Anyone Try the ILM Mod with Washers to Lower CPU Temp on ROG Maximus Z690 Hero?

[Augestflex]

Hello,

I was just reading this article on Toms Hardware and then followed through with reading the additional details on Igor's Lab to lower the Alder Lake CPU temp by 5° by installing four 1MM M4 washers between the ILM and motherboard.

Curious if anyone has tried this yet with the ROG Maximus Z690 Hero and i9-12900k (or similar) and what their results were?

Toms Hardware Article: https://www.tomshardware.com/news/easy-mod-reduces-alder-lake-cpu-temperatures-5-degrees-celsius
Igor's Lab Article: https://www.igorslab.de/en/alder-lakes-cooling-problem-bend-again-around-5c-ilm-mod-for-intel-lga170...

Thanks!

[Brigman]

Yes, I've tried it on a Hero/12900K (with a LOTES socket. Not sure if any of them have Foxconns)

First tried 0.5mm and then 1mm - and didn't find any real difference in temps. In fact, the best temps averaged out across all the cores were on the original setup without any spacers - albeit only by a fraction of a degree.


That said, I've had a custom-built solid backplate on the system since I first built it, one that directly contacts the ILM backplate and mounts the cooler, and prevents the socket or PCB distorting. The system's also been very good temp-wise since being built so I wasn't expecting a huge gain from the spacers anyway.

The CPU substrate and the IHS are still nice and flat - nothing like the banana-ing some of the pics in the articles (and various other reddit/etc threads) so I'm deducing that the backplate has helped keep everything flat and square.

I've left it with 0.5mm spacing on for the time being simply because I can't be bothered pulling everything apart again to take them out. Altho they're not doing any harm either so no biggie.

There seems to me to be a bit of oversight in Intel's specifications, in that the the thickness of the motherboard's PCB is going to affect the effective clamp pressure of both the CPU ILM *and* the thermal solution, as well as the structural integrity of the entire assembled unit: A motherboard with fewer layers (or simply thinner construction) is will distort more than a thicker one, but there is no ability for the socket or CPU clamp itself, or indeed the thermal solution mounting, to account for such variations. Beyond, at least, the metal backplate on the ILM, which is inadequately thick - and also doens't span across the mounts for the thermal solution.

The more the motherboard distots, the less consistent the clamp pressure will be, and the greater the tendency for undesirable bowing distortion in the socket/CPU area.

It might not have worked for me, but I have little doubt that there *is* potential for this idea to make a difference in installations where there's no equivalent reinforcement on the mounting structure. When I first got the Hero out of its box I was surprised/disappointed how warped/non-flat its PCB was (as well as that it was only 1.1mm thick), which was largely what prompted me to make the backplate in the first place.

Where this spacer mod does help, my own view would be that it's reducing the symptoms rather than curing the real underlying problem - which is flex in the motherboard PCB under pressure from the CPU clamp *and* cooling solutions. If my experience is any guide, coolers that are shipped with good structural backplate mounting solutions will have an inherent advantage. This is also an area that that the articles looking simply at the ILM clamp pressure, as interesting as they are, haven't fully explored.

[Augestflex]

I appreciate the detailed response and findings @Brigman. I'll be keeping an eye on this development before trying myself especially with your experience that you had the best temps with the out of the box spacing. I do hope that Intel, and motherboard manufacturers, release a statement or more details on this as a 5° delta isn't an insignificant reduction.

[Brigman]

Just to reiterate @Augestflex, it's possibly not realistic to describe my experience as "out of box". My temps with zero spacing were achieved with this mounted behind, and structurally connecting the socket and cooling solution (Corsair AIO):



That's 5mm aluminium, and the spacers are cut to 2.45mm (the ILM backplate and its plastic insulating spacer are 2.4mm thick) so there is minimal opportunity for the board to distort under pressure from either the CPU clamp *or* the attachment of the thermal solution.

[Augestflex]

Brigman wrote:
Just to reiterate @Augestflex, it's possibly not realistic to describe my experience as "out of box". My temps with zero spacing were achieved with this mounted behind, and structurally connecting the socket and cooling solution (Corsair AIO):



That's 5mm aluminium, and the spacers are cut to 2.45mm (the ILM backplate and its plastic insulating spacer are 2.4mm thick) so there is minimal opportunity for the board to distort under pressure from either the CPU clamp *or* the attachment of the thermal solution.

Good point, not about of the box, but an alternative scenario with the backplate with no chance for distortion or flex, plus likely better thermal conductivity. Hmm. Perhaps I should look into that alternative for my Z690 Hero. I want to both remove any chance of the CPU substrate and motherboard bending as well as reduce thermals from said solution if possible.