How are contact frames not the norm for lga1700?

[Drumroll]

Ι think it's mostly aios with flat contact that need a frame. My u12a keeps my 12900k at 76c peak in CBR23 stock with no contact frame. It can also do 5.7ghz in 2 cores and 5.2ghz in all core loads (although it climbs to the low 90s in this configuration).

Not always the case.
The Noctua NH-D15 and NH-D15S which should in theory be perfect for the bent CPU, despite good contact at the centre of the chip, wouldn't get good uniform mounting pressure here.
Using the frame allowed a uniform pressure, and considerable gains.
During the first few weeks it was fine without the mod, but after a while, the temperature increased daily.
Yes, the issue would be more noticeable with flat blocks, but even for Noctua coolers, the mod has proved beneficial. I guess that, after a while, the curvature of the CPU was greater than the cooler could compensate.

 

[Bencher]

Not always the case.
The Noctua NH-D15 and NH-D15S which should in theory be perfect for the bent CPU, despite good contact at the centre of the chip, wouldn't get good uniform mounting pressure here.
Using the frame allowed a uniform pressure, and considerable gains.
During the first few weeks it was fine without the mod, but after a while, the temperature increased daily.
Yes, the issue would be more noticeable with flat blocks, but even for Noctua coolers, the mod has proved beneficial. I guess that, after a while, the curvature of the CPU was greater than the cooler could compensate.

I noticed a slight increase in temperature as well with my u12a. Used to get 76c peak at 20c ambient and now im around 79-80c (but that was after changing the motherboard as well) , but it doesn't seem to be getting any worse. Ill try the contact frame at some point and see what's what but I don't expect anything dramatic

 

[arknor]

Not always the case.
The Noctua NH-D15 and NH-D15S which should in theory be perfect for the bent CPU, despite good contact at the centre of the chip, wouldn't get good uniform mounting pressure here.
Using the frame allowed a uniform pressure, and considerable gains.
During the first few weeks it was fine without the mod, but after a while, the temperature increased daily.
Yes, the issue would be more noticeable with flat blocks, but even for Noctua coolers, the mod has proved beneficial. I guess that, after a while, the curvature of the CPU was greater than the cooler could compensate.

I don't get how a cpu can be like that when it's under pressure crom a convex surface.. makes no sense your motherboard on the back would have to be like a pimple

 

[ChloeLouCaine]

I have the ROG Maximus Hero Z690 LGA1700 board + Intel i9 12900K + NZXT Kraken Z73 AIO w/LGA1700 Bracket that came with it. I'm not going to be OC'ing and i did query this issue with a UK based popular PC YouTuber (not naming him out of respect) and he said i'd be fine without the Thermal Grizzly contact frame. Is he right or sort of right? I'm confused now and unsure if i should just take his advice? I have taken out the 3yr cover on all the hardware mentioned just so if my CPU does bend within the 3yrs can i claim on it cause its not physically my fault. Also i did read somewhere that Intel said themselves that the bending is fine as long as your temps don't get to like 95c-100c

 

[Drumroll]

Mine took a while before temperature started to worsen. That was before the heatwave.
I went with the Thermalright one as the torque is more forgiving.
A bent CPU won’t be replaced, unless it stops working.

 

[Tetras]

A bent CPU won’t be replaced, even if it stops working.

Fixed it for you

Has anyone seen Intel say "if your CPU is bent so much it doesn't work we'll replace it"? Cos, I suspect their answer would be e.g. "ask Asus, we didn't do it"

 

[Drumroll]

Personally I was sceptical of doing any change unless necessary.
The main issue will be, but not restricted to, flat coldplates, but I noticed some degradation using Noctua NHD15 and NHD15S.
Also, depending on the supplier of the retention mechanism, you’re more likely to have the issue.
My MSI uses the more prone to issue retention mechanism.
I would say, unless you have problems, don’t bother.
Reason I decided to fix it, as many other choose to do so, is for peace of mind, and at the moment some gains.

 

[ChloeLouCaine]

Personally I was sceptical of doing any change unless necessary.
The main issue will be, but not restricted to, flat coldplates, but I noticed some degradation using Noctua NHD15 and NHD15S.
Also, depending on the supplier of the retention mechanism, you’re more likely to have the issue.
My MSI uses the more prone to issue retention mechanism.
I would say, unless you have problems, don’t bother.
Reason I decided to fix it, as many other choose to do so, is for peace of mind, and at the moment some gains.

So you saying i'll be fine being as i'm not OC'ing it?

 

[Drumroll]

So you saying i'll be fine being as i'm not OC'ing it?

Yes. I would say go without any mod/change. Keep one eye on temperatures. If they start to deteriorate, only then I would bother. Also, only if the few Celsius actually causes you any negative impact. Mine I went for it because my OCD and because I’m always trying to shave the very last bit of temperature, even if the actual one is within limits and doesn’t affect performance and boost. I classify myself as a temperature control obsessed person. Same for noise.

 

[Tetras]

So you saying i'll be fine being as i'm not OC'ing it?

The only thing is, since 12th gen cpus are quite new and the issue shows up usually only after awhile, we don't actually know how it might play out. Haven't seen any 13th gen boards to see if the ILM has been modified, but Intel aren't changing socket so they must be fairly confident

 

[Bencher]

Fixed it for you

Has anyone seen Intel say "if your CPU is bent so much it doesn't work we'll replace it"? Cos, I suspect their answer would be e.g. "ask Asus, we didn't do it"

Youve never went through Intel RMA apparently. Say what you want about them, their RMA process is nothing sort of excellent.

 

[Bencher]

So you saying i'll be fine being as i'm not OC'ing it?

You won't know unless you try. I have both asus apex z690 and unify x, both worked brilliantly without the frame, but I don't have an AIO but an air cooler.

 

[Tetras]

Youve never went through Intel RMA apparently. Say what you want about them, their RMA process is nothing sort of excellent.

We're talking about physical damage here though and we all know how manufacturers lovvveeee RMA for physical damage.

 

[arknor]

The main issue will be, but not restricted to, flat coldplates, but I noticed some degradation using Noctua NHD15

I had no degradation in 2 months with a D15, I got a bracket anyway and it only had a 2c difference

61c vs 63c

I posted pics in the other thread of the paste on my cpu/heatsink from the stock retention bracket

Thermalright LGA1700-BCF Frame Corrector.

Im not seeing much in quality difference between the thermalright and thermalgrizzly correctors. Some serious price gouging going on by the germans.

forums.overclockers.co.uk

looks like no signs of bending to me, if anything the paste was pushed to the edge that wouldn't be the one bending so all along the centre where the cores are must have had really good contact.

I did the spread paste on like your putting butter on bread method

 

[Major774]

It’s how Intel designed it, it’s supposed to bend to allow for air space. Thus, providing a free heater for your PC room.

I wished there was an (Apple) bendgate story on this… to see how Intel responds.

 

[Bencher]

It’s how Intel designed it, it’s supposed to bend to allow for air space. Thus, providing a free heater for your PC room.

I wished there was an (Apple) bendgate story on this… to see how Intel responds.

Cause if it wasn't bending then the heat would've evaporated into ether?

 

[Rob_mc_1]

I picked up the JEYI CPU Frame off Amazon. It is similar to the thermaltake where it rests on the board and does not need washers. I also Didn't use the thermal paste that came with the frame. I just wanted to throw in my results in case it is of any use.
CPU: i7-12700k
Mobo: Asus Strix z690-g
Cooler: NH-U12A
Thermal Paste: Arctic MX-4
Case: Antec Dark Cube

I had tested using Cinebench R23 before and after for both stock and oc and the results were within margin of error.

I didn't do research before this. It was also cheap entertainment for a day off so I do not feel like I'm out anything. It does look nicer even if no one can see it. I do prefer it over the stock mount. Also I usually don't overclock.

I can confirm that this does nothing for my Noctua cooler. The SecuFirm mount was good at holding the bracket back plate in place during the swap.

I do appreciate the explanation in this thread. It is very informative.

Edit: Just for reference the stock temps were 76c to 80c with 20c ambient. My OC was the AI option. It was bouncing between 4.9-5.1ghz. Sometimes it did go up to 5.3. It did thermal throttle briefly. CPU fans were full speed. It mainly thermal throttle as it approached 230 watts. I'm sure I could dial it in a bit.

 

[arknor]

I can confirm that this does nothing for my Noctua cooler

Yea not seen anyone with a noctua cooler see any benefit, seems to be only manufactures with crappy tolerances on the contact plate

 

[doyll]

First you say bracket lowered temps 2c.
Then say it's not helping and "only manufactures with crappy tolerances".
Can't have it both ways mate.

Yea not seen anyone with a noctua cooler see any benefit, seems to be only manufactures with crappy tolerances on the contact plate

and

I had no degradation in 2 months with a D15, I got a bracket anyway and it only had a 2c difference

61c vs 63c

I posted pics in the other thread of the paste on my cpu/heatsink from the stock retention bracket

Thermalright LGA1700-BCF Frame Corrector.

Im not seeing much in quality difference between the thermalright and thermalgrizzly correctors. Some serious price gouging going on by the germans.

forums.overclockers.co.uk

looks like no signs of bending to me, if anything the paste was pushed to the edge that wouldn't be the one bending so all along the centre where the cores are must have had really good contact.

I did the spread paste on like your putting butter on bread method

 

[arknor]

First you say bracket lowered temps 2c.

2c is margin of error...