Overheating chipset.

[NotHarry]

I have had a lot of trouble with my new 'Asus rog strix z690-A Gaming wifi D4' motherboard. My latest problem is a chipset that idles at 60c+.
I want to strip the heatsink for chipset in order to reseat it but you have this daft piece of plastic over the heatsink that has a cable tie attached to it, and I don't know how to remove it.
Have any of you guys striped the northbridge heat sink from one of these boards, how do you get the piece of plastic off first?

[Murph_9000]

The recent chipsets do run quite hot, it's just the way things are. The data clock rate inside the chipset doubles with each PCIe generation, so heat is just a thing that happens. Idling above 60C isn't unusual. You need to check the specific data sheet to be certain, but the max temp is very high these days, I think over 100C on some of the latest ones.

50C sounds pretty good, and I would leave the heatsink alone. It doesn't sound broken, so attempting to fix it may not be a good idea. Changing vendor is unlikely to make much difference, as they all implement the cooling spec provided by the chipset vendors (in general). Just keep a reasonable minimum case airflow and all should be good; it doesn't need that much to keep it under control.

[NotHarry]

Well, yes, but they must know that the chipset runs very hot yet they gave it a very thin bit of aluminium and then but a useless block of plastic on top of it for a even more useless cable tie. This must reduce the heatsink effectiveness by quite a lot. They should have fitted a good chunky heatsink on there over making it look pretty.
While search the net I come about many people having trouble with these boards, mainly one called the formula Z690, and a lot of those people striped the heatsink down and either refitted it properly or used another solution. They managed to drop the chipset temp down into the 30's. So Asus could have easily have done something similar.
It falls into the category of 'Bad design' They need to get away from the pretty and get back to what's practical.

[JohnAb]

A few months ago I started using 2x 8cm fans resting at the bottom of the case to blow air over the PCH heatsink which reduced the temperature by about 10C. I also reduced PCH voltages to their mid-points and chipset temp settled down to 59C idle with all NVME slots occupied.

I just set link power management to max savings (as suggested above) and now it's running at 51C. I had tried this before, but I was convinced it was affecting the stability of an external NVME drive. It turned out that the drive had some corruption and it just needed to be reformatted. Since then, I never thought to apply the link power savings again.

Thanks terrywany2k - great call 😉

I know that 59C was well within spec anyway and nothing to worry about, I just don't like things running that hot if avoidable. I'm extremely happy now with 51C and editing the Windows power plan settings is the most effective/easiest thing to do for sure. I'll see if this little change causes any other stability issues, but I'd be surprised if it did.

[JohnAb]

UPDATE

Since applying max link power savings to reduce chipset temp, I immediately started getting Windows hardware (WHEA) errors as reported by HWINfo (more specifically, PCI/PCIe bus errors). I got over 1,000 errors in less than 10 mins.

Before the change, I got 0 WHEA errors, even over weeks of monitoring.

I just set link power saving to moderate instead and the errors have completely stopped. PCH temp is now up from 51C to 56C. Since I'm happy with anything below 60C, that's still good, I just thought I'd let you all know that it might be something to keep an eye on.

[Murph_9000]

It's not "bad design" if it's running within the Intel thermal specs. I can't find the number right now, but I believe the Intel spec for max temp is over 100C. People need to recalibrate their expectations with the modern chipsets, the board vendors are not failing in their thermal design (in general), but people are expecting it to behave like historic chipsets (which is not reasonable). Yes, vendors could force them to run cooler, but they don't need to do that when Intel say they can happily run hot.

[NotHarry]

Murph_9000 wrote:
It's not "bad design" if it's running within the Intel thermal specs. I can't find the number right now, but I believe the Intel spec for max temp is over 100C. People need to recalibrate their expectations with the modern chipsets, the board vendors are not failing in their thermal design (in general), but people are expecting it to behave like historic chipsets (which is not reasonable). Yes, vendors could force them to run cooler, but they don't need to do that when Intel say they can happily run hot.

It's not about it running at safe temps, it's about how it heats up every other part of my computer that would normally stay cool. It's about the noise levels it creates.

[Silent_Scone]

Hello, 50-60 degrees Celsius is well within spec. Given the amount of data being thrown over the bus with modern storage devices and GPUs higher temps are to be expected. If one has an unnatural preoccupation with such things it would be better spent on temperatures that have noticeable impact on performance*

[NotHarry]

Silent Scone@ROG wrote:
Hello, 50-60 degrees Celsius is well within spec. Given the amount of data being thrown over the bus with modern storage devices and GPUs higher temps are to be expected. If one has an unnatural preoccupation with such things it would be better spent on temperatures that have noticeable impact on performance*

Correct, and that's what I'm trying to do. All my temps but the chipset are nice and low, GPU, CPU etc. What drives those temps up are the excess heat coming from the chipset. You see where I'm going here?
So if I can keep my chipset temps under control I keep all temps under control.
Anyway this is no longer a problem, my new Gigabyte board arrives today. I am going back to what's stable over what's flash.

[Silent_Scone]

NotHarry wrote:
Correct, and that's what I'm trying to do. All my temps but the chipset are nice and low, GPU, CPU etc. What drives those temps up are the excess heat coming from the chipset. You see where I'm going here?
So if I can keep my chipset temps under control I keep all temps under control.
Anyway this is no longer a problem, my new Gigabyte board arrives today. I am going back to what's stable over what's flash.

The chipset doesn't use enough current to have a negligible impact on surrounding components, so I believe you're wasting your time.

[NotHarry]

Silent Scone@ROG wrote:
The chipset doesn't use enough current to have a negligible impact on surrounding components, so I believe you're wasting your time.

That might have been possible if it wasn't for the fact my Son upgraded to Z690 at the same time as me but he went with Gigabyte. Yes, he has none of these problems. As far as the chipset goes Gigabyte have gave it adequate cooling, keeping the chipset around the 35c over the fancy but inadequate of the Asus board.
There are also many questionable stability issues with my Asus board that never appear on the GB board.
So, I don't believe I'll be wasting my time, my money maybe as I won't be able to return my Asus board and will take a hit in my wallet region by selling it second hand, but you have to do what you have to do.

Prey tell me this. Why in all it's wisdom did Asus, knowing the chipset gave of excessive heat did they stick a tiny bit of aluminium on there and then cover it with a plastic shroud. A recipe for disaster wouldn't you think. They should have doubled down on the heatsink, not reduce it.