Looking for a good AM4 cooler with size restrictions

[Bill_Bright]

I use HWinfo, and that’s the temp from the CPU diode.

And I am just those cost about 1 penny for 100 of them. While they are usually fairly accurate, I have also seen them reports temps so far off, it would defy the Laws of Physics if true - all while the computer was working just fine. That's why I suggested blasting a desk fan in there to see what happens. It is also why I asked about idle temps.

I think the only exhausts are dual 80mm.

If your PSU pulls air from inside the case, that will help (though typically not counted as part of case cooling). And 2 x 80mm fans provides almost the same area a 1 x 120mm so that's not bad. What is your system temp? If high too, then that could indicated insufficient case cooling too.

And for the record, Darmok was the best ST:TNG episode ever! Period. End of discussion.

 

[danbert2000]

I have the Shadow Rock LP, it's pretty great but you have to make sure your RAM works with the low clearance. I replaced the fan with a Noctua NF-A12 and it is much cooler and quieter now.

 

[kapone32]

I use HWinfo, and that’s the temp from the CPU diode. When I had a 150W tower cooler on this 2400G, it never got above the 70Cs, but with the stock cooler, temps keep climbing, and I get nervous once it’s at 85C. I know 95C is max, but I can’t bring myself to let it get that high!


I think the only exhausts are dual 80mm.

That is understandable given Mini iTX

 

[silentbogo]

. While they are usually fairly accurate, I have also seen them reports temps so far off, it would defy the Laws of Physics if true - all while the computer was working just fine.

That reminded me of AMD sFM2 temperature sensor fiasco. I don't think I've ever been able to get the correct temperature reading even within 20C of truth. Even in Crimson or Adrenalin.
I don't think even AMD's engineers know exact offset values and whatever magic multipliers they've set for each CPU model

 

[Bill_Bright]

That reminded me of AMD sFM2 temperature sensor fiasco. I don't think I've ever been able to get the correct temperature reading even within 20C of truth. Even in Crimson or Adrenalin.
I don't think even AMD's engineers know exact offset values and whatever magic multipliers they've set for each CPU model

It does not help us consumers that there is no industry standard for sensor placement for CPUs. They could be anywhere under the IHS. Even among CPUs within the same brand, it seems both AMD and Intel do it differently depending on which processor family they are dealing with.

But I guess, as far as the consumer is concerned, as long as the integrated thermal protection features consistently kick in to avoid thermal damage, then we should be happy. I wish I could point my laser guided infrared thermometer at the CPU die and get an accurate and useful reading. But that's pretty hard to do with a big block of copper and aluminum sitting on top of it.

 

[Darmok N Jalad]

Yeah, even if I can make myself comfortable with letting my CPU bump Tjunc, that still means I’ll be running into throttling and reduced performance. It does just seem limited to the stock cooler. Something with more mass seems to have no trouble keeping temps reasonable, even with an overclock.

I have the Shadow Rock LP, it's pretty great but you have to make sure your RAM works with the low clearance. I replaced the fan with a Noctua NF-A12 and it is much cooler and quieter now.

That was one I was considering, but I do have tall RAM sinks that I know would interfere, and I wasn’t sure if space was an issue if alternately installed. Can’t that be rotated to be over the VRMs instead?

Turns out if I’m willing to remove the big 200mm intake fan, I can install a 120mm or 140mm radiator in this case. Still not sure if that is ideal, as it looks like the fan would dump the heat into the case. Still, it’s impressive that one would even fit inside a mini-ITX case.

 

[Darmok N Jalad]

Just to update my own thread with an answer, I finally made a purchase. I went with the Deepcool GAMMAXX 300. It was less than $20, comes in just under 140mm, and doesn’t interfere with my tall RAM-sinks on my mini-ITX build. It’s rated for 130W, and was pretty easy to install. Loading my 2400G under prime95 now levels off around 70C, where before temps were well into the mid-80Cs and climbing with Wraith Stealth. Definitely quieter too.

 

[Bill_Bright]

One thing I have noticed over the years with very tall and heavy "side firing" coolers mounted on ("hanging off?") vertically oriented motherboards is stress damage/fractures around the coolers mounting points and around the motherboard mounting points. This is likely due to bounces and rough handling and the very heavy weight of the cooler pulling on those mounting points during transport of the computer.

To prevent this irreparable damage, when moving the computer - even outside for cleaning - try to immediately rotate the computer so the motherboard is horizontally oriented and the heavy cooler is now just sitting on top of and pressing down on the motherboard.

If you will be transporting the computer carry it and place it in your car in this horizontal orientation. If the computer will be out of your sight and someone else will be moving it, I recommend you remove the cooler and transport it separately. Then at your final destination, thoroughly clean the mating surfaces, apply a fresh new layer of TIM and re-mount it there.

Note I have never seen this sort of damage with the lower profile downward firing coolers.

 

[Darmok N Jalad]

In my case, the motherboard is installed horizontally, so there is no stress from gravity. Even so, it’s really not that heavy of a cooler.

 

[John Naylor]

just questions, have you tried lapping the current heatsink or improving the airflow?

CPU IHS surfacea are not flat, they are curved ... and CPU heatsinks are manufactured to meet that curve. Lapping removes the curve and voids warrantees. According to manufacturers, the curve actually helps the spread the TIM. All the before and after tests I have seen have proved inconclusive, sometimes things improve, sometimes get worse, but the magnitude of the change is lass than the standard deviation you get just remounting 10 times..

 

[Bill_Bright]

Even so, it’s really not that heavy of a cooler.

The point is stresses during transport when the case gets jostled about by the user, or when the vehicle flies over railroad tracks. Even a light weight cooler can leverage a lot of weight (and therefore stress on the mounting points) on vertically oriented boards under those transport conditions.

 

[John Naylor]

I actually think it would fit, too, based on my own measurements. I think if it comes in under 150mm, it will fit, but just be tight.

The case already has a massive intake fan. Biggest fan I’ve ever seen, and there are vents everywhere. Plus this computer is in my basement and it’s now wintertime. I won’t see better conditions than what I have now. I’m not sure why the WS cooler lets temps get so high. I know the Tjunc is 95C, but it seems like the cooler is more than happy to let the CPU ramp all the way there.

I have tried a 150W tower cooler on this CPU before, and it does a much better job keeping temps down in the 60Cs, even with an OC. The problem is, the cooler doesn’t fit in my case! The AMD cooler is installed correctly, and under gaming temps are more reasonable. But Handbrake sends temps into the 80Cs.

I actually think it would fit, too, based on my own measurements. I think if it comes in under 150mm, it will fit, but just be tight. .... The case already has a massive intake fan. Biggest fan I’ve ever seen, and there are vents everywher e..... .

Well if ya can squeeze in 150mm, then there's no better choice than the $45 Scythe Fuma Rev. B Beats the $90 flagship models from Noctua and Cryorig as well as most AIOs

https://tpucdn.com/reviews/Scythe/Fuma/images/temp_oc_aida64.png

That's a 200mm fan .... and those open grilles everywhere are a critical part of case cooling. Consider the following:

1. Fans will actually produce about 50% of the air flow stated on the packaging
https://martinsliquidlab.wordpress....w-specs-are-poor-measures-of-fan-performance/

2. Due to air inlet filter restriuctions, depending hoy dirty ya let the filters get, you will need at least 1.33 to 1.50 times more "rated" flow in than out. Thats what those rear grilles are for. While everyone knows about dust in negative pressure situations , the more important fact is when thi is presnt, you are sucking in hot GFX card and PSU exhaust in thru those rear grilles.

 

[Bill_Bright]

CPU IHS surfaces are not flat, they are curved ... and CPU heatsinks are manufactured to meet that curve.

Sorry, but not really true. Yes, the IHS surfaces for many CPUs are not flat. But the facts are, AMD processor IHS surfaces that are curved tend to be convex (higher in the middle) while Intel's are concave (lower in the middle).

A tiny bit of homework with our friend, Bing Google would show this. To make it easier to see, Tom's Hardware illustrated it here.

Because most aftermarket coolers are designed to support both AMD and Intel processors, including that Scythe Fuma and that DeepCool, they cannot be manufactured to meet both a convex and a concave curve - so they aren't. Aftermarket coolers are manufactured to be flat.

 

[sneekypeet]

Aftermarket coolers are manufactured to be flat.

As a guy who see's 100s of CPU coolers, most cooler bases are not flat. Most aftermarket CPU solid base coolers (even AIOs) are indeed convex. Most of the heat pipe direct touch coolers tend to be flatter. Put a straight edge against most bases and it is easy to tell right away as the light shines through on both edges and not the middle.

 

[Bill_Bright]

I think that mostly illustrates how imperfect they are - not intentional design - especially for those designed as [more or less] "universal" (supports both AMD and Intel).

As for heat pipe direct touch - like the OP's DeepCool choice - that style is what we are seeing most often lately in aftermarket coolers. And as you said, they tend to be pretty flat. Not sure if that style is more for efficiency, or cost by using less copper overall.

 

[sneekypeet]

that style is what we are seeing most often lately in aftermarket coolers.

In what I see, maybe 1 in 10 is an HDT cooler.

I also think the convex design is intentional, as with proper mounting pressure, it makes for better contact against any IHS, as it is "forced" to do so with the high pressure mounting most coolers require these days. All bases are milled/machined. I am near certain that the bases could be made flat as easily as they could convex, and you gain nothing by making them convex, as it has more copper in the base than a flat one would.

 

[Bill_Bright]

and you gain nothing by making them convex

Not sure I understand why they don't machine them flat either. These days with computer controls and precision laser guidance, even on the budget coolers, they could do it.

I have never seen a published spec on the curvature but for the best transfer of heat, the arc of the curve in the IHS and the arc in the heatsink would have to perfectly match. Not only that, both AMD and Intel would have to agree on that! So flat just makes more sense.

 

[eidairaman1]

Those direct heatpipe coolers need the line method of Thermal compound for best effect.

 

[Bill_Bright]

Those direct heatpipe coolers need the line method of Thermal compound for best effect.

Have you seen a study on that? I ask because if you look at the manuals for some (that Scythe Fuma, for example) they just put a blob on it. And the video on the DeepCool site either used no TIM, or had it spread evenly across (hard for my eyes to tell).

I just do it the way I always to - a small blob spread evenly across the die. I don't like having the pressure from the clamping mechanism do the distribution.

 

[sneekypeet]

Line method will provide better temperatures on the coolers where the base and the pipes have gaps. Reason being, much of the TIM will push into the gaps rather than across the IHS. You know, has to do with that perfection thing you speak of. Can be in the order of 3-5 degrees depending on how poor the build quality is.

 

[eidairaman1]

Have you seen a study on that? I ask because if you look at the manuals for some (that Scythe Fuma, for example) they just put a blob on it. And the video on the DeepCool site either used no TIM, or had it spread evenly across (hard for my eyes to tell).

I just do it the way I always to - a small blob spread evenly across the die. I don't like having the pressure from the clamping mechanism do the distribution.

Watch Time Frame 3:14+

@Bill_Bright
Here is my manual
ftp://ftp.scythe-eu.com/Manuals/CPU-Cooler/Ashura_Manual(new).pdf

Line method will provide better temperatures on the coolers where the base and the pipes have gaps. Reason being, much of the TIM will push into the gaps rather than across the IHS. You know, has to do with that perfection thing you speak of. Can be in the order of 3-5 degrees depending on how poor the build quality is.

Please note Both Fuma and Ashura Manuals state Apply Thermal Grease evenly.

For my builds that have a base plate hsf I use the spread method whether ihs or raw die.

Direct heatpipes i use the line.

 

[Darmok N Jalad]

I know folks here might faint when they read this, but I just used the preinstalled TIM on the new HSF. It was actually checkerboarded on the base instead of a solid coating like AMD’s WS cooler had. Like I said, so far so good, temps are much better managed now.

 

[NdMk2o1o]

I know folks here might faint when they read this, but I just used the preinstalled TIM on the new HSF. It was actually checkerboarded on the base instead of a solid coating like AMD’s WS cooler had. Like I said, so far so good, temps are much better managed now.

I've used the stock TIM on heatsinks before, totally acceptable unless it's really poor or if you want to get the absolute best out of your cooler then you can probably shave another 3-5c if your lucky with some aftermarket stuff, on the whole most of works fairly adequately though.

 

[sneekypeet]

I know folks here might faint when they read this, but I just used the preinstalled TIM on the new HSF. It was actually checkerboarded on the base instead of a solid coating like AMD’s WS cooler had. Like I said, so far so good, temps are much better managed now.

By all means use the factory application, as long as there is no dust/debris in it. Typically OEM pastes are equivalent to MX2.

 

[Bill_Bright]

Line method will provide better temperatures on the coolers where the base and the pipes have gaps. Reason being, much of the TIM will push into the gaps rather than across the IHS.

??? That would depend on the direction you lay down the lines, the placement of those lines, and how much TIM is applied. I don't see where it makes a difference as long a full coverage is achieved, no insulating air is trapped between mating surfaces, and excess TIM is kept to a minimum. And to me, that is the same criteria regardless the method used.

I know folks here might faint when they read this, but I just used the preinstalled TIM on the new HSF.

10 -15 years ago, it would have been frowned upon to use the OEM TIM. But cooler and CPU makers heard the complaints. So not only are today's OEM coolers much more efficient (and quieter) but the OEM TIM is much more efficient these days too.

These days, I use the OEM TIM and only use after market TIM if I have to pull the cooler for some reason (never reuse old TIM).

then you can probably shave another 3-5c if your lucky with some aftermarket stuff

That's the thing. If you "need" that extra 3 - 5°C to keep from reaching or crossing the thermo-protection thresholds, you have other pressing matters in need of your attention first - like case cooling or the CPU fan itself.

Watch Time Frame 3:14+

Yeah, I have watched that before. Looks like they are using 10 times too much to me. TIM should NOT be a layer sitting between the mating surfaces. The best (and fastest) transfer of heat occurs with direct metal to metal contact. All the TIM needs to do is fill the tiny gaps and microscopic pits and valleys in the mating surfaces so insulating air cannot get in. Any excess is in the way regardless how efficient the TIM is. Not to mention, if you apply too much, you risk it getting squished out and onto some place it does not belong.