Best Cooling

[bit_user]

Here's the only long term testing conducted that I know of.

...

Summary, IMHO: filters, filters, filters!

Ugh. Those clowns ran positive pressure with some of their intake fans (the top ones) unfiltered. The result was predictable from the day their "experiment" began.

I guess Linus doesn't care, as long as he gets views. Fell into that trap, I did. And controversial videos generate even more views, which economists refer to as a "perverse incentive".
: (

 

[Darkbreeze]

In a negative pressure setup, it also gets in through other cracks and holes. Positive pressure only helps with dust buildup if you have dust filters in front of all your intake fans.

This basic principle applies not only to computer cases, but to clean rooms as well. They are positively-pressurized so that unfiltered air can't leak in and bring contaminants.

In hospital wards where patients might have infectious diseases, they do the opposite and run negative pressure in those areas. This is so that infectious microbes don't leak out into the general population areas.

What I'm saying isn't new. You can find other people with the same experience as me, with a little searching.

​

View: https://www.reddit.com/r/buildapc/comments/2rike2/discussion_how_is_positive_pressure_supposed_to/​

​

I'm familiar with positive and negative pressure aerodynamics and principles. And you are right, in truth. I was just kind of being facetious when I said air can only enter through the fans. No worries.

 

[Darkbreeze]

Ugh. Those clowns ran positive pressure with some of their intake fans (the top ones) unfiltered. The result was predictable from the day their "experiment" began.

I guess Linus doesn't care, as long as he gets views. Fell into that trap, I did. And controversial videos generate even more views, which economists refer to as a "perverse incentive".
: (

Not only was their result predictable, it was wrong from the start because not only did they not have filtration on those fans, they didn't have filtration on those fans by INTENT of the manufacturer because it is NEVER a best practice to configure the fans located where the hottest air in the case is to be intake fans anyhow. Top fan locations in any standard tower case, and honestly, in pretty much any kind of enclosure you want to look at, should be configured as exhaust in 99.99% of instances, as I know you know. So why they were doing that is senseless and a waste of time. Not to mention the fact that dust in the air tends to settle downward, so any dust settling to the top of the case is of course ALSO going to be sucked in AND it is likely that there will be MORE dust present at locations on top of the case than in front or below it, so they were not only using a poor configuration for cooling, but also for dust mitigation. Clowns was the right word too, because that's exactly what I feel like they are. Good at marketing to the unwashed masses, but not particularly bright as compared to even most of the veteran members around here.

 

[bit_user]

Not to mention the fact that dust in the air tends to settle downward, so any dust settling to the top of the case is of course ALSO going to be sucked in AND it is likely that there will be MORE dust present at locations on top of the case than in front or below it, so they were not only using a poor configuration for cooling, but also for dust mitigation.

Well said. I thought about adding something about that. It was just a hilariously clueless experiment. They should have at least understood the principle they were trying to test, but I long-ago realized these folks actually aren't as incentivized to "get it right" as some people might assume.

I wonder if it's a phenomenon particular to video, since you can distract from a lot of fuzzy thinking by being engaging and charismatic. Something a lot harder to do in writing. And in the writing of an article or technical blog post, you have to structure your thoughts and provide more explanations during which you're more likely to catch some flawed assumptions or erroneous conclusions.

I think the problem is only likely to get worse, since TickTock seems to favor even shorter videos and that means cutting out even more theory and explanation. Even for videos still hosted on Youtube, if the audience is becoming accustomed to shorter videos, Youtube creators will probably react by making their videos shorter as well.

 

[Darkbreeze]

They already do. See: Youtube shorts. It's becoming a popular thing now.

 

[AndrewJacksonZA]

wonder if it's a phenomenon particular to video, since you can distract from a lot of fuzzy thinking by being engaging and charismatic. Something a lot harder to do in writing. And in the writing of an article or technical blog post, you have to structure your thoughts and provide more explanations during which you're more likely to catch some flawed assumptions or erroneous conclusions.

I think the problem is only likely to get worse, since TickTock seems to favor even shorter videos and that means cutting out even more theory and explanation. Even for videos still hosted on Youtube, if the audience is becoming accustomed to shorter videos, Youtube creators will probably react by making their videos shorter as well.

It worries me that the most visible representation of the Gen Z crowd is one where they don't WANT to read and want to be automatically spoonfed by The Algorithm™ of the platform of their choice, mostly tiktok. So they are, basically, IMHO, surrendering their brains to the sway of whoever is programing the recommendation engine. :-(

Read this thread from the top. I felt sad for the future watching these people's responses in the video.

https://twitter.com/i/web/status/1557885524295270402

View: https://twitter.com/AdriSheares/status/1557885524295270402?t=Ne7zGtuA2uVo2lL2SR9oSw&s=09

 

[AndrewJacksonZA]

Fell into that trap, I did.

Do you know of any other long term experiments for this please? Not theory, documented experiments please.

 

[bit_user]

Do you know of any other long term experiments for this please? Not theory, documented experiments please.

Good question. Unfortunately, I don't have any references I can provide. That's not to say they don't exist, but I'd be searching for them just like you.

There was once an article I saw which showed an example of positive pressure directly resulting in lower CPU temps, but it might've been case-specific and was possibly sponsored by a case manufacturer. It was a long time ago and I've not been able to find it, in previous searches.

For me, my own experience is good enough. I once had a Pentium VI in this room. It had a Zalman "flower"-style cooler, like this:

I had to dust the thing with compressed air, between 1-2 times per year, or else I'd get CPU overheating alarms. That thing was a major dust trap!

Today, I have a machine in the same room (slightly different location), with a CPU that burns a similar amount of power. Bigger heatsink and slower fan, but the prior experience shows there's plenty of dust in this room. Also, there's lots of dust on top of the case, as added proof. Yet, there's none inside! It's pretty surprising just how clean the inside stays. The case & PSU are both black, so I'd see the dust if it were there.

The newer machine has got a 140 mm intake fan and I think a 90 mm exhaust. I forget the fan models, but I checked their CFM to ensure the case would maintain positive pressure. The PSU is an 860 W model with a fan that doesn't spin up until it hits at least 50% utilization, which means it spends most of its time spinning slowly or not at all.

 

[Darkbreeze]

There are plenty of experiments on this although most of the better ones were on sites that are defunct now. The bottom line is that positive pressure is the best configuration for dust suppression and negative pressure is the best configuration for cooling performance, SO LONG AS there is sufficient airflow to adequately achieve the desired level of cooling performance to begin with.

So, for example, in a case with a fairly high TDP processor and graphics card, where perhaps both are overclocked, if you have only a single rear exhaust fan you will have a negative pressure configuration BUT that alone does not mean you are going to have good cooling performance because you do not have adequate airflow, regardless of the pressure arrangement, to provide the required influx of cooler ambient air to the graphics card and CPU coolers. It will however have better cooling performance than, say, just a single intake fan, which by itself (So long as there are no massive leakage points in the case) would tend towards a positive pressure arrangement.

So long as there is enough overall airflow to feed the coolers and offer some residual airflow over other components, a negative airflow arrangement will offer better cooling performance because not only will there be a higher exchange rate of air through the case but there will also be a more immediate removal of trapped stack effect heat. On the other hand, with a positive pressure arrangement there will be a reduction in the rate of exchange within the case because there is less heat being forcibly exited from the enclosure and the air that cannot escape will spend more time in the enclosure where it will tend to become hotter since it cannot get out as easily.

I like a balanced or slightly positive pressure configuration. Positive pressure can still mean more than adequate cooling performance while still suppressing dust from entering the system so long as you are creating the positive pressure arrangement by ADDING an additional intake fan and not by removing a recommended exhaust fan. For the vast majority of cases out there, having two exhaust fans, one in the rear location and one in the top-rear location, and having two or preferably three intake fans in front, is more than sufficient to keep practically any hardware configuration cool so long as the actual coolers in use are capable enough for the hardware.

Most custom systems these days have overkill or poorly configured case cooling and CPU coolers that are not as capable as they ought to be for the hardware and configuration.

https://www.tomshardware.com/reviews/cooling-air-pressure-heatsink,3058-5.html

https://www.overclockers.com/dust-prevention-positive-pressure-negative-pressure-does-it-really-matter/

The Big Air Cooling Investigation | bit-tech.net

Ever wanted definitive answers on how best to air cool your PC, then look no further

bit-tech.net

Site Suspended - This site has stepped out for a bit

 

[HkrYq]

Best 360mm Closed-Loop Liquid Cooler With a Display :
MSI MEG CoreLiquid S360? You didn't even consider testing the Corsair's H150i Elite LCD ? It might have the same performance as H150i Elite Capellix tho since it uses the same fan.

 

[Darkbreeze]

So, I know they regurgitate this article to simply include updates, which makes commentary sometimes seem "off" because it might not still be relevant, but I just commented a few months ago on this and feel the need to again.

You NOW have the Thermalright Peerless Assassin 120 SE as the #1 "Big air CPU cooler" on the list, and once again, Matt, this is just (insert favorite expletive here, if you like) ludicrous. And, it's not just an opinion this time, although, I didn't bother to bring the empirical evidence because, well, I'm too damn lazy these days. I might have to do it though.

So, how do you put that cooler in #1, when you KNOW damn well it can't compete with the Deepcool Assassin III, Noctua NH-D14/15/15s, or any of the other big coolers? I would REALLY like to see some direct comparison, empirical data, that show this, or else this cooler (Much as I am a HUGE fan of Thermalright and use AND recommend them more than almost anybody else on the TH forums) needs to not be in the top spot unless the top spot is "best performance for the price" in which case I'd agree. Purely top performance though, not so much.

I've just taken a 12700k with a Noctua NH-D15 on my bench, which was only using that cooler temporarily while I was setting up the OS installation and applications for this owner, and replaced it with the Peerless Assassin 120 SE, and I even pulled it, cleaned it again and did the installation over just to be sure, and it simply didn't even come close to the performance of the D15. There was like an 8°C difference. So then I tried the NH-U14S I have on hand as well, and it was close, very close, but the U14S lost to it. But barely, and it's so much smaller overall. But bottom line is, unless it IS "bang for the buck" performance, which the article simply does not say, then it at the very least it's misleading. At worst, it's simply "flavor of the week" journalism.

I understand the need to keep things fresh, but I also understand that this is a long time well respected site when it comes to reviews, or at least it used to be, and it doesn't feel like we are living up to that standard in terms of being accurate when we say "best" of something, and are maybe losing sight of being accurate in our reporting. And I think I'm being lenient in that regard.

There is nothing wrong with the writing. The problem is obviously coming from a higher position than that I think. I'm just saying. Let's get back to at least being reasonably accurate when we say "best" about something. Even if it's just categorically accurate.

 

[bit_user]

negative pressure is the best configuration for cooling performance, SO LONG AS there is sufficient airflow to adequately achieve the desired level of cooling performance to begin with.

That's not what one study showed. Can't find the link, but they claimed positive pressure reduced the number and magnitude of hotspots. Sort of makes sense, if you think about it, because positive pressure should result in more chaotic airflow, whereas negative pressure is going to simply stream air from the main intakes to the output fans.

That said, I seem to recall the study involved only one case. And I doubt they tried more than one negative or positive pressure configuration. So, drawing generalizations from it is a bit risky.

a negative airflow arrangement will offer better cooling performance because not only will there be a higher exchange rate of air

Huh? No, you're presuming positive pressure is accomplished by lowering the exhaust rate! If you hold the exhaust rate constant and simply increase the intake pressure, then the exchange rate should actually go up slightly (i.e. because the efficiency of the exhaust fan will improve if it's not having to work as hard).

https://www.tomshardware.com/reviews/cooling-air-pressure-heatsink,3058-5.html

The claims in this article aren't supported with data.

https://www.overclockers.com/dust-p...sure-negative-pressure-does-it-really-matter/

I think this is flawed, probably because the filter they used is too coarse for the size of the sawdust particles they used. Now, you may say: "well, dust particles in the air are even smaller, so that should be worse", however what I think happens is that the larger dust particles (hair & fibers) would normally form a finer mesh that helps trap some of the smaller particles.

The reason I'm so confident about positive-pressure is that I've used positive vs. negative on 2 machines in the same room. That's how I know the benefits of positive pressure + intake filters are no lie.

https://scholarlygamers.com/tech-talks/2017/11/14/building-pc-positive-vs-negative-air-pressure/

This article matches my experience:

"Previously I had purposely created a negative pressure environment, as I believed this would give me the lowest temperatures possible. Sure enough the system ran nice and cool, but I found myself having to clean my PC quite often due to the accumulation of dust. After changing out a few components (replacing the huge fan on the top of the case with a window), installing filters on all of the intake fans, and configuring my fans so I had more air entering the case than exiting; my system still has not required any dusting since. The ambient temperature within the case has only risen by a degree or two, and the end result is much more optimal for my configuration."​

 

[Darkbreeze]

Lot of assumptions there. But ok, whatever works for you man.

 

[bit_user]

Lot of assumptions there. But ok, whatever works for you man.

It's fine with me if you're done with the subject, but we could focus on two specific claims you made:

1. "negative pressure is the best configuration for cooling performance, SO LONG AS there is sufficient airflow to adequately achieve the desired level of cooling performance to begin with."
2. "a negative airflow arrangement ... because not only will there be a higher exchange rate of air"

If you start with a negative-pressure setup and then add more intake flow, what's your theory on why it hurts cooling performance? I've already offered my hypothesis that positive pressure should help by creating more turbulent airflow, which I expect should then reduce hotspots.

On point #2, if you make the configuration into positive-pressure by simply increasing intake, why wouldn't that result in higher exchange rate? Adding intake should should reduce load on the exhaust fan, slightly boosting its output. Hence, higher exchange-rate! Right?

 

[zailscode576]

You guys forgot the phantom spirit 120, it's like the peerless assassin 120 but better. (from what I've seen)

Id really appreciate if i could get a reliable source to benchmark the cooler since currently i only have random youtube benchmarks from lesser-known channels. and also myself since i have one.

 

[Darkbreeze]

You guys forgot the phantom spirit 120, it's like the peerless assassin 120 but better. (from what I've seen)

Id really appreciate if i could get a reliable source to benchmark the cooler since currently i only have random youtube benchmarks from lesser-known channels. and also myself since i have one.

I'm sure they'll jump right on that. LOL.

 

[belimo]

As per Matt Safford's advice, bought the PA120 SE CPU cooler and loving it! Thank you Matt and to the team of tom's HARDWARE!

 

[mcdonnell155]

Does any company make R410-A refrigerated CPU cooler for e.g. (dual overclocked) Xeon boards? Would love to try 5GHz at double BCLK (extra 2-3 cycles of RAM latency).. Edit: I have a FinTek superio that turns CPU fan off at low temps, shouldn't worry about AC mode.

 

[bit_user]

Does any company make R410-A refrigerated CPU cooler for e.g. (dual overclocked) Xeon boards? Would love to try 5GHz at double BCLK (extra 2-3 cycles of RAM latency).. Edit: I have a FinTek superio that turns CPU fan off at low temps, shouldn't worry about AC mode.

We see coolers incorporating peltier devices pop up every few years, but the're remarkably inefficient and don't tend to work out all that well.

Perhaps you should look into a custom-loop water cooling setup with an external chiller.

 

[mcdonnell155]

We see coolers incorporating peltier devices pop up every few years, but the're remarkably inefficient and don't tend to work out all that well.

Perhaps you should look into a custom-loop water cooling setup with an external chiller.

I have a Prescott IBM with a massive aluminum block as a CPU cooler, I think it will work. Once saw tiny AC compressors on Aliexpress, probably too small though.

 

[bit_user]

I have a Prescott IBM with a massive aluminum block as a CPU cooler, I think it will work. Once saw tiny AC compressors on Aliexpress, probably too small though.

You mean you'd essentially use a custom loop with Freon instead of water?

I've never heard of that, but I'd imagine it's possible if you know what you're doing and design a system with the right coolant, pressure, and boiling point. That said, I assume there are good reasons it's not done.

 

[RyanKent]

NH-U14S product page The dimensions listed for the Noctua NH-U14S cooler are incorrect. As per Noctua, it is 165mm high, not 171mm. From the manufacturer: "Standing 165mm tall, the NH-U14S is short enough to fit most modern high-end cases"

 

[Albert.Thomas]

That said, I assume there are good reasons it's not done.

condensation is why it's not done. You don't want a freon based cooling system anywhere near a PC unless it's been fully insulated.