01-17-2024 04:24 AM
I've been dealing with weird instability issues (been going on over a year) and after running around in circles and practically replacing everything, I've narrowed the issue to my Contact Frame being the cause.
A little background and specs
Maximus Z790 Hero 13900K, G.Skill 32g 6400, EVGA RTX 3080ti, Samsung 990 Pro 1T SSD, Corsair HX1500i PSU. Dual Loop custom Watercooling (Including the SSD)
My system will run fine and pass all benchmark and stress tests I've ran, but when I would attempt to play a game I get freezing every now and then and/or my monitors going black (no signal) and hanging the system. This issue is very inconsistent. I had originally upgraded from a Z690 Hero and 12900KF and my PSU was a Corsair AXi1500 that I've had since they came out. At first the system was quite unstable and I started having USB issues and system startup issue causing Windows 10 to corrupt over time. My voltages were a little low, but still acceptable, but I was having alot of issues with instability related to my USB devices mostly. I thought I had a bad motherboard until the PSU started having issues turning the system on. Replaced the PSU and all seemed great for a few weeks, until this Freezing and Black screen issue started manifesting.
Long story short I ended up replacing, my memory, SSD, swapping in spare PSU a few times in between, and then finally RMA'ing my video card. Each time I would take one of these troubleshooting steps the system would work great for a few weeks to a couple of months, and then start the freezing and black screen issue again.
I hadn't originally considered the TG Contact Frame being the cause because I felt I had plenty of pressure considering the screws were slightly to short and I had to push hard to get them started, so that couldn't be the issue, and no mishaps to cause any bent socket pins ( I have a stack of MB's with bent pins, every one an accident over the years).
I then loosened my waterblock (Aquacomputer Next Vision Vario) in case I had too much pressure and because my temps started creeping. When I did this the system started freezing ALOT more, so I tightened it back down and the issue almost completely stopped. I then started my Vario adjustments, which put extra pressure on the transfer plate in the corners, and I've been able to get my system almost completely stable, but most importantly has allowed me to be able to recreate my issue at will.
Has anyone else had issues with Contact Frame's? I've also heard that you can tighten them too much causing the same problems, and I may just tightened too much, or unevenly, and putting pressure on one corner is relieving too much pressure on another.
ANY thoughts would be appreciated, and does anyone know the screw size(6-32?) and length that is used with original socket clamp.
I want to get screws a little longer and redo the plate, I also need to put my water block on my RMA'd video card and this is all a pretty entailed process with my water cooling and not as easy as an AIO cooler air or water, so I need to make sure I've done it correctly, so any info would be useful. I've not seen anyone else having issues with Contact Frames.
01-17-2024 06:59 AM
I recently had issues with my 14900k - erroring out in OCCT, and some other strange things... tried remounting a few different times, didn't fully fix it.
Then, I replaced my Thermalright contact frame with the stock latch... problems solved. Rock stable. I honestly think these contact frames are causing a lot of silent issues with people's systems, and maybe getting attributed to other things, including the CPU.
My system wouldn't error out in R23, but much more intensive tests would detect the issue after extended testing. Most people probably aren't doing this kind of testing, so it goes unnoticed... until it doesn't, then the blame gets passed to other parts of the system. Undervolting is another thing that gets recommended far too casually.
01-17-2024 09:04 AM
My issues have been real intermittent, and I think you may have hit the nail on the head. My post is the short version of what I've tried and been through.
Since I started playing with the Vario it's been the most stable since I got this setup, however the Windows Reliability History is really poor unless I don't play any games. There have been times that I couldn't game more than 15 mins or so before my screens would go black. Then I would try a hardware fix and it'll work great for a few weeks. I've lost count on how many times I've reloaded windows, but I now think that's another symptom.
I'm going to try OCCT and see what happens, I haven't used it in a long while. I have to admit though, I haven't done any extensive stress tests. Only with Memtest, and only once with the Hammer test. I've done countless passes without the hammer test and haven't had one error. R23 and XTU are my other goto's for stress testing or overclocking. I don't use an overclock, except for XMP for daily usage. That said whether I'm overclocking and/or XMP or not hasn't seemed to have any effect, nor has undervolting or overvolting.1.45v has been the highest I've used
I'm using the Thermal Grizzly frame, but I also have a Themalright. Got em both at the same time and when I saw the the Thermalright frame will actually flush against the MB, I used the TG Frame to make sure I could get enough pressure. I got them for my 12900KF and Z690 Hero, but gave that setup to my son for last Christmas and got the13900K and Z790 hero for myself. I used the Thermalright on his and he hasn't had any issues supposedly.
When I used the TG Frame on the 13900K I had to barely start one screw and it literally popped up the opposite corner and I had to use quite a bit of force to push it down to start the next screw, and I tightened even more. I was thinking that the pressure may be uneven since I couldn't back the screws off to better apply even pressure from an equal start point.
01-17-2024 09:54 AM
If you're using the original Thermal Grizzly frame, the tolerances can be tight. I'd recommend getting another Thermalright, which is much easier to install, or using the stock frame. My temps are slightly higher with the stock frame but not enough to cause performance degradation as I'm using a moderate power limit.
01-17-2024 12:13 PM
I actually still have a Thermalright. I had gotten two of them at the time.
There's new info out there since the last time I researched. Found this on overclockers forum.
"With the Intel 13th/14th Gen CPU Contact Frame by der8auer we have updated the well-known mounting aid for Intel mainboards with socket LGA1700. Compared to its predecessor, the assembly of the frame has been significantly simplified by utilising a revised inner contour. For example, there is no longer any need to use a specific torque during assembly."
Problem is no one has the 13th/14th gen TG Frame, however I finally found the reddit with the screw size for confirmed 6-32 x 3/8. There are quite a few out there that have finally surfaced with the same issue. Short screws causing instability
This issue has plagued me so badly I'm still trying to come to grips that I should just use the cheap Chinese engineered device over the German one. I even watched der8auers video yet again warning about over torquing.
01-17-2024 01:35 PM
The first gen Thermal Grizzly is an inferior design, I honestly wouldn't even bother with it, especially considering the issues you're having. The question becomes whether or not you want stability, or instability with slightly better temps - I chose stability. By the way, TG took the idea for their 13/14 gen version from Thermalright, so if you want basically the same thing you should just use the Thermalright.
01-17-2024 07:50 AM
I have the 12th gen TG frame on the Apex Encore and 14900K. No issues. 8600MT stable using Intel TVB to 5.9.
01-17-2024 09:05 AM
Did you need and/or use longer screws?
01-17-2024 10:12 PM
No, I’m using the stock screws as designed. I’d wager the tolerance is out or the torque is.
01-18-2024 10:40 AM
Were they long enough?
I've seen quite a few reports, all from 13th gen owners, who had the same issue I did and the screws weren't quite long enough.
I'm pretty positive it's not torqued correctly, even though I spent quite a bit of time and effort to get it right.
I'm going after it again this weekend and I appreciate everyone response. I've been doing more adjustment to my Vario to hopefully get a better handle on why it's out wack. My system and design, unfortunatly, doesn't make it an easy task.