all noctua Black versions

Noctua NF-A14 PWM chromax rated at 1500rpm full speed. I get 1480rpm so that seems to be fine at full speed connected to CPU_OPT, fan control use temp input set to system 1

heres the strange thing

Noctua NF-A15 HS-PWM chromax rated at 1500rpm full speed, I get only 1300rpm and that is set to full speed connected to CPU, fan control use temp input set to CPU. so thats 200rpm under speced

and even more strange is this

Noctua NF-S12A PWM chromax rated at 1200rpm full speed, I get 1510rpm set to full speed connected to SYS_Fan1, fan control use temp input set to system 1. I get 310rpm over the rated fan speed ???? which is very strange.

so one fan is correct the other fan is under and the other fan is over the rated speeds, i dont understand that.

the other questions that i would like to know what do you all Have your CPU_OPT Control mode set to ? mine is set to auto and noticed you can have it voltage or PWM. I am not to sure what Auto selects as default. could this cause the fans rampping up for a couple of sec when i get cpu spikes ?

and waht is CPU_opt fan stop ?

Are the fan headers i am using correct I know two of the fasn are but not to sure about the CPU_OPT as i set this to system fan.



thanks
 

·
Registered
Joined
·
286 Posts
What do you mean randomly?
I thought it was just when going down to the lowest RPM for the profile.
Do they stop suddenly even when running at high RPM?

If so, doesn't look like the issue afflicting other people.

Did you try using Argus Monitor to see exactly at which value they stop?
It has a 30 days trial.
I don't know when exactly it happened, and I'm currently not using them. As far as I remember the problem was acknowledged by @GBT-MatthewH, and was no mention of it getting fixed since then.

Edit: Here is the last relevant reply I can find: https://www.overclock.net/forum/11-...570-aorus-owners-thread-266.html#post28147692
 

·
Registered
Joined
·
485 Posts
The problem is, for some reason, the fans randomly stop working. This is an issue with the custom PWM controller used by Noctua, and a long time ago Gigabyte supposedly was working with them to get it fixed.
So the problem is Noctua's fan design not complying with Intel's PWM spec?

If so, fault should be on them for selling fans not fit for purpose. Advertised as PWM yet not compliant with its spec.
 

·
Registered
Joined
·
2 Posts
Cold boot DRAM issue fixed

I just wanted to share this for anyone experiencing the cold boot issue (DRAM causing POST to fail - resetting the BIOS in the process).

Assuming it hasn't been posted already, this guy's solution fixed it for me:

https://www.reddit.com/r/gigabyte/comments/ez4b1m/gigabyte_aorus_elite_x570_cold_boot_issue_solved/

Summary:

Set 'Power Loading' to 'Enable'.

You can find it in the following section:
Settings tab>Platform Power>Power Loading

I'm currently on BIOS F12E, Aorus X570 Ultra, 16gb Aorus 3600mhz Ram with 'Memory Boost' enabled, 3700X with PBO.

It doesn't appear to work for everyone, but fixed mine.

Apologies if this has been posted before.
 

·
Registered
Joined
·
794 Posts
Bought a X570 I Aorus Pro Wifi, 3700X and Corsair LPX Vegeance 32GB(2 x 16GB) 3200mhz (CMK32GX4M2D3200C16).



Getting a lot of BSODS during installation and booting into desktop with Windows.

BSODs include:
- System Service Exception
- Irql_not_less_or_equal
- APC Index Mismatch
- Kernel Security Check Failure
- Kernel Auto Boost Invalid Lock Release
- Resource Not Owned (most frequent)


Ran memtest86 on a USB stick and it completed overnight with no issues. F5 Bios and CMOS reset so far seems to be the stablest, but still crashing whilst doing nothing on the desktop after installing the latest AMD chipset driver. The latest BIOS seem to cause more unstability and crashing on booting.


Was going to go through the BIOS options and see if there's any stable with default settings. (F3, F4, F5n, F6bn, F10, F11, F12e). I'm running out of ideas of what to do. The store offered to take them back and test them, but I'm scared that they won't have the same issues I have or can't replicate them. It's also busy for them and would take some time.
Have you tried mem test hcl to test inside Windows to ensure RAM settings are 100% stable? Mem test86 is great for catching hardware defects in RAM while mem test hcl is better at catching stability errors.

Also be sure you are using slots A2 and B2 (2nd and 4th slot going left to right starting at CPU)

Link to memtest hcl : https://hcidesign.com/memtest/

also try aida64 stability test. I had an issue where 1 dimm was faulty but only presented in AIDA64 stability test (stopped very soon after starting test). Turns out the dimm would throw errors when above 1.2volts but passed memtest 86 with flying colors. Was also coincidentally a Corsair Vengeance LPX kit. Had a hell of a time creating a Corsair account in order to submit a ticket to get a RMA.

Most of bsods, crashing on Ryzen systems comes from memory instability or faulty Dimms that's why focusing so hard on your RAM. Does it fail if at bios safe defaults? Be sure to manually set DDR Voltage to 1.35 or 1.37 as there was a bug that left it at 1.2 when XMP was loaded

Also try Blue Screen View here: https://www.nirsoft.net/utils/blue_screen_view.html this tool will help you diagnose exact cause of BSOD and allow easy Googling to resolve if not memory related

EDIT: Did you do a fresh install of Win 10 or just reuse OS drive from old build? Best practice is do do a fresh install when changing out major system components to prevent random issues.
 

·
Old crazy guy
Joined
·
3,157 Posts
I don't know when exactly it happened, and I'm currently not using them. As far as I remember the problem was acknowledged by @GBT-MatthewH, and was no mention of it getting fixed since then.

Edit: Here is the last relevant reply I can find: https://www.overclock.net/forum/11-...570-aorus-owners-thread-266.html#post28147692
Yes they are talking about the power surge spike when they start.
Aqua computers is working on a solution but they still haven't got one.
It's a tricky issue because depending on the Super IO controller used and how it's integrated in the board a fix could be or could be not possible.
Mitigating the issue for the Noctua fans could introduce issues to other fans which are conforming to Intel spec.

So the problem is Noctua's fan design not complying with Intel's PWM spec?

If so, fault should be on them for selling fans not fit for purpose. Advertised as PWM yet not compliant with its spec.
There is a spec but it's not mandatory; it's good practice to be compliant but if Noctua and EK are not, well there problem is yours sadly.
PWM is a generic term, nothing you can hang on.
That's the reason why for HDMI, Bluetooth, etc you have to make a contract that enforces you to comply to the specs and provide a certification.
Incidentally you end up very often in the same issues but that's another story...

Noctua seems mostly available to swap with older models.
But there's no legal ground to force them to do so; if you are not satisfied and they don't want to, it's a boomer.
If you can't get a swap or refund from the reseller, you have to take the toll and avoid any new purchase from them.

Hi all just had a quick check for my fans

all noctua Black versions

Noctua NF-A14 PWM chromax rated at 1500rpm full speed. I get 1480rpm so that seems to be fine at full speed connected to CPU_OPT, fan control use temp input set to system 1

heres the strange thing

Noctua NF-A15 HS-PWM chromax rated at 1500rpm full speed, I get only 1300rpm and that is set to full speed connected to CPU, fan control use temp input set to CPU. so thats 200rpm under speced

and even more strange is this

Noctua NF-S12A PWM chromax rated at 1200rpm full speed, I get 1510rpm set to full speed connected to SYS_Fan1, fan control use temp input set to system 1. I get 310rpm over the rated fan speed ???? which is very strange.

so one fan is correct the other fan is under and the other fan is over the rated speeds, i dont understand that.

the other questions that i would like to know what do you all Have your CPU_OPT Control mode set to ? mine is set to auto and noticed you can have it voltage or PWM. I am not to sure what Auto selects as default. could this cause the fans rampping up for a couple of sec when i get cpu spikes ?

and waht is CPU_opt fan stop ?

Are the fan headers i am using correct I know two of the fasn are but not to sure about the CPU_OPT as i set this to system fan.



thanks
I'd avoid using CPU_OPT if it's not a CPU Heatsink fan.

If it's on the CPU use the CPU temperature as input sensor, not System1.
If you know the fan attached to the output is PWM set control mode to PWM just to be sure; usually if it's not working in Auto the fan either doesn't start or runs always full speed.

Fan stop means that below a trigger temperature, you should see a vertical bar on the graph, the fan fully stops instead of running at the lowest RPM.
I'd highly recommend you NOT to enable it with your fans!
To be on the safe side you should also use the Normal profile instead of Silent.

These are the problems you can get into with their special PWM controller sadly.
You probably have the fans reporting a lower and higher speed not running at full speed.
Don't worry about having it running above 100%, it's not; it's just wrong the reporting.
 

·
Registered
Joined
·
2 Posts
I just wanted to share this for anyone experiencing the cold boot issue (DRAM causing POST to fail - resetting the BIOS in the process).

Assuming it hasn't been posted already, this guy's solution fixed it for me:

https://www.reddit.com/r/gigabyte/comments/ez4b1m/gigabyte_aorus_elite_x570_cold_boot_issue_solved/

Summary:

Set 'Power Loading' to 'Enable'.

You can find it in the following section:
Settings tab>Platform Power>Power Loading

I'm currently on BIOS F12E, Aorus X570 Ultra, 16gb Aorus 3600mhz Ram with 'Memory Boost' enabled, 3700X with PBO.

It doesn't appear to work for everyone, but fixed mine.

Apologies if this has been posted before.
Just to add to this, it only works with F12E. I tried it in F11 and F10 but the issue persists.
 

·
Registered
Joined
·
221 Posts
Yes they are talking about the power surge spike when they start.
Aqua computers is working on a solution but they still haven't got one.
It's a tricky issue because depending on the Super IO controller used and how it's integrated in the board a fix could be or could be not possible.
Mitigating the issue for the Noctua fans could introduce issues to other fans which are conforming to Intel spec.



There is a spec but it's not mandatory; it's good practice to be compliant but if Noctua and EK are not, well there problem is yours sadly.
PWM is a generic term, nothing you can hang on.
That's the reason why for HDMI, Bluetooth, etc you have to make a contract that enforces you to comply to the specs and provide a certification.
Incidentally you end up very often in the same issues but that's another story...

Noctua seems mostly available to swap with older models.
But there's no legal ground to force them to do so; if you are not satisfied and they don't want to, it's a boomer.
If you can't get a swap or refund from the reseller, you have to take the toll and avoid any new purchase from them.



I'd avoid using CPU_OPT if it's not a CPU Heatsink fan.

If it's on the CPU use the CPU temperature as input sensor, not System1.
If you know the fan attached to the output is PWM set control mode to PWM just to be sure; usually if it's not working in Auto the fan either doesn't start or runs always full speed.

Fan stop means that below a trigger temperature, you should see a vertical bar on the graph, the fan fully stops instead of running at the lowest RPM.
I'd highly recommend you NOT to enable it with your fans!
To be on the safe side you should also use the Normal profile instead of Silent.

These are the problems you can get into with their special PWM controller sadly.
You probably have the fans reporting a lower and higher speed not running at full speed.
Don't worry about having it running above 100%, it's not; it's just wrong the reporting.
I use the CPU fan header for the CPU and the CPU_opt set to system 1 as its for a case fan as its next to the other case fan so it would be

CPU fan header as a cpu fan set to custom as i need to stop the fan spikes
fan control use temp input set to CPU (CPU fan is conntected to this header)

CPU_OPT header used as a case fan set to normal
fan control use temp input set to system 1 from CPU (as this is no longer a cpu fan as its a case fan hence the reason i set it to system 1)

SYS_Fan1 header used as a 2nd case fan set to normal
fan control use temp input set to system 1

the problem is the sys_fan2 is at the other end of the motherboard and theres no way it will reach.

but the cpu_opt seems that you can alter it to a system 1 fan by using the fan control use temp input option and this stops it acting like a cpu fan so thee is no speed ramp ups when you get a cpu spike and keeps the fan at the same speed all the time just like a system fan.
 

·
Old crazy guy
Joined
·
3,157 Posts
I use the CPU fan header for the CPU and the CPU_opt set to system 1 as its for a case fan as its next to the other case fan so it would be

CPU fan header as a cpu fan set to custom as i need to stop the fan spikes
fan control use temp input set to CPU (CPU fan is conntected to this header)

CPU_OPT header used as a case fan set to normal
fan control use temp input set to system 1 from CPU (as this is no longer a cpu fan as its a case fan hence the reason i set it to system 1)

SYS_Fan1 header used as a 2nd case fan set to normal
fan control use temp input set to system 1

the problem is the sys_fan2 is at the other end of the motherboard and theres no way it will reach.

but the cpu_opt seems that you can alter it to a system 1 fan by using the fan control use temp input option and this stops it acting like a cpu fan so thee is no speed ramp ups when you get a cpu spike and keeps the fan at the same speed all the time just like a system fan.
Sorry I wasn't clear and didn't consider you were already out of free headers :)

Indeed you should be able to use it as any other header setting the temperature control to something else than the CPU.
I usually avoid using the CPU_OPT header unless it's really needed; there are 2 reasons for that.

First, often I've seen odd behaviors due to some boards considering it as an actual CPU fan. It shouldn't happen on this board but you never know...
Second reason it's that usually headers close to each other share some circuitry.
You have already issues with the Noctua fans and there's a chance that a power draw spike shuts off both the CPU_FAN and CPU_OPT.

I'm quite surprised you don't have plenty of 4-pin extenders, there's usually at least one together with the LNA and/or ULNA cables for each fan in the Noctua packaging.
If you don't have one, I'd suggest you buy it on Amazon they are quite cheap. Then you could move it to the SYS_FAN2 header.

But it's just a precaution that I would get; maybe check if without the fan connected on CPU_OPT you can get the CPU fan at full speed.
Then could be it does matter.
 

·
Registered
Joined
·
276 Posts
I suppose i am lucky with my noctua fans. Most of them are with splitters. 3 front nf12 to fan4, 3 back 80mm(in the 2nd compartment of my Air740 case) to fan5, 1 x 3000rpm infront of my sas enclosure (2nd comp.) to fan2, 1 back 140 and 2 x 140 top to fan6, 140mm to cpu and 2nd 140mm to cpu_opt.

All start ok.
 

·
Registered
Joined
·
485 Posts
These are the problems you can get into with their special PWM controller sadly.
You probably have the fans reporting a lower and higher speed not running at full speed.
Don't worry about having it running above 100%, it's not; it's just wrong the reporting.
My NF-A14s run perfectly fine at any speed and correctly report their max speeds inline with that advertised by Noctua


It's not unusual for fans speeds to be out by around +-10% when calibrated by fan controllers and motherboards. It's not like Gigabyte or any board or fan controller vendor is going to go to the expense of putting high end ICs, sensors and so forth just to ensure fans read at bang on their real speed.

1450-1550rpm would be perfectly normal readings for something like a NF-A14

But there's no legal ground to force them to do so; if you are not satisfied and they don't want to, it's a boomer.
If you can't get a swap or refund from the reseller, you have to take the toll and avoid any new purchase from them.
Consumer laws differ obviously, but in my country I'd have easy recourse to return them as 'not fit for purpose' .
 

·
Old crazy guy
Joined
·
3,157 Posts
My NF-A14s run perfectly fine at any speed and correctly report their max speeds inline with that advertised by Noctua


It's not unusual for fans speeds to be out by around +-10% when calibrated by fan controllers and motherboards. It's not like Gigabyte or any board or fan controller vendor is going to go to the expense of putting high end ICs, sensors and so forth just to ensure fans read at bang on their real speed.

1450-1550rpm would be perfectly normal readings for something like a NF-A14



Consumer laws differ obviously, but in my country I'd have easy recourse to return them as 'not fit for purpose' .
I may have been lucky but I've never seen a 10% error at high RPM; it does happen often at low RPM but even that is usually well below 10%.
The 10% tolerance in Intel PWM specs is about the reported speed and the actual fan rotations.
This is very common, manufacturers "like" to cheat about performances and often the advertised and reported speed is actually much lower.

AFAIK the Tachometer output signal is coming from the fan, two pulses for revolution.
It doesn't need calibration as it's a PWM digital signal and purely depends on the ICs and circuitry of the fan.

Indeed there could be national laws that could make easier to return Noctua fans.
Under EU consumers laws it can't be considered a manufacturer defect and I guess in USA it's even worse.
 
7101 - 7120 of 16635 Posts
Top