Radiator Fan Orientation - The Workshop

[nicklmg]

Amazon: http://geni.us/1DE8
NCIX: http://bit.ly/1o7ZkIA

 

Push? Pull? Fans on top? Fans on the bottom? Does any of it really matter? Let's find out!

 

 

 

[Bajantechnician]

first

[ddrkreature]

Funny. I was forced to do the pull air from the top method because of my case/motherboard layout. Good to know I was forced into a preferred method

[Shahnewaz]

I'd like this test re-done with something that kicks out A LOT of heat, like a dual GPU, and running the test for an hour or two instead of 10 minutes. That might make a significant difference.

[Jarf10]

Does anyone know which case he used?

[iPolymer]

6 minutes ago, Jarf10 said:

Does anyone know which case he used?

I think it was the Cooler Master MasterRace 5 

[iPolymer]

I never thought of placing both fans aiming at each other, I assumed it would be better. 

 

Also this video reminds me if when Linus did something similar. It's also part of his Final Answer segment now I guess. 

 

 

 

[N.tony]

So, Luke reacted to a silly comment, but not to the actually valid point, that was definitely raised last time? Saying "this configuration is 3° cooler" is not exactly as good sounding as "this configuration is 10% more efficient" (which is more or less the difference between 35° and 38°).

 

If you were to use a much hotter CPU, this difference could

1. be larger in the absolute value (70° vs 76°?)

2. be larger even in the relative value (it could very well be closer to 70° vs 80°, but we'll never know)

 

Besides, you never commented on the Turbo frequency. I mean, of course for the temperatures of ~40° you could expect the CPUs to run at max turbo consistently, who is to say that this holds at higher temps? So, having three scenarios all capped at ~70° would not mean shit if two of them didn't turbo to the same degree. Besides, if turbo boost is governed not only by temperature, but also by power draw, it could still be affected at any temperature, and thus should be controlled.

[me99771]

Anyone know what that song is that starts playing around 4:00?

[aflyinghotpocket]

5 hours ago, Shahnewaz said:

I'd like this test re-done with something that kicks out A LOT of heat, like a dual GPU, and running the test for an hour or two instead of 10 minutes. That might make a significant difference.

I really did like this video. I was actually in an argument with one of by friends the other night over fan placement on the radiator and he was on the "dual fans on outside pulling out team" and I was on the "dual fans inside pushing out" team. Crazy to know the science behind it. 

 

I would like to see a follow-up with different types of fans. Are all fans created equal? OEM case fans? Noctua? NZXT? Thermaltake? and the channel's sponsor Cooler Master? When it comes to system cooling, who is the king? Do different fans matter?

 

I know there was a recent video about air flow where we found out that you could stuff your case will all sorts of crap and not really effect temperatures all that much, but what if you change the fan manufacture? does that effect performance?  Many fans claim to be the best, but does it really mater which one you buy? or should you stuck with OEM fans? 

[mikeeginger]

Linus did this 3 Years Ago why are you Redoing it?

 

[twisties]

8 hours ago, Shahnewaz said:

I'd like this test re-done with something that kicks out A LOT of heat, like a dual GPU, and running the test for an hour or two instead of 10 minutes. That might make a significant difference.

I second this. Dealing with heat in my tight case (Node 204 with a just-fits 980Ti) can be challenging. I think this test would have a different conclusion if you were dealing with the difference of drawing hot air from the GPU over the rad vs fresh air from outside the PC.

[doomsriker]

3 hours ago, me99771 said:

Anyone know what that song is that starts playing around 4:00?

Durude Sandstorm 

[brandishwar]

Nice of you to jab at your critics when they raise a legitimate point.

 

As I and several others have said, you need to account for the variables and control for them to ensure that the claims you're making are actually being tested. Herein you did just that, making sure that fan speeds are held constant. The pump is controlled via a 3-pin header, so it should've been running at a constant speed as well. Thermal throttling on the CPU wasn't going to be a concern at the temperatures you were observing, and the only thing you were manipulating was fan placement on the radiators.

 

We raise these criticisms to help you make better videos wherein the claims are being tested with valid methods. Many others are citing your videos, so that provides an implicit responsibility to get it right. Patronizing your critics in the manner presented in the video rather than listening to your critics is also a good way to alienate your audience.

[JMJenkins]

So what Noctua fans did Luke use, also, what fans should you get for a 3x 120mm/140mm for the front of the case on the CM MasterCase 5 Pro for a quiet operation?

[Razor512]

The pull config works best. Extremely easy to clean, especially if you make a custom vacuum cleaner hose attachment that matches the width of your radiator. I made one out of cardboard in order to clean my kraken x40 in 1 swipe. (Very ghetto, but it works well)

 

PS, it is also useful to make a slim cardboard rectangle with a hose attachment on one end, and a circle cut out (with 1 flap left sticking out on the widest face of the rectangle). This makes it easy to clean videocard heatsinks as it pulls the dust out while also stopping the fan from spinning.

 

This is how the radiator gets after I have not cleaned it for a long time:

 

Radiators clog faster than regular heatsinks since their fins are closer together. They clog at a similar rate to laptop heatsinks (which is why you should clean your laptop heatsink every 2 months or so if you frequently use it.

[stanielz]

What CPU was he using or how did he get temps at 38 degrees under a full load? Am I missing something?

[brandishwar]

2 hours ago, stanielz said:

What CPU was he using or how did he get temps at 38 degrees under a full load? Am I missing something?

Part of the reason for the temperatures was the ambient temperature in the room, a chilly 18C (65F).

[Opferlamm113]

On 24.3.2016 at 8:24 AM, Jarf10 said:

Does anyone know which case he used?

 

On 24.3.2016 at 8:32 AM, iPolymer said:

I think it was the Cooler Master MasterRace 5 

It´s the Coolermaster MasterCase 5 Pro

 

My 2 cents for this video: You only look at one variable, the CPU-temperature. You need to look at the whole system. If you heat up the other components when you draw hot air into the case its less favourable.

I have the same case and I use a Corsair H110i GT. One configuration I tried was the radiator in the front with the fans in push. As I have 3 HDD´s for storage and the RAD blocking the room for the 3-HHD-bay. I placed 2 HDD´s in the HDD bay under the power supply cover. BAD IDEA! I was monitoring the Temps with the Corsair Link software, and it does not refresh HDD temps..., which I found out later. After a 2 hours gaming session during which I monitored the temperature and noticed nothing alarming. I turned off the PC, but I had forgotten something and restarted the PC thats when I saw my HDD in the bottom had temps over 45° Celsius! Reversing the air flow and pull the air through the RAD out of the case in the front yielded lower Temps, but they were still above 40° C. I added a 80cm fan under the cover pushing air through the HDD´s dropping the Temps to about 38°C.

Shortly afterwards one of my drives died.

I then DIYed the the 280mm fan into the top again. The Mastercase 5 officially supports 280mm RADs only in the front.

Now all Temps are in a desirable range.

Ambient Temps are around 20°C.

 

So please LTT dont show simplified versions.

 

[Jarf10]

13 minutes ago, Opferlamm113 said:

I have the same case and I use a Corsair H110i GT. One configuration I tried was the radiator in the front with the fans in push. As I have 3 HDD´s for storage and the RAD blocking the room for the 3-HHD-bay. I placed 2 HDD´s in the HDD bay under the power supply cover. BAD IDEA! I was monitoring the Temps with the Corsair Link software, and it does not refresh HDD temps..., which I found out later. After a 2 hours gaming session during which I monitored the temperature and noticed nothing alarming. I turned off the PC, but I had forgotten something and restarted the PC thats when I saw my HDD in the bottom had temps over 45° Celsius! Reversing the air flow and pull the air through the RAD out of the case in the front yielded lower Temps, but they were still above 40° C. I added a 80cm fan under the cover pushing air through the HDD´s dropping the Temps to about 38°C.

Shortly afterwards one of my drives died.

I then DIYed the the 280mm fan into the top again. The Mastercase 5 officially supports 280mm RADs only in the front.

Now all Temps are in a desirable range.

 

Thanks for this information man! 

 

I use the Spec-02 because I was on low budget at the time I bought it. Officially it only supports 120mm RAD's but my H100i GTX first fitted fine at the top. Then I bought new RAM-sticks and my RAD didnt fit any longer at the top. Now I push air through the front but never look at my HDD temps . I actually was planning to find out how to put the fans under 40% because my cpu is most of the time 27 degrees in the lowest configuration in corsair link. 

 

Do you know any software to monitor the HDD's better? 

[Opferlamm113]

3 minutes ago, Jarf10 said:

Do you know any software to monitor the HDD's better? 

 

Speedfan should do the job.

 

A workaround to get the HDD Temp is to exit the Corsair Link Software by right clicking its symbol in the taskbar and starting it again.

 

 

[Jidonsu]

On March 24, 2016 at 5:06 AM, N.tony said:

So, Luke reacted to a silly comment, but not to the actually valid point, that was definitely raised last time? Saying "this configuration is 3° cooler" is not exactly as good sounding as "this configuration is 10% more efficient" (which is more or less the difference between 35° and 38°).

 

If you were to use a much hotter CPU, this difference could

1. be larger in the absolute value (70° vs 76°?)

2. be larger even in the relative value (it could very well be closer to 70° vs 80°, but we'll never know)

 

Besides, you never commented on the Turbo frequency. I mean, of course for the temperatures of ~40° you could expect the CPUs to run at max turbo consistently, who is to say that this holds at higher temps? So, having three scenarios all capped at ~70° would not mean shit if two of them didn't turbo to the same degree. Besides, if turbo boost is governed not only by temperature, but also by power draw, it could still be affected at any temperature, and thus should be controlled.

Celcius doesn't scale proportionally like that.  Kelvin does. In absolute terms, it's more like 1% difference in thermal energy. 

[SolarNova]

From personal experience, the last thing u want to do with rad fan layout, is dump that air after going through the rad, inside your case, especially if u have additional radiators using that air for cooling, and a large res inside the case sitting in that warm air.

 

I used to run a 240x80 monsta rad in push pull in the front pulling air into the case, then a 240x40 rad on the top pulling air out, then another 240x40 rad bolted to the back of the case pulling air out. The result, a toasty internal case temp and reduced cooling efficiency. So I changed the airflow design so that all the rads are pulling air out from inside the case (negative pressure bad for dust management I know) and temps improved and opening the side panel of the case is no longer like opening an oven door  

 

I still think they need to use a higher heat load in these types of tests, high 30's is nothing, hell, sitting here typing this my system idles in the low to mid 30's  granted my room temp is a tad higher at 23c (no air con). At load using OCCT PSU test(CPU + GPU at load), my CPU after a 10 minute run maxes out at 69c on the hottest core, average high of 65.5c, and GPU maxed at 46c, though the OCCT PSU test cant fully maximise the GPU due to 100% CPU usage, so GPU does get a little hotter during gaming. Even so, the OCCT PSU test puts the most thermal load on a water loop cooling system.

 

The point is, a higher thermal load and thus higher temperature ranges may well show up differences in things like different TIMs, application method, fan orientation, loop order, loop type(parallel vs serial), and so on.

[N.tony]

On 3/28/2016 at 5:49 AM, Jidonsu said:

Celcius doesn't scale proportionally like that.  Kelvin does. In absolute terms, it's more like 1% difference in thermal energy. 

I would agree if we were talking about a thermal output of a device in the first place. Here the question is regarding the heat removal, which (without active components as in a fridge or the phase change compressor) is only affected by the relative difference in the temperature vs the ambient. So yes, I might have been wrong with the 10%, but in the other direction - it's (35°-18°) vs (38°-19°), which is... ~11%

Of course, there are also other factors in place, like how much the temperature of the coolant is cooling on its own while reaching the radiator, how long will it take to reach a steady state (the temperature inside the case would rise to some degree, even if we assume outside room temperature is kept at 18°), how the thermal leakage increases with the components heating up, and so on and so on...

 

The point here is that none of these factors contribute too much, but their relevance is further decreased if the CPU temperature is way higher than the ambient. When you're "testing" at 35°, how can you be sure it's the fault of the fan orientation and not the increased in-case temperature after previous tests, for example?