Crazy PCH temperature on Z690 Carbon

JJG

New member
Joined
Dec 31, 2019
Messages
3
Ended up returning the MSI MPG Edge DDR4. Funky wake on sleep/power off issues. Cheap Chipset Heatsink, average audio, limited bios options and not even 1 temperature sensor on the mobo. Asus Z690 Apex and Gskill DDR5-6000 arrived on X-mas and couldn't be happier. The quality difference is night and day. And the amount of options in the bios for tweaking on a high-end Asus board is unparalleled. PCH runs over 10C cooler, probably because of bigger heatsink. Thanks for the 3 week filler board MSI- adios!
 

Attachments

citay

Pro
SERGEANT
Joined
Oct 12, 2016
Messages
8,530
Asus Z690 Apex and Gskill DDR5-6000 arrived on X-mas and couldn't be happier.
Hmmm... Asus Z690 Apex is ~650 EUR, compared to ~320 EUR for the MSI MPG Z690 Edge DDR4.
Gskill DDR5-6000 32 GB kit should be ~580 EUR, compared to ~170 EUR for a G.Skill DDR4-3600 CL16 32 GB kit.

That's roughly 1230 EUR vs. 490 EUR, for not having standby issues (already getting fixed in beta BIOS for MSI), for a better chipset cooling, for better onboard audio (never mind the soundcard you could buy from the difference), more BIOS options (?), and another temperature sensor? Of course you also get slightly higher performance from your faster RAM, depending on the use case.
But is this really worth a 250% higher price than the previous setup, for very similar performance? I guess it counts that you are happy.
 

ElderLake

New member
Joined
Jan 2, 2022
Messages
4
I noticed this problem

I bought a 12600kf and Z690 tomahawk ddr4. My brother bought the same
but we have different temp on the motherboard.
His, is about around 50°c, while mine is around 63°c

We have the same case, cooler, ram and main ssd nvme.
Room ambiant temp are the same
Difference is power supply , graphic card and case fans ( mine are all 120mm)
and i have one hdd et one ssd more than my bro

Why is there , a temperature difference between same motherboard ?
Is it the same as cpu, silicon lottery ?
 

citay

Pro
SERGEANT
Joined
Oct 12, 2016
Messages
8,530
No, it's most likely the differences that you still have in your hardware, as well as the case airflow, perhaps in combination with a slight sensor divergence.

If the PCs are in the same room or if you can put them together, take off the side doors and put your hand on each chipset heatsink, that large temperature difference should be easy to feel. Then you have to look at the airflow (the power supply's fan is also part of the fan/airflow setup) and how much heat the graphics card generates.

If you can't feel such a big difference by touching the chipset coolers, then you can also look closely at power draw using HWinfo64 Sensors. In Windows there can obviously be different things going on in the background, depending on what you have installed. Even different drivers etc. can make a difference.

Of course, if you compare the temperatures in the BIOS after being in there for a while, then they should be somewhat similar. But you don't have the same airflow and graphics card.

If you want to find out if the graphics card has anything to do with it, swap the cards between the systems. If there's no difference from that, maybe the air from his fans hit the chipset heatsink better.
 

ElderLake

New member
Joined
Jan 2, 2022
Messages
4
ok thanks
We have a Fractal R4
I have a msi gtx 1070 gaming X, 45°c at idle
My bro has a zotac 970 crappy dual fan, same temp. I own also this card so i will take a look
 

kmoore39151202d2

New member
Joined
Jan 28, 2022
Messages
3

Color_Fx77

New member
Joined
Apr 27, 2022
Messages
5
Hello everyone! I have tried next MSI models: TOMAHAWK WIFI DDR4, MPG Z690 FORCE WIFI , MPG Z690 EDGE WIFI DDR4, MPG Z690 EDGE WIFI DDR4, MEG Z690 UNIFY-X alredy, all of them got the same problem to HOT PCH! SO, i've read a lot of topics about this theme , so i've even opened up the Unify-x motherboard and checked if it was a bad thermal pad - no way everything is useless.
I conclude that based on the fact that the bridge between the PCH and the processor in the 4th generation is so fast, so productive that it now heats up 2 times more than before, so that's why it's burns! and it need's to be cooled with something more substantial, rather than a passive radiator. So this is a technological miscalculation, while MSI and Asus are trying to survive their hardware to the maximum, and the rest, as usual, are not conscientious. - that's my theory.
Currently I have:
5 disks 2x m2 evo 970 plus 1tb, evo 980 pro 1tb, 1 evo 870 1tb and 1 hdd 1TB. And for the clearens i have turn on wifi!
12700KF (game mode) ddr5 6200 (try it)
gigabyte Aorus master 3080ti (with water block from the aquatuning)


And finally My SOLUTION to solve this problem!
So I managed to achieve a decrease in temperature from the PCH of the diode 73 to 54 using the following manipulations with the BIOS settings. WITHOUT LOSS OF PERFORMANCE IN GAMES.
Required first to update the BIOS to the latest version, in my case this is the version "H30" 25.03.22, otherwise the network will shut down!
and than do like at the screenshots below!
1.jpg
2.jpg
3.jpg
 
Last edited:

citay

Pro
SERGEANT
Joined
Oct 12, 2016
Messages
8,530
So I managed to achieve a decrease in temperature from the PCH of the diode 73 to 54 using the following manipulations with the BIOS settings. WITHOUT LOSS OF PERFORMANCE IN GAMES.
The first photo shows the Active State Power Management modes, Auto should default either to L0s only or maybe to L0s+L1. The best would be to select L0sL1 manually, this is maximum energy saving. Like i mentioned before, it can only work with the "Balanced" power plan in Windows, and even then, Windows often tends to deactivate it for compatibility reasons (presumably a few rare devices don't play nice with it). The energy reports under "C:\ProgramData\Microsoft\Windows\Power Efficiency Diagnostics" will show if the power savings actually get applied or not. I would say, this might not have much of an impact.

The second photo shows the Package C State Limit set to C6. This is a power-saving mode for the CPU package, the higher the number, the deeper the power savings. Actually, you can set it to C10 or the highest number that's available. Much like the ASPM modes, Windows will automatically use the deepest supported C-state, so that will end up being C6 or C7 if you're lucky (C10 is only supported on notebook CPUs). Also, you should enable C1E support, more CPU power-saving. But all of this is really just for the CPU.

Therefore, i think most of your power-saving for the PCH/chipset actually comes from the option in the third photo:
You are restricting the DMI Link Speed, which basically consists of PCIe lanes, from PCIe 4.0 x8 to 3.0 x8.

DMI.png


As you can see on this picture, everything that is wired via the PCH / Z690 chipset, it all has to go through this DMI link between PCH and CPU.
This explains why there is no loss of performance in games, because games mostly rely on the graphics card, the CPU and the RAM, all of which are connected directly to the CPU.

However, everything going through the PCH first, meaning, all M.2 SSDs except the first one, all the USB ports, the LAN, the Wifi, all this is now restricted to PCIe 3.0 x8 speed instead of PCIe 4.0 x8.
With two M.2 PCIe 3.0 x4 SSDs such as your 970 EVO PLUS's, you can theoretically fully saturate this DMI link with your setting. In practice, it won't happen often. But Intel apparently saw the need to upgrade this DMI link from DMI 3.0 x8 (Z590) to DMI 4.0 x8 (Z690). You can have a loss in performance, but not in games.


12700KF (game mode)
You mean Game Boost? I would turn that off again. I explain the reasons here: https://forum-en.msi.com/index.php?threads/bios-game-boost-cpu.373168/#post-2112970
 

Color_Fx77

New member
Joined
Apr 27, 2022
Messages
5
However, everything going through the PCH first, meaning, all M.2 SSDs except the first one, all the USB ports, the LAN, the Wifi, all this is now restricted to PCIe 3.0 x8 speed instead of PCIe 4.0 x8.
With two M.2 PCIe 3.0 x4 SSDs such as your 970 EVO PLUS's, you can theoretically fully saturate this DMI link with your setting. In practice, it won't happen often. But Intel apparently saw the need to upgrade this DMI link from DMI 3.0 x8 (Z590) to DMI 4.0 x8 (Z690). You can have a loss in performance, but not in games.
2 citay
Sure, i know about all of that, but do you have any better sollution right now?
 

thescopema151402d6

New member
Joined
Jun 1, 2022
Messages
1
Hi,
I have a MEG Z690 UNIFY-X and have it a bit more extreme. The PCH is usually around 75 to 77 degrees at idle. I have also tried all possible bios settings but it does not help.

Bios version is the last official A3.

2022-06-01_193631.jpg
 

plutomate

Well-known member
MAJOR GENERAL
Joined
Jul 30, 2015
Messages
5,472
It's not a problem, it is just the way it is, PCH temperature gets high when you have more M.2 device. Furthermore, the "high temperature 76C" you describe will not damage anything.
No need to worry at all :beerchug:
 

Color_Fx77

New member
Joined
Apr 27, 2022
Messages
5
Hi,
I have a MEG Z690 UNIFY-X and have it a bit more extreme. The PCH is usually around 75 to 77 degrees at idle. I have also tried all possible bios settings but it does not help.

Bios version is the last official A3.
1)
I don't think you did exactly what I recommended in my comment. The temperature you are talking about - I reached after 30-50 minutes of playing with 4 M.2 on unify-x without all the settings that I mentioned above with screenshots. So recheck it please.
after all the adjustments, the temperature will be somewhere around 60-70 .. also high but not 80


2)
Yes, this is the reason why, in my researches, I abandoned Unify-x in favor of Carbon and Edge
If you do not use stock memory DDR5 for overclocking 6600+, YOU just using those that are already configured for XMP, SO - you do not need a Unify-x.
That;s right if you use more than 2 M.2 u will got such hight temp,
There is only one way to fix it somehow, that that i've done at all my current 4 PC's look at the photo...

3)
The fact that at such temperatures nothing will suffer is stupidity. It's like keeping a piece of metal constantly hot - do you think sooner or later what will happen to it? it will change its properties and break. If you play games once a week for 1-2 hours. - then right you can not bathe, and if not - It is better to change the motherboard to Carbon\Edge\Force, or at least do what I advise with bios and fans. Otherwise, your M.2 possibly be destroyed about in 6 or 12 months. Well, power circuits will slowly burn out ect.

P.s BTW interesting fact that at the lowest model - z690 TomaHawk DDR4 - there is no problem with such hot PCH at ALL!!!
I've Build the comp for my freind with this MB and just dumbfounded
 

Attachments

Last edited:

citay

Pro
SERGEANT
Joined
Oct 12, 2016
Messages
8,530
3)
The fact that at such temperatures nothing will suffer is stupidity. It's like keeping a piece of metal constantly hot - do you think sooner or later what will happen to it? it will change its properties and break. If you play games once a week for 1-2 hours. - then right you can not bathe, and if not - It is better to change the motherboard to Carbon\Edge\Force, or at least do what I advise with bios and fans. Otherwise, your M.2 possibly be destroyed about in 6 or 12 months. Well, power circuits will slowly burn out ect.
While i agree that ~75°C idle is hot for a desktop motherboard PCH temp and i would also raise an eyebrow when i saw that on my board, i don't think there is any risk of something on the board or even your SSDs being "destroyed" soon. There is no component - neither the PCH itself, nor anything in the vicinity, and neither on the SSDs - whose lifespan is negatively affected in a major way. The main components whose lifespan depends on the ambient temperature are capacitors. That's why old/cheap PSUs are often among the first things to fail in a PC, because they use a bunch of electrolytic capacitors which are close to hot components like MOSFETs. There are no such capacitors near the PCH, and the PCH itself, or the SSDs, don't change their properties from being subjected to 80°C temperatures.

In notebooks, PCH temps of 80°C or higher are not uncommon at all. Yet they tend to work for many years, even a decade or longer. SSDs have throttling mechanisms if their temperature sensors detect any overheating. Changing your mainboard model because you are worried of some sort of "metal fatigue"-like failure from the PCH temperature is not based on a realistic scenario.

Better airflow through the case is always good. But in case of the PCH, it would be more for peace of mind, not to prevent a catastrophic failure.
I don't want to play down the high temperatures - i wouldn't like that either on my board - but there also shouldn't be any unfounded fear about early failures.
@thescopema151402d6, what are your PCH temps with data transfers going over the PCH, for example when running a disk benchmark?
 
Last edited:

Color_Fx77

New member
Joined
Apr 27, 2022
Messages
5
While i agree that ~75°C idle is hot for a desktop motherboard PCH temp and i would also raise an eyebrow when i saw that on my board, i don't think there is any risk of something on the board or even your SSDs being "destroyed" soon. There is no component - neither the PCH itself, nor anything in the vicinity, and neither on the SSDs - whose lifespan is negatively affected in a major way. The main components whose lifespan depends on the ambient temperature are capacitors. That's why old/cheap PSUs are often among the first things to fail in a PC, because they still use a bunch of electrolytic capacitors which are close to hot components like MOSFETs. There are no such capacitors near the PCH, and the PCH itself, or the SSDs, don't change their properties from being subjected to 80°C temperatures.

In notebooks, PCH temps of 80°C or higher are not uncommon at all. Yet they tend to work for many years, even a decade or longer. SSDs have throttling mechanisms if their temperature sensors detect any overheating. Changing your mainboard model because you are worried of some sort of "metal fatigue"-like failure from the PCH temperature is not based on a realistic scenario.

Better airflow through the case is always good. But in case of the PCH, it would be more for peace of mind, not to prevent a catastrophic failure.

On 99% of laptops, long-term use in games leads to the irreversible degradation. The performance of laptops is cut down and inevitably decreases over time.
any experienced gamer knows this.
In general, I do not want to breed discus here, this is complete absurdity. There are facts that if hardware works for you forever 80+ degrees, then it will soon be it slowly dies
. and its performance is questionable. just like video cards work at temperatures of 80+ but if you manage to make it 40-60 you will see that it has become more productive.

Believe me, all your components in the composition can deteriorate faster or slower. -the only matter of your choice

Yes, of course, a lot of hardware operates at exorbitant temperatures today. but this is not because it should work this way, but because the manufacturer does not worry about how you will cool it, it is generally beneficial for them that your hardware fails faster and you'll buy a new one! as with cars and othere tech
 

Hexenhammer

Glory to Ukraine!
Joined
Nov 7, 2020
Messages
4
Hello guys.
I have Unify X, 12900K, 6000Mhz DDR5, and i have 5 m.2, x2 2tb EVO Plus, x1 2tb XS70 Gen 4.0 [the only gen 4.0 i have], x1 2TB MP510, x1 U.2 4TB Intel Enterprise Drive connected with cable to M.2
I have no GPU in the first slot i use a riser so GPU is vertial and doeant heat up anything
but in second slot i have Intel Optane 480gb

I found 2 issues.

#1 Both M.2 drives that conencted across the chipset run HOT, i have them in the biggest heatsinks, is thoise black tall 3 heatpipe heatsink and they IDLE at 48-50 [becuase of the chipset]
#2 chipset Idleas at 70 with Power saving turned on [on auto, will change manually and see how it runs]

I found a solution and i recommend it to everybody its not that ugly and its the only solution that we can just install, easily,without losing speed and performance.

With one of my otehr PC, ASUS board i received a metal fan bracket that screws into one of the motherboard standoff screws that hold the board it can hold a small fan, i think 40mm and 50mm, right now i installed it for the VRM [im on custom loop and have no air flow]
But im going to move it and buy another one, here is a link on aliexpress its like 2USD, BUT don't buy their fan, get a noctua fan



Also i recommend to swap the heatsink thermal pad, its up to you but if you want a recommendation for good one let me now, these are like paste but they thermal pads you put them on screw the heatsink and they squize and form just the right size, i use this "paste" everywhere now [its not a paste but Thermal pad replacement, Linus Tech tips had a video about it, sold on amazon].
If you afraid and think it might void the warrnaty, at least Tighten up the screws that hold the heatsink

If you want a better solution, here is a Copper universal chipset heatsink with fan, its chinese so replace the fan with Noctua, it should be smallest and least visible solution
 

Hexenhammer

Glory to Ukraine!
Joined
Nov 7, 2020
Messages
4
Im happy to report that i managed to reduce the PCH temp by 40%
1. Changed thermal pad
2. installed the tiny noctua fan that runs 5400rpm, but because of its size i cant hear it [even without the fan the reduction from thermalpad swap was amazing]
I was idling at 70 with power saving features and 80 without them and 85 under load.
Now i idle at 48 to 55 and 65 under load
 

Nichrome

Friday? Deploy to prod!
Global Moderator
Joined
Nov 7, 2013
Messages
20,407
What thickness was the thermal pad? Same size as the one you replace, or thicker?
 

Hexenhammer

Glory to Ukraine!
Joined
Nov 7, 2020
Messages
4
What thickness was the thermal pad? Same size as the one you replace, or thicker?
I used smaller [thermalight 12W or 14W]
2mm
I also tightened the screws on the back and cleaned both surfaces with alcohol

Also i tested with K5-PRO from amazon, its Thermal pad replacement in Paste like, not liquid more like goo
This one you apply on the die and around the the die on the chip with good amount 2mm thick and when you press the heatsink it squishes it to right size and you get contact and heat transfer from the die and from the around area too, this one is great too, good results
 
Top