News Noctua's Offset Mounting Bars Can Lower AM5 CPU Temps

[Admin]

Noctua has released offset mounting bars for AM5 coolers to lower temperatures of AMD's Ryzen 7000-series processors .

Noctua's Offset Mounting Bars Can Lower AM5 CPU Temps : Read more

 

[TechieTwo]

I'm not sure who "all the people" are that are bitching about AMD's CPUs running hot as they are totally reasonable temps for modern performance oriented CPUs. Intel CPUs ran hotter than AMD CPUs for years and few people complained. I recall seeing an egg cooked on an Intel I386 many moons ago.

 

[Deleted member 2838871]

I'm not sure who "all the people" are that are bitching about AMD's CPUs running hot as they are totally reasonable temps for modern performance oriented CPUs. Intel CPUs ran hotter than AMD CPUs for years and few people complained. I recall seeing an egg cooked on an Intel I386 many moons ago.

Exactly. This is my first AMD build since 2001... and believe me... after running Intel for the last 20 years these AMD processors don't get hot at all.

In a 68F ambient room my idle temps are a bit high (45-50C) on my 7950x3D... but temps while gaming and doing productivity are quite normal... 60-70C. A max 30 min all core burn test peaked at 89C with the D15 cooler. It went to 89C after 1 min and stayed there for 29 mins.

Temps are fine... but this is pretty interesting and I just ordered the offset bracket. For $5 why not?

 

[Kamen Rider Blade]

Temps are fine... but this is pretty interesting and I just ordered the offset bracket. For $5 why not?

Since you're going to have to re-mount anyways, you might also want to try the new hotness in Thermal Paste.

The Corsair XTM70

Reddit - Dive into anything

www.reddit.com

Review: Corsair XTM70 Thermal paste - Overclocking.com EN

Test / Review of the XTM70 thermal paste from Corsair. The best thermal paste in our comparison sold at a gold price.

en.overclocking.com

Every review I've seen states that it's one of the top performing Thermal Pastes and has only recently came to market.

 

[Deleted member 2838871]

Since you're going to have to re-mount anyways, you might also want to try the new hotness in Thermal Paste.

The Corsair XTM70


Every review I've seen states that it's one of the top performing Thermal Pastes and has only recently came to market.

Done. I used MX-4 on the build but for $8 I'll try it... $13 total for the paste and the bracket. ETA Thursday... will update this thread with my thoughts after installation.

Perfectly happy with temps as is... the D15 is a beast cooler... but for $13 why not?

 

[deesider]

Exactly. This is my first AMD build since 2001... and believe me... after running Intel for the last 20 years these AMD processors don't get hot at all.

In a 68F ambient room my idle temps are a bit high (45-50C) on my 7950x3D... but temps while gaming and doing productivity are quite normal... 60-70C. A max 30 min all core burn test peaked at 89C with the D15 cooler. It went to 89C after 1 min and stayed there for 29 mins.

Temps are fine... but this is pretty interesting and I just ordered the offset bracket. For $5 why not?

I needed to undervolt my 12900K to get it down to 89C with a D15

 

[hotaru.hino]

In a 68F ambient room my idle temps are a bit high (45-50C) on my 7950x3D... but temps while gaming and doing productivity are quite normal... 60-70C. A max 30 min all core burn test peaked at 89C with the D15 cooler. It went to 89C after 1 min and stayed there for 29 mins.

You're also using an X3D version, which is has a lower default operating voltage, a lower TDP profile, and a lower Tj Max spec.

I'm not sure who "all the people" are that are bitching about AMD's CPUs running hot as they are totally reasonable temps for modern performance oriented CPUs. Intel CPUs ran hotter than AMD CPUs for years and few people complained. I recall seeing an egg cooked on an Intel I386 many moons ago.

My issue is the CPUs are using a lot of power for very little gain. Even dropping the power profile to 105W TDP barely drops performance (https://arstechnica.com/gadgets/202...ew-zen-4-starts-off-expensive-but-impressive/)

However I would argue temperature is irrelevant here, because it depends on the cooling performance. But using less power means the same cooler doesn't have to work as hard.

 

[Deleted member 2838871]

I needed to undervolt my 12900K to get it down to 89C with a D15

Hahah... you're reminding me of my recent 7700k and 10900k builds. The 7700k was great after the delid... I lost 15C across the board.

You're also using an X3D version, which is has a lower default operating voltage, a lower TDP profile, and a lower Tj Max spec.

My issue is the CPUs are using a lot of power for very little gain.

Yep... and that's just one of the reasons I went with the x3D variant. Why burn all the added power and have the higher thermals with a 7950x or a 13900k? AMD should really be using this as part of their marketing strategy.

 

[10tacle]

I needed to undervolt my 12900K to get it down to 89C with a D15

Surely you know that is way out of spec norm with that cooler. You have a serious contact and/or paste issue with your heatsink on the processor. I also have a D15 with my overclocked 11700K (5.0GHz on all 8 cores) sitting in a high volume air cooled case, and right now in summer indoor ambient temps of 79C, it doesn't get more than 72-75C on the warmest cores. That's with the assigned PWM motherboard fan curve setting of 75C at 80% fan speed (90% at 80C, 100% at 85C). You should be in high 70s/low 80s under load with that cooler if overclocked to 5.0GHz all cores, let alone stock. And I'm basing that on Tom's own review of the 12700K with a Corsair H115i AIO cooler which is comparable to the D15 in performance.

 

[Deleted member 2838871]

Surely you know that is way out of spec norm with that cooler.

Yeah the D15 is really an amazing cooler. I heard all about the Peerless Assassin but I went with Noctua based on their reputation and service.

I am looking forward to reseating the cooler with this new bracket and new paste... will also give me another look at how well my "pea sized drop" in the middle of the CPU turned out. This will be the 2nd time after I swapped out a dead processor and board as part of the implosion drama back in April.

Paste covered the entire IHS and looked good without being overboard... so expecting the same this time.

 

[10tacle]

I am looking forward to reseating the cooler with this new bracket and new paste... will also give me another look at how well my "pea sized drop" in the middle of the CPU turned out.

And that reminds me that another thing about overclocking makes a difference: what paste do you use? I use Thermal Grizzly Kryonaut which is among the top performers. For reference it performs about 3C cooler than the popular Arctic Silver 5. I applied it to my RTX 3080 Ti as well.

 

[Deleted member 2838871]

And that reminds me that another thing about overclocking makes a difference: what paste do you use? I use Thermal Grizzly Kryonaut which is among the top performers. For reference it performs about 3C cooler than the popular Arctic Silver 5. I applied it to my RTX 3080 Ti as well.

I used Arctic MX-4 with the build and with the 2nd CPU.

A post upthread linked some new and exciting paste so for $8 I figured I'll try it with the new bracket.

 

[Kamen Rider Blade]

Done. I used MX-4 on the build but for $8 I'll try it... $13 total for the paste and the bracket. ETA Thursday... will update this thread with my thoughts after installation.

Perfectly happy with temps as is... the D15 is a beast cooler... but for $13 why not?

I think you might've bought the wrong Thermal Paste.
The XTM70 is different formulation from the XTM50.

You might want to double check because all the reviews says there is a different formulation and that there is only one XTM70 that is the over priced $25 package, that I linked you.

Corsair doesn't sell the XTM70 Thermal Paste in a stand-alone Tube.

Why don't they, go ask Corsair.

 

[deesider]

Surely you know that is way out of spec norm with that cooler. You have a serious contact and/or paste issue with your heatsink on the processor. I also have a D15 with my overclocked 11700K (5.0GHz on all 8 cores) sitting in a high volume air cooled case, and right now in summer indoor ambient temps of 79C, it doesn't get more than 72-75C on the warmest cores. That's with the assigned PWM motherboard fan curve setting of 75C at 80% fan speed (90% at 80C, 100% at 85C). You should be in high 70s/low 80s under load with that cooler if overclocked to 5.0GHz all cores, let alone stock. And I'm basing that on Tom's own review of the 12700K with a Corsair H115i AIO cooler which is comparable to the D15 in performance.

I probably will repaste it next time I go delving into the case and see if that makes any difference - but my temps seem fairly typical: https://www.overclock.net/threads/taming-my-i9-12900k-with-air-help.1797265/

I believe the issue is that the D15 heatpipes become saturated at around ~220W - anything above that is cooled only by the more limited heat conduction through the copper rather than water evaporation.

- Also, I think the 11700K maxs out at 180W, depending on your clocks of course

 

[Deleted member 2838871]

I think you might've bought the wrong Thermal Paste.

You're right... leave it to Amazon to have an item page and 3 different links leading to 3 different items... (XTM70, XTM50, TM30)

Good catch. I didn't even notice. Good news is it hadn't shipped yet so I just went and cancelled it without a hitch. Unfortunately there appears to be an issue with the XTM70. I've tried to order it to ship to my home... my work... and the local locker and Amazon says it's unable to deliver but shows the paste in stock.

Weird... never seen this happen.

So I found it at my local Best Buy right down the street (same price, XTM70) and just picked it up. Ready to go for when the bracket gets here.

So it's not a $13 upgrade... it's a $30 upgrade. Still want to try it. I was talking with a guy on YouTube earlier who has the same CPU and cooler and he asked me what my Cinebench R23 score is... which is 35,000. His was slightly higher but he told me that he had seen AIO users with 36k and that he thinks we are thermal throttling with the D15 cooler.

I don't believe that is the case but I am going to run benchmarks again once the new bracket and paste are installed. It's my understanding that the 7950x3D throttles at 89C and I haven't gotten anywhere close to that in gaming/productivity... only in an all core burn test did I hit 89C. Everything else as mentioned earlier has been in the 60-70C range.

 

[Kamen Rider Blade]

Good catch. I didn't even notice. Good news is it hadn't shipped yet so I just went and cancelled it without a hitch. Unfortunately there appears to be an issue with the XTM70. I've tried to order it to ship to my home... my work... and the local locker and Amazon says it's unable to deliver but shows the paste in stock.

Weird... never seen this happen.

How is that possible?
It's Amazon, they should be able to deliver to you in the middle of no-where.

I don't believe that is the case but I am going to run benchmarks again once the new bracket and paste are installed. It's my understanding that the 7950x3D throttles at 89C and I haven't gotten anywhere close to that in gaming/productivity... only in an all core burn test did I hit 89C. Everything else as mentioned earlier has been in the 60-70C range.

But Benchmarks have a tendency to push the temperature range well past gaming loads.

Gaming usually doesn't load all CPU cores to 100%. So you might be closer to the 89°C limit than you think.

 

[Deleted member 2838871]

How is that possible?
It's Amazon, they should be able to deliver to you in the middle of no-where.

I know... haha... it's probably a glitch in the system. Item shows in stock but unable to deliver to multiple locations. I order 40 times a month from Amazon and have never seen this. At any rate wasn't a huge deal. I had to run down to Best Buy to pick up some movies and they had it in stock there... picked it up earlier. The XTM70 for $24.99... thanks for those review links I'm definitely looking forward to seeing how it works out with the new bracket.

 

[Kamen Rider Blade]

I believe the issue is that the D15 heatpipes become saturated at around ~220W - anything above that is cooled only by the more limited heat conduction through the copper rather than water evaporation.

This is why I'm very excited about the new cooling tech shown at Computex.

MSI's new Bi-Metallic HeatSink fins are very exciting. That can be a HUGE game changer for Heat Pipe based cooling.

MSI's DynaVC new Vapor Chamber with integral Heat-Pipes is very cool tech that could push CPU & GPU cooling to the next level.

Spoiler: Computex Preview of AlphaCool's new Metal Fans.

If the presented cooling values are to be believed, it's a GAME CHANGER for Case-Fan design and cooling performance.

Spoiler: Alpha Cools claimed Apex Stealth Noise levels

Spoiler: AlphaCool's claimed Volumetric Airflow

If Alpha Cool's to be believed, they just literally "Changed the game" on Case Fan design and EVERYBODY will be copying Alpha-Cool's Metal Case Fan designs by the end of next year to match the performance #'s for a "High-End" SKU for their Case Fans.

Imagine what Alpha Cool could do with a "Tip Magnetic Drive" version of the Apex Stealth Fan mounted on a larger fan than the old YS Tech TMD fan.
The old YS Tech TMD fan was a 70 mm fan running at 5800 RPM from the early 2000's.
That's why it sounded like a jet engine.

Imagine what you could do with the "Sound Dampening" effects of a modern 80 mm / 120 mm / 140 mm / 200 mm Case Fan using the "Apex Stealth" Metal Design + TMD design for a larger case fan aperture.

I'm just salivating at what can be done.

I don't think it's possible to make a Rim Driven Hubless fan that is "Cheap to Produce" w/o MagLev and cheap NeoDymium magnets to hold up the Hub.

Spoiler: Major Hardware - I made a HUBLESS PC fan and it’s so loud!

He tried a home made "HUBLESS PC fan".
It's LOUD AS EFF!!
It's also a First Iteration experiment.

If you're wondering why Rim Driven Hubless fan can't be done for "Cheap?".
1) It'd require a MagLev solution to avoid bearing noise.
2) Any moving bearing on the outside of the Rim would be spinning so fast that you would need fancy materials to make the Rims survive the crazy RPM's exerted on it.

=======================

So I think TMD is the more "Cost Effective" Mass Production solution that needs to be re-explored.

We have SOOO much more cooling knowledge and engineering know-how compared to the early 00's and what they tried.

 

[Kamen Rider Blade]

I know... haha... it's probably a glitch in the system. Item shows in stock but unable to deliver to multiple locations. I order 40 times a month from Amazon and have never seen this. At any rate wasn't a huge deal. I had to run down to Best Buy to pick up some movies and they had it in stock there... picked it up earlier. The XTM70 for $24.99... thanks for those review links I'm definitely looking forward to seeing how it works out with the new bracket.

Be-sure to report back on the "Before & After" for your temps.

 

[Deleted member 2838871]

Be-sure to report back on the "Before & After" for your temps.

Yep... gonna do some before/after tests and will report back with the results.

 

[TerryLaze]

I'm not sure who "all the people" are that are bitching about AMD's CPUs running hot as they are totally reasonable temps for modern performance oriented CPUs. Intel CPUs ran hotter than AMD CPUs for years and few people complained. I recall seeing an egg cooked on an Intel I386 many moons ago.

What you did see was a CPU being cooled by a single egg, and for a pretty long time at that, until the egg wasn't liquid anymore which meant that it lost most of its ability to soak up heat.
An egg will also cook at 95 degrees or at 89.
Also I guess you forgot the giant topics about the intel tim being bad because it wasn't indium but paste, or something along those lines, that only kept the overclockers from going to the extremes but didn't matter to anybody else.

View: https://www.youtube.com/watch?v=xL40UN5p6IY

 

[RodroX]

There are $0 ways to keep the temp on the non-X3D models in check.

Adjusting the way the CPU volatge and/or boost work can lower the temp way over 3°C on the non-X3D models.

I guess for people who want to pursue even lower temps this new noctua frame may be interesting. Perhaps someone trying a no-fan system.

 

[Deleted member 2838871]

what paste do you use?

Be-sure to report back on the "Before & After" for your temps.

I've got the bracket and the paste...

Don't feel like messing with it tonight... long day at work. Will make the change with before/after testing and report back some numbers before the end of the weekend.

 

[Kamen Rider Blade]

I've got the bracket and the paste...

Don't feel like messing with it tonight... long day at work. Will make the change with before/after testing and report back some numbers before the end of the weekend.

Sweet! I look forward to see the change.

Be sure to first try the bracket by itself with your old existing Thermal Paste (Fresh Application if possible).

Then wipe off that old thermal paste and add in the XTM70 to see the difference between your old Thermal Paste and the "New Hotness".

 

[Deleted member 2838871]

Sweet! I look forward to see the change.

Be sure to first try the bracket by itself with your old existing Thermal Paste (Fresh Application if possible).

Then wipe off that old thermal paste and add in the XTM70 to see the difference between your old Thermal Paste and the "New Hotness".

Busier weekend than I thought and didn't have time... planning on tackling this after work today.