Tower vs downdraft cpu cooler

[Macro Device]

they are recommened for Ryzen 5 cpu s

Yet you will gain more if you go for something NOT featuring DirectCU technology. This only works well with single-die CPUs which is almost never a scenario for newest AMD CPUs. So, say, Gammaxx 400 makes your CPU run at 85C and a similar indirect contact cooler will make it 75C. Numbers ain't meant to be accurate, they're just for general understanding.

 

[maxfly]

You are taking 2 isolated "anecdotal" examples and trying to imply that applies to all scenarios. I have been careful to work my replies to include exceptions all around.

The OP asked a very general and generic question. Had the OP asked about one of your specific example compared to a specific downdraft cooler, then sure your examples may be correct. But he didn't.

LMAO. Really? Trying to moderate a mod. Hilarious.

You do realize the c14 and tf2 are among the very best downflows money can buy right? Perfect examples for making comparisons and coming to intelligent conclusions in our, um "general, generic" topic.
Someone speaking from/with hands on experience is exactly the type of information this topic needs. "Isolated or anecdotal" it is not. You may not realize this but you come off as if you have an ulterior motive or worse yet...as if your trolling.

The tower moves the air from front to back, so it complements the airflow in the case.



View attachment 305086

Good point. In the most common types of cases today, front to back flow is how they're designed to perform best. With solid fans this configuration keeps the mb, GPU and m.2 drives supplied with plenty of cooling. Tower coolers do a great job of using that flow to their advantage.

But the Snowman T4 V my Gammaxx 400 are on the same leval and they are recommened for Ryzen 5 cpu s

It depends on your case and specific CPU. If you already have the gammaxx 400, try it out. If it doesn't perform up to your needs or expectations look into something like the Thermalright pa120 or one of their other excellent towers (top notch cooling at a great price), Noctua u12s (great cooling but very expensive) or one of the idcooling towers, (something for everyone and tons of reviews to help you pick) just to mention a few of my personal favorites. The good thing with towers is the selection of high quality units is virtually endless.

 

[eidairaman1]

Downdraft heatsinks have the advantage of being compact and also cooling vrms around them, you can always shroud them to put a 140mm fan on them.

Then there is the downdraft tower

Thermaltake Big Typhoon

Thermaltake Big Typhoon Review

Thermaltake has been around for only seven years, but every one of them has been spent turning the heads of the hardware community. From the original Golden Orb's effective cooling at near silent levels to the current array of water and air coolers, Thermaltake continues to impress with both...

www.techpowerup.com

Scythe Susanoo

Giant from Scythe: Susanoo CPU Cooler Review - XBitLabs

Today we are going to talk about the largest and heaviest processor cooler Scythe has ever made. Just recently the Japanese Scythe Co., Ltd. introduced their Mine 2 CPU cooler for those who value super-efficiency, which we have recently tested and which made a great impression, despite a few...

www.xbitlabs.com

Heatsink – Thermalright

thermalright.com

 

[freeagent]

Thermalright has some nice ones

Low-Profile – Thermalright

 

[tabascosauz]

For all the testing Hardware Canucks does with coolers, the tower vs. downdraft video must be one of the most worthless ones they've ever made. You pit the downdrafts against the towers.........in the very scenario that downdraft coolers despise and tower coolers love. Thanks, Sherlock?

Good point. In the most common types of cases today, front to back flow is how they're designed to perform best. With solid fans this configuration keeps the mb, GPU and m.2 drives supplied with plenty of cooling. Tower coolers do a great job of using that flow to their advantage.

In a big enough, empty case (and easy to cool CPU) I find the airflow to not matter much, but as soon as it's actually filled up with GPUs etc. and cooling meaningfully difficult hardware (100W+) it really starts to show that the C14S, L12, L12S, L9x65 and Big Shuriken 3 start fighting against everything else and send exhaust wherever they friggin please.

More than anything, they really don't like solid side panels unless they also have copious (ie. 10cm or more) free clearance above them to breathe. If you're lazy, you run a side intake and let exhaust go wherever (which isn't possible with a solid panel). If you're super efficient, you run a rear/front intake and a side exhaust (which isn't possible with a solid panel).

I find there was more of an argument for big downdraft in the DDR4 days given that B-die's *actual* performance supremacy (good speeds, tightest timings) was never really challenged. Nowadays, yes Hynix A still likes to be cooled, but it doesn't seem to be a must-have until it gets closer to 7000.

My previous mobo had VRM (with puny small heatsinks) at 100C when I had a tower cooler and after ONLY swapping it to a top-down one, VRM decided to never exceed 67C. CPU (i5-8600K@stock) operated at equal wattages in both scenarios. No-name 120 mm 1500 rpm fan.

I don't know the details here (what board? what load?), but it's 2023. 3+ generations of continuous VRM upgrades later, VRMs even on cheaper boards are a far cry from the state of Z370/Z390/X470 where the vast majority of even high end boards were just flat out garbage.

I mean sure, at the very very bottom of the AM5 or LGA1700 lists there are probably still a handful of fire hazard boards, but it seems like the obvious answer is not to go out of your way (these days) to buy absolute dumpster fire boards.

 

[Greenslade]

Yet you will gain more if you go for something NOT featuring DirectCU technology. This only works well with single-die CPUs which is almost never a scenario for newest AMD CPUs. So, say, Gammaxx 400 makes your CPU run at 85C and a similar indirect contact cooler will make it 75C. Numbers ain't meant to be accurate, they're just for general understanding.

I am very new to all this

Yet you will gain more if you go for something NOT featuring DirectCU technology. This only works well with single-die CPUs which is almost never a scenario for newest AMD CPUs. So, say, Gammaxx 400 makes your CPU run at 85C and a similar indirect contact cooler will make it 75C. Numbers ain't meant to be accurate, they're just for general understanding.

I am pretty new to all this having only had Dell.s and OEM Pc, s for the last 16 years.So never thought about fans etc.And having only built mr first PC eight months ago,I am just learning about all this.It seems like it is a good subject.

 

[A Computer Guy]

Whats your preference Tower or Downdraft cpu cooler? It seems like Down-firing ones are in the minority now. and from the videos on youtube are more for small cases.

In a case like NR200/NR200P with vented side panel (not glass) NH-C14S makes sense and fits and works beautifully (not that there aren't other tower coolers that would work well also). It mostly overlaps the entire area of my ITX motherboard and as an intake it provides some airflow to SSD, RAM, and VRM's and is aided by immediate top exhaust of the case. It handles a stock 5950x just fine but for OC maybe not that great. If you have the case clearance I've read giving it a push/pull fan configuration can help it work a bit better however much more compact down draft coolers won't have that option. Also ducting a drown draft cooler closer to a vented case panel or between the fan and heatsink can help a bit too to ensure your getting the freshest air (assuming an intake configuration).

Some picture to illustrate. In my config I opted to mount the fan to the side panel rail but this leaves a small gap between the fan and heatsink were I need to run a cable. Ideally one would not want a gap like that for best performance.

I think for very specific setups or real SFF cases downdraft coolers can be a good choice (for real SFF mostly the only choice) but in general I read good tower coolers will usually match or outperform most downdraft coolers. I wouldn't really recommend putting a down draft cooler in a case with a solid or glass side panel or in large cases where a better tower cooler can work better. Maybe large cases that already had a side intake fan to feed the cooler would be ok.

 

[tabascosauz]

Some picture to illustrate. In my config I opted to mount the fan to the side panel rail but this leaves a small gap between the fan and heatsink were I need to run a cable. Ideally one would not want a gap like that for best performance.

View attachment 305112View attachment 305113

I think for very specific setups or real SFF cases downdraft coolers can be a good choice (for real SFF mostly the only choice) but in general I read good tower coolers will usually match or outperform most downdraft coolers. I wouldn't really recommend putting a down draft cooler in a case with a solid or glass side panel or in large cases where a better tower cooler can work better. Maybe large cases that already had a side intake fan to feed the cooler would be ok.

The gap between the fan and heatsink is not ideal in the case of the C14S. Fortunately your C14S still looks pretty new and straight, you can probably seal that gap with some foam, tape or other material if you wanted. NCASE and Cerberus both have a very similar side bracket and it has never proven to be that useful in a push config, may as well just mount it on the cooler.

There's consensus that the best setup for the C14S in this layout is a rear intake and C14S as exhaust, though the NR200's wacky rear panel throws a wrench into the works. Honestly, you may as well consider a strong 120mm if you wanna use the side bracket - A14 doesn't like being pressed up against restriction + being close to restriction increases risk of resonance and noise in the A14 + T30-120 generally provides similar performance all around while not having significant area blocked by the bracket.

Though, if you are running a generally lower performance and lower power system (ie. 4750G), it doesn't matter much. In a smaller case I would still give the nod to the C14S since the SS135 and D12L aren't that impressive against it, but where there is room for regular towers like in the NR200, downdraft is harder to justify.

 

[A Computer Guy]

The gap between the fan and heatsink is not ideal in the case of the C14S. Fortunately your C14S still looks pretty new and straight, you can probably seal that gap with some foam, tape or other material if you wanted. NCASE and Cerberus both have a very similar side bracket and it has never proven to be that useful in a push config, may as well just mount it on the cooler

There's consensus that the best setup for the C14S in this layout is a rear intake and C14S as exhaust, though the NR200's wacky rear panel throws a wrench into the works.

Yea the rear panel doesn't jive well for rear intake.

Honestly, you may as well consider a strong 120mm if you wanna use the side bracket - A14 doesn't like being pressed up against restriction + being close to restriction increases risk of resonance and noise in the A14 + T30-120 generally provides similar performance all around while not having significant area blocked by the bracket.

Yea that side bracket isn't the best for the A14. I found a video also suggesting a T30 + A12x25 do seem to be a good combo with C14S.

Though, if you are running a generally lower performance and lower power system (ie. 4750G), it doesn't matter much. In a smaller case I would still give the nod to the C14S since the SS135 and D12L aren't that impressive against it, but where there is room for regular towers like in the NR200, downdraft is harder to justify.

Indeed. I think one also needs to consider the dynamics of the case where you are using the cooler as well. The NR200 is probably more of the exception than the rule where C14S isn't the absolute best but can help the system overall balance between CPU/GPU thermals. Well at least in an example from this older video.

Also suggests swapping out the A14 for A12x25 for better noise profile.

 

[tabascosauz]

@A Computer Guy the side intake C14S only works well there because of the deshroud mod, however, and specifically where GPU deshroud fans are setup as bottom exhaust. Otherwise any axial GPU cooler will dump exhaust heat directly against the side panel, whereupon the C14S happily laps it up. The side intake C14S is by far the easiest and laziest setup, but the consequence is that unless you are able to deshroud like that, temps under GPU load will inevitably suffer.

Machines and More knows his stuff, great SFF content creator especially with air cooling and in the NR200.

 

[Greenslade]

Yet you will gain more if you go for something NOT featuring DirectCU technology. This only works well with single-die CPUs which is almost never a scenario for newest AMD CPUs. So, say, Gammaxx 400 makes your CPU run at 85C and a similar indirect contact cooler will make it 75C. Numbers ain't meant to be accurate, they're just for general understanding.

I did not know what DirectCU technology.was, until i just looked it up.How did we mange before compters Now you can look up anything and find the answer.I was a late starter getting one.

 

[AusWolf]

I go for down-draft for anything except for my main PC, but only because the only down-draft cooler that nips at the heels of my Dark Rock 4 is probably the Dark Rock TF2, but it's just as heavy, if not heavier, and maybe a bit more expensive as well.

Bottom line: VRM cooling is important, except when no down-draft cooler can sensibly handle your CPU.

 

[Greenslade]

I go for down-draft for anything except for my main PC, but only because the only down-draft cooler that nips at the heels of my Dark Rock 4 is probably the Dark Rock TF2, but it's just as heavy, if not heavier, and maybe a bit more expensive as well.

Bottom line: VRM cooling is important, except when no down-draft cooler can sensibly handle your CPU.

I have just been readiing your Small Form facter post.You obviously are a very knowlodgeable person on the subject.It is a very good post. I all ways thought Ryzen,s stock coolers were good.

 

[AusWolf]

I have just been readiing your Small Form facter post.You obviously are a very knowlodgeable person on the subject.It is a very good post. I all ways thought Ryzen,s stock coolers were good.

They are, just not in every situation, and not for keeping your CPU actually cool (unless you have a R3 3100).

Thanks for the praise. I don't consider myself too knowledgeable, though. I just buy way more PC hardware than I should.

 

[Greenslade]

They are, just not in every situation, and not for keeping your CPU actually cool (unless you have a R3 3100).

Thanks for the praise. I don't consider myself too knowledgeable, though. I just buy way more PC hardware than I should.

I bought a ready built Ryzen 5 2600x from someone on eBay last year ,and that has the stock cooler plus it is in a very small case but not a SFF one.I have thought about putting it in a biger case.It is pretty noisey in there,I think that goes for most of us on here.

 

[Deleted member 232534]

I bought a ready built Ryzen 5 2600x from someone on eBay last year ,and that has the stock cooler plus it is in a very small case but not a SFF one.I have thought about putting it in a biger case.It is pretty noisey in there,I think that goes for most of us on here.

the CPU cooler makes the noise, not the PSU or GPU?

i would consider replacing the cooler first provided you can use it in the other pc case you want to try.

 

[Bill_Bright]

I see no use of downdraft coolers in matx/atx.

Did you mean to say "matx" too? Assuming you mean µATX (microATX), many, if not most, µATX cases are small and therefore may not have room for a tall cooler. Now of course, most mid and full tower cases usually support µATX boards too. In that case, then no issues unless a very "slim" case.

I all ways thought Ryzen,s stock coolers were good.

Contrary to what many want us to believe, both Intel and AMD OEM coolers are fully capable of keeping the CPUs they came with "adequately" cooled, with default settings, in a properly cooled case. And note as explained back in post #5, "adequate" does not mean "as cool as possible". Point being, it doesn't have to be more than adequate.

But it wasn't always that way. Years ago, OEM coolers got a bad reputation for being noisy and inadequate, a reputation they sadly, but rightfully earned. Thanks to a lot of bad press and user complaints, both eventually wised up. And both fully understand, for many of their CPUs, their OEM coolers are totally inadequate! And in those products, no OEM cooler is included. You must go 3rd party.

They are, just not in every situation

I totally agree. When case cooling is lacking, changes to the default voltages and timings are more than just very minor, or during extreme taskings, the OEM coolers may not be enough. I will add, however, if the user is doing extreme tasking, then perhaps they should have picked a different CPU in the first place - if the budget permitted. I will also add that "marketing weenies" for both AMD and Intel, as well as the factory computer builders overhype the realistic capabilities of many of those products.

You do realize the c14 and tf2 are among the very best downflows money can buy right?

LOL - You are so eager to criticize another, you don't even understand what you think you are defending! We were discussing potential problems the "weight" being enhanced by the "leverage" created by tall tower coolers hanging off motherboards. That has absolutely nothing to do with how good (or bad) the C14 and TF2, or any other cooler is.

And FTR, no where was I being critical of the C14 or the TF2. They are indeed, excellent coolers - even though some might argue for downdrafts, they too are tall and "top heavy" and "IF" the case is bounced about or transported in a tower orientation, that leverage/weight "may" (I never said "will") put undue stress on the board.

Regardless, taking two specific examples out of 100s of possible products and 1000s of different scenarios does not prove a hypothetical - one way or another.

Clearly there are times when a tower cooler is better suited for the job just as clearly there are times when a downdraft cooler may be better suited.

Regardless, the OP asked, which do we "prefer" among the two categories. He did not ask about specific models - also my point.

A quality tower will last virtually forever. Where as a downflow cooler...nope.

Hogwash!

 

[Greenslade]

Some are more critical, calling them walls of text - often without even reading . For others who complain, well, no one forced them to read them. Still, I admit I may may get a bit wordy, but I try to hard to be factual and informative.

My goal is to share what I have learned over the years before those little gray cells completely deteriorate. Glad you got something out of it. Thanks.


That is true WHEN the tower cooler is oriented that way. Sadly, often users orient the tower cooler so the fan(s) blow at a 90° angle to the air flow provided by the case fans. While sometimes "turbulence" provides better cooling for a discrete device, I feel a "smooth" flow of air through the case is more effective for overall case cooling. That said, I have not tested that personally.

I think you are right there about them not reading them.
I do read all your posts there very intresting
Good for you Bill, keep up the good work you are doing on here.

Yet you will gain more if you go for something NOT featuring DirectCU technology. This only works well with single-die CPUs which is almost never a scenario for newest AMD CPUs. So, say, Gammaxx 400 makes your CPU run at 85C and a similar indirect contact cooler will make it 75C. Numbers ain't meant to be accurate, they're just for general understanding.

I get your point there about numbers that is

Beginner Micro Device ​

the CPU cooler makes the noise, not the PSU or GPU?

i would consider replacing the cooler first provided you can use it in the other pc case you want to try.

The CPU cooler yes. But there is a lot of dust in there ,i have hardly used it since i got it and not got round to getting my air duster out.I am trying to get an OS on ECS KN1SLI
Extreme .It goes through the motions of getting XP on there and goes off and starts once again to get XP on there ,i give up on ,it. The board was unused i bought if for a fiver at the local market.The cpu fan is as noisey as hell,and going like the clappers.The board failed to work with the gpu on 642 and yet it worked on another board weird stuff.This is the only fully built PC i have done so far , the CPU fan is really quite and was only £8 .I did not know about getting a speciel fan before this.It ws only since seeing videos of the Snowman fans which people rave about on you tube It has
a Xeon E3-1246 V3 and the 1070.

 

[A Computer Guy]

Contrary to what many want us to believe, both Intel and AMD OEM coolers are fully capable of keeping the CPUs they came with "adequately" cooled, with default settings, in a properly cooled case. And note as explained back in post #5, "adequate" does not mean "as cool as possible". Point being, it doesn't have to be more than adequate.

I have 4750x and 3800x (running in 65w TDP) runs just fine with wraith stealth cooler. 3950x with Wraith Prism. Sure they can get a bit warm like 80c at full load but they are usable when you don't have a better cooler around. (Disclaimer: A caveat to my given examples is I wasn't dumping any meaningful GPU heat into the system during my observations of those combinations.)

 

[nomdeplume]

In a big enough, empty case (and easy to cool CPU) I find the airflow to not matter much, but as soon as it's actually filled up with GPUs etc. and cooling meaningfully difficult hardware (100W+) it really starts to show that the C14S, L12, L12S, L9x65 and Big Shuriken 3 start fighting against everything else and send exhaust wherever they friggin please.

Big Shuriken/NF-A15 has been a nice pairing for my admittedly low end build. I let the room warm up around 100 F and it still keeps temps low (34 C) with hardly a noise.

In a perfect world I'd have another inch of memory clearance that allowed using Ripjaws or Trident Z. Basically any heat shield is a no go which greatly limits your options.

 

[Bill_Bright]

I let the room warm up around 100 F

Yikes! I would melt.

it still keeps temps low (34 C)

Sorry, but that is impossible. No fan based cooling system is capable of cooling any non-living object cooler than the ambient temperature. 34°C = 93.2°F.

"Feels like" or the "wind chill factor" only affects living flesh, not hunks of metal and silicon.

 

[Greenslade]

Obviously, below a certain height clearance, downdraft is all you can have. But they basically top out around NH-U12S level performance, so there aren't any downdraft coolers that offer real high end performance (nothing beyond the C14S and TF2). You inherently lose some cooling performance compared to a tower, at comparable heatsink mass + heatpipe count + fan config.

I used a C14S for a long time and have basically exhausted all its potential. On the test bench it's still useful, but in an actual case where height clearance doesn't strictly prohibit more powerful single and dual towers, it doesn't offer much.

Whether in a SFF or normal case, downdraft also requires a lot of planning and attention to airflow to maximize performance. It's not enough just to assume that a conventional front intake rear exhaust is enough.

The VRM and RAM cooling benefits aren't that significant anymore. A single push fan (ie. stock C14S) doesn't provide any meaningful amount of airflow. When VRM quality and performance is shitty enough that VRM cooling *actually* matters, a C14S or TF2 will never be nearly as effective as just strapping a small ghetto fan to the VRMs, and will never save a VRM doomed to fail. When doing mem OC at high VDIMM on temp sensitive ICs, a dedicated fan is way more effective than a C14S or TF2 that just heats up the sticks with CPU exhaust under load (e.g. TM5 or Linpack where CPU uncore load is significant).

Considering how heavy the C14S (esp. push-pull at about 1.2-1.3kg) and TF2 get, I don't see a reason to consider them "healthier" or less strain on a motherboard.

Have the Snowman CPU coolers been reviewed on here

tabascosauz ?​

 

[Shrek]

I let the room warm up around 100 F

I also expect my computers to tolerate a non air-conditioned room, but I'm normally below 90°F

 

[A Computer Guy]

Yikes! I would melt.


Sorry, but that is impossible. No fan based cooling system is capable of cooling any non-living object cooler than the ambient temperature. 34°C = 93.2°F.

"Feels like" or the "wind chill factor" only affects living flesh, not hunks of metal and silicon.

You were oddly specific there "any non-living object" ?

 

[Dan.G]

I prefer tower coolers - I believe they exhaust the hot air more efficiently. My MB doesn't have VRM cooling, but I do have a 120 mm static pressure fan (I think, attached image) that blows air over them, from the front (small case). Trying to have the same airflow direction inside the case.