Asus ROG CROSSHAIR VIII HERO (WI-FI) random shut downs/reboots, high temps/ voltages

[derridada]

My system shuts down randomly, and then just reboots. There is no bsod, no log, no dump, no diagnostics. There is no way to tell why this happens, but it seems as if the CPU hits the 95 C max and reboots for safety, but that is just my guess.
I’m running the latest chipset driver and BIOS.

Specs:
- AMD Ryzen 9 3900X on ROG CROSSHAIR VIII HERO (WI-FI); with latest AMD chipset driver and BIOS 0803
- GPU: ASUS ROG STRIX GeForce RTX 2080 O8G-GAMING
- RAM: G.Skill F4-3600C16D-16GTZNC DDR4-3600 on CL16-19-19-39 at 1.35V
- PSU: Seasonic Prime Ultra 1000W 80 Plus Titanium
- Cooling: CPU: Noctua NH-U12S, with 2 Noctua NF-F12 industrial PPC 2000 PWM fans
- OS Drive: Samsung 970 EVO NVMe M.2 1TB with Windows 10 Pro, version 1903

This happens both in Windows balanced / Ryzen balanced power plans. Temperatures are around 50-55C idling/low usage, which are on the high side. When idling/low usage, the CPU shows erratic voltage jumps, with a high, average ‘baseline’ of 1.464V. It does cycle below 1.0V, but only very sporadically, and very erratically.

The problem is that under load when gaming, the temps shoot up to 80 and beyond. Ultimately the system shuts down. I am running everything stock – no OC profiles.
The last time it happened was while playing Destiny 2 yesterday. Which is ironic, since the idling/power issue was supposedly patched with the latest AMD chipset driver and Asus BIOS.

I’ve cleared CMOS, flashed BIOS a few times with optimized defaults, reinstalling Win 10, uninstalling Armory Crate and disabling it right in the BIOS, disabling AURA.
I’m not using Asus Ai Suite, or anything like it; BIOS is stock, except for fan profiles.
It’s running on Power saver mode now, which is extremely stable, but not exactly blindingly fast.

[AngryAce]

In my experience, random reboots like your describing indicate a PSU issue. Under what conditions are you hitting 95c? It seems to me the Noctua CPU cooler should be doing a better job than that under gaming loads. I would redo the TIM and try to get those temps under control. If you're still having random reboots, try a different PSU.

[derridada]

I do hope it's not the PSU; just bought this Seasonic Prime Ultra 6 months ago, and was working fine in my previous build, an i-7 6700k on Asus 170 deluxe - and I do not have an older psu i can test, unfortunately. Though I will reapply the TIM - thanks for the tip.

[AngryAce]

derridada wrote:
I do hope it's not the PSU; just bought this Seasonic Prime Ultra 6 months ago, and was working fine in my previous build, an i-7 6700k on Asus 170 deluxe - and I do not have an older psu i can test, unfortunately. Though I will reapply the TIM - thanks for the tip.

Computer issues... can be anything and everything.. :D. It is possible when you hit TJ max the CPU crashes the system but it is suppose to throttle. I have the same motherboard but with a 3600x. I only get to 95c under prime95 (small FFT) and the CPU throttles. Gaming, its 55 to 60c with Corsair h115i. I realize I have 1/2 the cores of the 3900x but 95c gaming seems high to me. Let us know how it goes!

[derridada]

Well, it's not the thermal paste. If only it were this simple!
I reinstalled the heat sink and fans, this time using Noctua NT-H2 instead of my usual Kryonaut, but I'm not seeing any significant changes.

It just seems that the system never truly idles down; voltage remains at 1.46V most of the time, even when idling/low usage (it occasionally dips below 1.0V, but only for a second). This despite the fact that I'm running the latest chipset driver and BIOS, which were supposed to fix the idling issues.

I ordered a PSU just to exclude it; if it is the PSU I'll have to RMA and then return the new unit for a refund if this does indeed turn out to be the issue - I've already invested too much already.

Back to 'power saver mode' and 2.1 GHz for now...

[AngryAce]

derridada wrote:
Well, it's not the thermal paste. If only it were this simple!
I reinstalled the heat sink and fans, this time using Noctua NT-H2 instead of my usual Kryonaut, but I'm not seeing any significant changes.

It just seems that the system never truly idles down; voltage remains at 1.46V most of the time, even when idling/low usage (it occasionally dips below 1.0V, but only for a second). This despite the fact that I'm running the latest chipset driver and BIOS, which were supposed to fix the idling issues.

I ordered a PSU just to exclude it; if it is the PSU I'll have to RMA and then return the new unit for a refund if this does indeed turn out to be the issue - I've already invested too much already.

Back to 'power saver mode' and 2.1 GHz for now...

Mine never really clocks down either, even with the Ryzen Balanced power plan. With that said, hitting 95c and shutting down under gaming loads is simply not normal. Every review I've seen of the 3900x, the thermals are in the 65 to 75c range whether air or AIO cooled during gaming. Pauls hardware did a aida64 burn in test and his 3900x averaged 79c with Noctua NH-U12A on an open test bench.

I don't know your level of building experience but I assume you applied enough TIM with good coverage, removed any protective stickers on the CPU cooler block, cleaned both CPU and cooler block with IPA and have good air flow through your case.

While random shutdowns are usually the result of a faulty PSU, I am beginning to think you have a defective CPU if all those conditions above are met. The CPU should thermal throttle when it hits tjmax not shut down altogether. it is difficult to diagnose these types of problems without having extra gear to swap out and test with.

Perhaps others on this forums have some insight to help you! Keep us updated!!!

[derridada]

Turns out it was the PSU after all...

I was getting some high voltage and power readings in HWiNFO64, and decided to go ahead and order a different PSU thinking that I’d always be able to return it to Amazon within 30 days. Well, seems like my original unit, the Seasonic Prime Ultra Titanium 1000W, was way off :
- CPU core current: average: 20.89 A
- Core power: core 1 power average 1.12 W / core 2: 0.94 W / core 3: 4.34 W core 6: 1.80 W / core 9: 1.34 W etc.
- CPU core power average: 30.12 W
CPU die/package temps from 55-62, with a minimum of 48 with no apps running.

Compare this to the (temporary?) replacement PSU, Corsair AX850:
-CPU core current: average: 9.34 A
- Core power: core 1 power average 0.50 W / core 2: 0.27 W / core 3: 2.22 W core 6: 0.13 W / core 9: 0.10 W etc.
- CPU core power average: 13.08 W
CPU die/package temps from 45-50 on average load, 40-45 when idling/low load. *

So... I invested quite a bit in the*Seasonic, and I am rather disappointed. This is a 6-month old unit, and Titanium to boot. It’s possible the unit was already off in my previous i6700k build, but I never noticed anything on a lower core CPU.

Anyhow, I’m relieved as this seems the solution; still waiting to hear back from Seasonic to see what they can do.
As the temps/voltages have gone down, so has my blood pressure!

[Matos_]

derridada wrote:
Turns out it was the PSU after all...

I was getting some high voltage and power readings in HWiNFO64, and decided to go ahead and order a different PSU thinking that I’d always be able to return it to Amazon within 30 days. Well, seems like my original unit, the Seasonic Prime Ultra Titanium 1000W, was way off :
- CPU core current: average: 20.89 A
- Core power: core 1 power average 1.12 W / core 2: 0.94 W / core 3: 4.34 W core 6: 1.80 W / core 9: 1.34 W etc.
- CPU core power average: 30.12 W
CPU die/package temps from 55-62, with a minimum of 48 with no apps running.

Compare this to the (temporary?) replacement PSU, Corsair AX850:
-CPU core current: average: 9.34 A
- Core power: core 1 power average 0.50 W / core 2: 0.27 W / core 3: 2.22 W core 6: 0.13 W / core 9: 0.10 W etc.
- CPU core power average: 13.08 W
CPU die/package temps from 45-50 on average load, 40-45 when idling/low load. *

So... I invested quite a bit in the*Seasonic, and I am rather disappointed. This is a 6-month old unit, and Titanium to boot. It’s possible the unit was already off in my previous i6700k build, but I never noticed anything on a lower core CPU.

Anyhow, I’m relieved as this seems the solution; still waiting to hear back from Seasonic to see what they can do.
As the temps/voltages have gone down, so has my blood pressure!

Hey Derridada.... how is your system now?

I am still waiting for the right moment to take the leap of faith and buy Asus Crosshair VIII Hero (WI-FI). Can you please tell me if the system is completely stable with the latest BIOS (I think it is version 1001)? I will us it with:
Ryzen 3700x
G.Skill Trident Z RGB 32GB DDR4 Kit 32GTZR (F4-3200C16D-32GTZR)
Samsung SSD 970 Evo Plus M.2 NVMe 1 TB
ASUS GeForce RTX 2060 SUPER DUAL
Corsair power supply.

Can I expect any problems with AMD build? Any problems at all?

Thanks a lot!

[grayhex]

Matos_ wrote:
Hey Derridada.... how is your system now?

Can I expect any problems with AMD build? Any problems at all?

Thanks a lot!

Have very similar problem

3700x (and on 3600 same), all auto, no OC
C8H wi-fi (latest bios and chipset driver)
16 gb g.skill 4000cl17 at 3733 1.35v
2080ti evga ulta
samsung 970 evo plus (new, same on old 960 evo)
PSU be quite dark power pro 11 1200W (probably OK)
case be quite 700 with 4 fans silint wings 3
fan be quite dark rock pro 4 (checked and reinstalled may times)
monitor aw3418dw, logitech g pro kb, mouse and headphones

have unexpected reboots with and without load, on forza 4 for ex., and in chrome with youtube.
have on stable 1903 build W10, 1909 preview, and currently on 191010-1436 build.
temps is high( ~40 idle, ~80 at full load, ~60 in games), but is OK I guess, because 7nm, 2 chiplets, strange temp sensors and bla bla bla
memtets86 is OK (1 pass...), aida and OCCT is OK.

I have no idea what check and waiting new BIOS\Agesa.

[rbmorse]

Grayhex:

One pass with memtest isn't sufficient. Let it run overnight.

But, before you do that, pull your memory sticks and swap sockets. When you re-install them, make sure they are absolutely fully seated. The pin spacing is so tight that even the slightest mis-alignment can lead to a momentary short circuit when things heat up. .

Spontaneous reboots are hard. Usually caused by a short circuit somewhere...the AXT spec motherboard monitors voltage on one pin of the main ATX power connector (pin 14?) and when it goes below threshold it thinks you've pressed the reset button.