Thermal Grizzly Announces Contact Frame For Intel Alder Lake, Promises to Reduce Temperatures by up to 10º

[Raevenlord]

Thermal Grizzly has developed a new Contact Frame designed specifically to fix bending issues present with Intel's latest 12th Gen, Alder Lake CPUs. Developed in partnership with overclocking extraordinaire Der8auer, the new Contact Frame promises to lower operating temperatures on Intel's Alder Lake. According to the company, this improvement is achieved by fixing that platform's independent loading mechanism (ILM), which has been proven to slightly deflect the integrated heatspreader (IHS), reducing its heat transfer capability.

As tested by Igor's Lab, the new contact frame for LGA 1700 reduced the operating temperature of Intel's Core i9-12900K by as much as 10.19 °C - from 70.48 °C without the Contact Frame and towards 60.29 °C after it was installed. The CPU was configured to run popular stress test Prime95, with Small FFT at a fixed 5 GHz frequency on its P-cores. The processor's E-cores were deactivated so as not to compromise the results, while the memory subsystem was run at DDR5-7000. Thermal Grizzly's Contact Frame isn't the only product in this category, and the company is introducing their product at €39.90 for the German and European markets (~$36). Enthusiasts have likely spent more in cooling upgrades that delivered a lesser final operating temperature improvement.



View at TechPowerUp Main Site | Source

 

[zlobby]

You know you screwed badly when 3rd party companies make crutches for your CPUs, and make profit out of it.

 

[ZoneDymo]

sooo the exact same thing Thermalright made a month or so ago?

 

[HisDivineOrder]

If you watch the video, you see this only gets 3-4 degrees on average. Seems like a, "Lap and you'll get better temps" kinda situation to me. Which, yeah, sure, but still. Keep in mind he was using the best case scenario cooler, too. If the guy who made it can't even get much out of it, then I don't think there's much to discuss except someone trying to make money doing something a few extreme users actually need.

 

[napata]

sooo the exact same thing Thermalright made a month or so ago?

Der8auer's video is more than a month old so it's not a new thing.

 

[GerKNG]

in a german forum a bunch of people are talking about it, bought it and tested it.
many of them said that it did nothing, one said it got worse.

the whole "IHS BENDING!!!!" fuzz is completely out of proporting just like the "bad caps" on early ampere cards even if it was a driver bug.
if you have bad contact to the IHS open and close the mounting mechanism 10 times in a row and it gets softer to the point where the mounting pressure and distribution is fine.

 

[P4-630]

in a german forum a bunch of people are talking about it, bought it and tested it.
many of them said that it did nothing, one said it got worse.

the whole "IHS BENDING!!!!" fuzz is completely out of proporting just like the "bad caps" on early ampere cards even if it was a driver bug.
if you have bad contact to the IHS open and close the mounting mechanism 10 times in a row and it gets softer to the point where the mounting pressure and distribution is fine.

You got a i7 12700K? How are it's temps in your system?

 

[GerKNG]

You got a i7 12700K? How are it's temps in your system?

30 Minutes R23 loop:

 

[P4-630]

30 Minutes R23 loop:

Temps look fine @ stock , I won't be OC'ing my i7 12700K for a while, I have to cool it with a NH-U12A.

 

[GerKNG]

Temps look fine @ stock , I won't be OC'ing my i7 12700K for a while, I have to cool it with a NH-U12A.

i used the NH U12A in both tests as well.

 

[Eskimonster]

Man they should be forced to send one out to every single buyer of the said chips.

 

[ZoneDymo]

Der8auer's video is more than a month old so it's not a new thing.

I mean sure but its just a piece of metal and they are bringing it to market now, which again, Thermalright already did so I fail to see the point really.
Especially seeing as the Thermalright version is a 3rd of the price.....

 

[GerKNG]

I mean sure but its just a piece of metal and they are bringing it to market now, which again, Thermalright already did so I fail to see the point really.

it was actually patended by roman and thermalright broke the patent (at least in germany)

 

[ZoneDymo]

it was actually patended by roman and thermalright broke the patent (at least in germany)

got a link to that info?

 

[P4-630]

i used the NH U12A in both tests as well.

Not bad at all!

 

[GerKNG]

got a link to that info?

it has been at least 4-6 weeks and what i could fine (right now) is him talking about it in the HWluxx forum about it:

Stabiles Alder-Lake-Package: Der Thermal Grizzly CPU Contact Frame zeigt sich

Falsch. Er liegt am ganzen Rand der CPU auf, also alle Seiten. Bei Romans wohl etwas zu viel, weshalb er dazu rät nicht voll anzuziehen. Da hat er wohl 1 mm zu viel weg genommen am Material. viel Erfolg beim demontieren ohne Schäden. Aber ja, grundsätzlich korrekt. 1650800378 Nicht korrekt...

www.hardwareluxx.de

he had a patent on it since 2015 and they cloned it without permission.
here is his youtube comment.

 

[ir_cow]

I want to buy this and try it for myself. I haven't peronally had this mounting problem. But I am using EK blocks and the backplate is solid. I still think its a cooler issue, rather than socket.

 

[erocker]

I wonder if this issue is happening more to cheaper motherboards, or boards with less PCB layers. Doesn't seem like's it's really the socket but what the socket is mounted to.

 

[ir_cow]

I wonder if this issue is happening more to cheaper motherboards, or boards with less PCB layers. Doesn't seem like's it's really the socket but what the socket is mounted to.

Probably. We all seen the PCB bend on budget MBs when a cooler is mounted.

 

[Sithaer]

I wonder if this issue is happening more to cheaper motherboards, or boards with less PCB layers. Doesn't seem like's it's really the socket but what the socket is mounted to.

I have a rather mediocre Asus Prime Plus D4 B660/12100F and used both the stock cooler and the current ID-Cooling SE-224-XT with no bending issues to my knowledge.
Temps are where they should be and I tried to check but from what I can see the contact is proper between the IHS and the cooler.

 

[P4-630]

Probably. We all seen the PCB bend on budget MBs when a cooler is mounted.

I got an Aorus Master incoming tomorrow, hope I won't have this issue with my NH-U12A on this board.

 

[Solid State Brain]

With the stock unmodified frame I had an irregular thermal paste footprint with my 12700K and Noctua NH-D15S. The CPU was clearly bending. Adding 1 mm washers as suggested by Igor's Lab and Buildzoid decreased temperatures by 6 °C on average at about 200W and visibly improved cooler contact.

Lapping the CPU (which definitely voided warranty) decreased temperatures by maybe 1-2 °C compared to the washer-only solution, so not much more. I plan getting a Thermalright frame in a few weeks, which might improve temperatures by a few °C more, and then hopefully I will be done torturing this CPU.

I would have probably got a Thermal Grizzly Contact Frame if it wasn't so expensive.

 

[Ed_1]

I don't like that there no set mounting tension, you have to turn the screw x deg and test, and if not great retry with more or less amounts.
This might work ok setting up on water-cooled but would be a major pain with big HS as you need to remount HS each time to get at the screws, plus once you find the right amount you need to mark, fix the screw head so it doesn't ever move.

IMO different sized washers should have been included so screw mounting is secured properly.

 

[Nanochip]

Dialing in the proper voltage settings instead of relying on "AI Overclocking" features also reduce temps significantly. Asus' AI Overclocking might be stable but at the expense of unnecessary heat and temps. I was able to dial in a stable overclock of my 12700K @ 5.0 all core, 5.3 on 1 core, 5.2 on 3 cores, 5.1 on 4-7 cores... running cool at 74 degrees in Cinebench. 20 degrees below Asus' "AI Overclocking" nonsense.

 

[Solid State Brain]

I don't like that there no set mounting tension, you have to turn screw x deg and test and if not great retry with more or less amounts.

The mounting tension does not seem to be important for temperatures, but it can affect system stability. Best results seem with minimal tension. See:

German Engineered Bend Aids for Intels LGA1700 – Thermal Grizzly CPU Contact Frame and Alphacool Apex Backplate Thermal Testing | Review | Page 2 | igor´sLAB

The LGA1700 socket, which Intel introduced for the 12th generation of their Core CPUs last November, is known to have one or two problems with bending hardware. For context, I recommend our other two…

www.igorslab.de

In his video on the frame, der8auer gives the advice to first tighten the screws between thumb and index finger and then to continue turning them exactly 90 °C using the markings. In my tests, however, this was already too much contact pressure, which resulted in instability of the DDR5 OC – more about this on the next page. Instead, it has worked best for me to really only tighten the screws so lightly that they just don’t come loose on their own. If the frame can still be moved back and forth minimally on the CPU without the screws turning, the RAM also still works like with Intel’s standard ILM – less is more here.

From my own testing, mounting the CPU without ILM at all also works and gives good results (better than with the washers), but then the CPU will get stuck to the cooler due to suction cup effect. And now that I lapped the CPU such effect has become rather strong, so I haven't tried it anymore. In this case mounting pressure was just due to the cooler.