- last edited on by
ROGBot
So, I'm not entirely happy to see the PCH on my board ticking between 68-70 celsius in the middle of winter - and that with a slight undervolt applied. And yes, I get that it is still running in spec at this temperature, but if there's anything that can be done to easily reduce it, I'm up for that. So, let's see what's possible.
First let's take the heatsink assembly off and check it out:
The placement of the chipset right in the shadow of the graphics card isn't ideal, but it is what it is and is largely dictated by the signal paths to the PCIe bus and storage anyway. An improved heatsink assembly could be machined if I really felt like it. However, I'd rather give myself an easier time...
The heatsink is nicely finished and has a significant amount of thermal mass, but is very poorly optimised for heat dissipation, having no finning to speak of, as evidenced by the fact that the chipset actually runs very cold over the first few minutes of the system's use - and then slowly warms as the heatsink heats up.
The anodising also looks to be finished with an acetate film-type sealer on the underside, which will have a small thermal insulating effect - especially in the actual interface with the chipset.
Turning the heatsink on its side, we can also see that the textured finishes on its upper surface are appliqués that are (presumably) stuck on with adhesive tape, and have cosmetic value only. This is, on one hand, theoretically undesirable as it will reduce the thermal efficiency of the heatsink itself somewhat. On the other, it may reduce heating caused by waste air blown from the graphics card.
So, there are a few ways in which this situation can be improved. It is not difficult to envisage a way in which a fundamentally superior thermal solution could have been implemented by Asus - a heatpipe with a fin array, for example. It is also not difficult to see how the monolithic aluminium heatsink could have been improved via the addition of some finning to increase its surface area.
In the short term, I'm going to lap/polish off the anodising in the interface area, and replace the stock thermal pad with a TG Minus Pad and see what, if any, difference we see.
Let's see what that does.
And, good, that's made more of a difference than I had figured. Chipset temps are down a good 10-15 celsius with this change. Now idling in the 52-55c range once the system has fully warmed, and peaking 58-60 under heavy load.
So anybody else interested in a quick win in this area, there are a couple of things to try there. Next, I'm going to add some ducting round my (existing) SSD cooler to direct airflow specifically under the PCH heatsink - and then maybe machine some fins in it.
First let's take the heatsink assembly off and check it out:
The placement of the chipset right in the shadow of the graphics card isn't ideal, but it is what it is and is largely dictated by the signal paths to the PCIe bus and storage anyway. An improved heatsink assembly could be machined if I really felt like it. However, I'd rather give myself an easier time...
The heatsink is nicely finished and has a significant amount of thermal mass, but is very poorly optimised for heat dissipation, having no finning to speak of, as evidenced by the fact that the chipset actually runs very cold over the first few minutes of the system's use - and then slowly warms as the heatsink heats up.
The anodising also looks to be finished with an acetate film-type sealer on the underside, which will have a small thermal insulating effect - especially in the actual interface with the chipset.
Turning the heatsink on its side, we can also see that the textured finishes on its upper surface are appliqués that are (presumably) stuck on with adhesive tape, and have cosmetic value only. This is, on one hand, theoretically undesirable as it will reduce the thermal efficiency of the heatsink itself somewhat. On the other, it may reduce heating caused by waste air blown from the graphics card.
So, there are a few ways in which this situation can be improved. It is not difficult to envisage a way in which a fundamentally superior thermal solution could have been implemented by Asus - a heatpipe with a fin array, for example. It is also not difficult to see how the monolithic aluminium heatsink could have been improved via the addition of some finning to increase its surface area.
In the short term, I'm going to lap/polish off the anodising in the interface area, and replace the stock thermal pad with a TG Minus Pad and see what, if any, difference we see.
Let's see what that does.
And, good, that's made more of a difference than I had figured. Chipset temps are down a good 10-15 celsius with this change. Now idling in the 52-55c range once the system has fully warmed, and peaking 58-60 under heavy load.
So anybody else interested in a quick win in this area, there are a couple of things to try there. Next, I'm going to add some ducting round my (existing) SSD cooler to direct airflow specifically under the PCH heatsink - and then maybe machine some fins in it.
