Overclock.net banner

1 - 20 of 250 Posts

·
Registered
Joined
·
8,273 Posts
Discussion Starter #1
There are frequent questions regarding the Hyper 212+ and whether or not push pull is worth it for the heat sink, so I did a more complete test to find out.

There are also frequent questions regarding whether or not positioning the heat sink vertically would affect performances, so I tested that too.

Note:
These tests are done using Prime95 at 20 minute intervals. The graphics card was not under load during the tests therefore results may vary.

Also the case used was a CM Elite 335 with the rear and front fan grills cut out.

The testing is done using a i5 760 overclocked to 3.8GHz at 1.353V using IC Diamond TIM

The wiring may seem like a mess, but it is the best I can do, all of the extra power cables are up top and there is a hole where the fan in the 5.25" bays blow cool air into the case

Testing Results

All temperatures are recorded in degrees Celsius

Horizontal Tests

First run
900x600px-LL-9b7a62fd_100_0590.jpeg

Push/Pull using CM 2000RPM Blademasters with a rear fan
Ambient Temperature: 23.0
CPU Temperature: 61, 59, 57, 58
Average Core Temperature: 58.75
Temperature Over Ambient: 35.75

Second run
900x600px-LL-94957f83_100_0585.jpeg

Push/Pull using CM 2000RPM Blademasters without a rear fan
Ambient Temperature: 23.0
CPU Temperature: 59, 56, 55, 56
Average Core Temperature: 56.5
Temperature Over Ambient: 33.5

For the rest of the tests, the rear fan was left off

Third run
900x600px-LL-ff776e6d_100_0591.jpeg

Push using 2000RPM Blademaster
Ambient Temperature: 23.2
CPU Temperature: 61, 57, 55, 59
Average Core Temperature: 58
Temperature Over Ambient: 34.8

Fourth run
900x600px-LL-44c0730f_100_0594.jpeg

Pull using 2000RPM Blademaster
Ambient Temperature: 23.3
CPU Temperature: 61, 57, 55, 58
Average Core Temperature: 57.75
Temperature Over Ambient: 34.45

Fifth run
900x600px-LL-202f1489_100_0592.jpeg

Push/Pull using 2500RPM Sickleflows
Ambient Temperature: 22.3
CPU Temperature: 59, 57, 54, 56
Average Core Temperature: 56.5
Temperature Over Ambient: 34.2

Sixth run
900x600px-LL-ff08e20a_100_0595.jpeg

Push using 2500RPM Sickleflow
Ambient Temperature: 23.5
CPU Temperature: 61, 58, 56, 58
Average Core Temperature: 58.25
Temperature Over Ambient: 34.75

Seventh run
900x600px-LL-24e0b322_100_0598.jpeg

Pull using 2500RPM Sickleflow
Ambient Temperature: 22.6
CPU Temperature: 58, 55, 54, 55
Average Core Temperature: 55.8
Temperature Over Ambient: 33.9

Vertical Tests

Eighth run
900x600px-LL-bcf13beb_100_0599.jpeg

Push/Pull using 2000RPM Blademasters
Ambient Temperature: 22.4
CPU Temperature: 59, 56, 54, 58
Average Core Temperature: 56.75
Temperature Over Ambient: 34.35

Ninth run
I forgot the picture, but I think you get the idea
Push using 2000RPM Blademasters
Ambient Temperature: 22.2
CPU Temperature: 61, 58, 55, 58
Average Core Temperature: 58
Temperature Over Ambient: 35.8

Conclusion:
http://www.overclock.net/image/id/1599646/width/900/height/600/flags/LL

http://www.overclock.net/image/id/1599645/width/900/height/600/flags/LL

The push/pull sickleflow was best since it had the highest cfm, but the push pull blademasters are also really good. But if someone buys a brand new Hyper 212+ or a Hyper 212 Evo, I would suggest just using the one fan, as shown above it is pretty close in temperature.

As for the being horizontal or vertical, either is good but you would need ball bearing fans because when I tested it, the blademaster was vibrating audibly. However with the cooler vertical, it is best to have either a gaping hole up top or to have a faster fan than the ones on the cooler.

Questions and suggestions are welcome!

Some more results thanks to Captain Mayhem
Quote:
Originally Posted by Captain Mayhem View Post

been digging around for a 140 and the one that looks like it would work very well is the BGears B-blaster 2 B, unless anyone knows of a better option.
Second test is complete.
Methodology has changed in one aspect, and that was the measurement of ambient temperature. Instead of measuring five centimetres away from the intake fan, I took the measurement directly off the cpu fan blades, resulting in what I think is a better ambient measurement. The other change that was made was the orientation of the HDT from horizontal to VERTICAL.
With that said, here are the results.
Ambient: 23.4
============
PUSH direction

Case panel off
Idle:
Core temp 1: 38.0
Core temp 2: 38.0
Core temp 3: 40.0
Core temp 4: 35.0
TOA: 14.35
Load:
Core temp 1: 75.0
Core temp 2: 74.0
Core temp 3: 70.0
Core temp 4: 73.0
TOA: 49.6

Case panel on
Idle
Core temp 1: 37.0
Core temp 2: 36.0
Core temp 3: 38.0
Core temp 4: 34.0
TOA: 12.85
Load
Core temp 1: 73.0
Core temp 2: 72.0
Core temp 3: 69.0
Core temp 4: 71.0
TOA: 47.85
==============
PULL direction

Case panel off
Idle
Core temp 1: 39.0
Core temp 2: 38.0
Core temp 3: 40.0
Core temp 4: 36.0
TOA: 14.85
Load
Core temp 1: 75.0
Core temp 2: 74.0
Core temp 3: 71.0
Core temp 4: 73.0
TOA: 49.85

Case panel on
Idle
Core temp 1: 38.0
Core temp 2: 37.0
Core temp 3: 38.0
Core temp 4: 34.0
TOA: 13.35
Load
Core temp 1: 74.0
Core temp 2: 73.0
Core temp 3: 69.0
Core temp 4: 72.0
TOA: 48.6
Afterthoughts: well, it's looking like the numbers are speaking for themselves, and this orientation certainly came as a suprise. After comparing both datasets, I'm starting to lean towards saying that the best orientation and direction for the hyper 212+ with side panel fans pulling in air, is VERTICAL with a PUSH configuration using a high static pressure fan such as the Blademaster.
Can someone be so kind as to create a graph for these results?
Next test will be with the Slipstream 1900RPM model.
Quote:
Originally Posted by Captain Mayhem View Post

Test three has been completed. This is with the Scythe SY1225SL12SH 110.3CFM model (rated at 1900rpm)
This time, I omitted the case open portion, as it made no discernible difference to TOA. Also, judging by the numbers from the last test, I would expect a significant drop in temperatures from vertical to horizontal orientation using this fan again. Ambient was measured at both fans and averaged out to 22.53 C, well within testing parameters.

push configuration:
idle:
35
34
35.7
32
avg: 34.175
load:
68
65
64
65
avg: 65.5
TOA: 42.97

pull configuration
idle:
35
33.3
35
31
avg: 33.6
load:
66.7
63
63
62.3
avg: 63.75
TOA: 41.22
Thoughts: This further reinforces the theory that using a high CFM fan in the pull configuration/horizontal orientation works just as well, if not better than the Blademaster in the push configuration/vertical orientation, provided that the HDT fan has a supply of fresh exterior air.
Oh, and I really like the tone the scythe gives off. quite soothing even at full RPM. I'm almost willing to go so far as to mount it permanently in this current orientation using double sided weatherstrip tape.
Stay tuned for more results. This time, there will be a Delta TFB1212GHE involved. yes, the 220 CFM MONSTER!
Graphical results thanks to axipher
Quote:
Originally Posted by axipher View Post

Here are results and graphs:
From First post testing @ 3.8 GHz
Results:
d5ec7d1c_3.8GHzdata.PNG

Graph:
14e3a961_3.8GHzgraph.PNG

From First post testing @ 4.2 GHz
Results:
fb4276c7_4.2GHzdata.PNG

Graph:
6c8dddd2_4.2GHzgraph.PNG
Also similar tests with a grill by mytoyotaco here
 

·
Retired Reviewer
Joined
·
10,092 Posts
I love data. +rep

Those pull fan results are very very interesting -- quite the opposite of other heatsinks. And you had consistent results: both fans produced better cooling on pull alone than with push/pull or push alone.

Opening out the back really shows the effect of an obstacle, even at some distance. I have put my hand on the output side of fans and heard the fans slow down. Now you have shown the open back model makes a big difference.

Those 2500 rpm sickleflows have ball bearings. They also come with their own speed adjustments, so they can be toned down to almost any rpm you like. With ball bearings they can point their airflow up. Pretty good for $4.12 fans.
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #3
they are, however without the controller they will not go above 800 rpm, and the only way to get them to max rpm is to press the 2 little prongs they have for the controller together, and then you have to have them on max. That is the only downside, I have not figured out how to get them to all be controlled. Also when i took out the back fan, I felt SO much air coming out simply from the blademaster itself
 

·
Retired Reviewer
Joined
·
10,092 Posts
Quote:
Originally Posted by chinesethunda View Post

. . . when i took out the back fan, I felt SO much air coming out simply from the blademaster itself
I noticed that myself, feeling the wind on my hand inside my case, the feeling the lack of wind outside the case. The rear grill basically shut down any air getting through. Then I cut out that rear grill . . .
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #6
possibly that might be why.

Also ehume, I had no rear grill but with a fan, maybe it was just because the fan was lower RPM than the blademaster or it was just a bad fan, but i didn't feel much air so I took it out
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #8
lol, I have 4 of the 2500RPM sickleflows on my main rig, I think il pass lol. Also I have decided against selling it, as it contains most of my first rig parts, so i feel sentimental. Also cuz my roomy is a haggling cheapskate lol.

but that aside let me know if I I can do more tests
 

·
Premium Member
Joined
·
5,295 Posts
+rep for good and useful material.

I like to contribute to this thread by making adjustment for ambient temperature variance:-

Some previous studies have shown that the relationship between ambient temp and average core temp is linear with a slope of approximately 1.5C. This slope of 1.5 means, for every degree change of ambient, average core temp rises by 1.5C rather than by 1C.

Raw data from OP is summarized below:-

Test Ambient T Core Temp Ave

1st 23.0 58.75

2nd 23.0 56.5

3rd 23.2 58

4th 23.3 57.75

5th 22.3 56.5

6th 23.5 58.25

7th 22.6 55.8

8th 22.4 56.75

9th 22.2 58.0

Raw data is adjusted by normalizing to the highest temp (ie 23.5C) of the set:-

Test Normalized Amb T Adj. Core Temp Ave (Temp Over Amb)

1st 23.5 59.5 (toa=36)

2nd 23.5 57.25 (toa=33.75)

3rd 23.5 58.45 (toa=34.95)

4th 23.5 58.05 (toa=34.55)

5th 23.5 58.3 (toa=34.8)

6th 23.5 58.25 (toa=34.75)

7th 23.5 57.15 (toa=33.65)

8th 23.5 58.4 (toa=34.9)

9th 23.5 59.95 (toa=36.45)

Charting the above Temperature Over Ambient:-

adjusted toa chart.JPG

Observation

From Tests 2, 3 and 4, the result exhibits the expected pattern:

a. Push/Pull has the lowest delta.

b. Push only and Pull only trail Push/Pull by 1.x C

c. Push only and Pull only differ by 0.5C

But from Tests 5, 6 and 7, the result is odd and unexpected. Push/Pull gives the highest delta while Pull only is the coolest.

Looking over the raw data, I wonder if the ambient temp of 23.5C for Test 6 was actually 22.5C. A mis-recording?
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #10
No they were recorded directly and the temps were lower. But keep in mind there might be a point where too fast of a fan might not perform as well as one. Both fans were on a splitter connected to the mobo. So they are equal in speeds

Sent from my brain using my fingers
 

·
Registered
Joined
·
5,813 Posts
Nice graphs and results!
+rep!

This has been needed for a while!

But all in all it doesn't really make a difference? Your biggest fear is the rear fan/top fan being an obstacle?

This should be link'd on the Hyper 212+ page/club! (;

Goodwork!
biggrin.gif
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #12
Pretty much, I might do this test again in a week to see if the results match still. But yeah for the most part however you want to do it it still cools about the same.
 

·
Registered
Joined
·
248 Posts
I have some Gelid GC2, a couple Ultra Kaze 2000rpm and YL medium speeds on the way, plus I might be able to get that one Nidec I dug up on the way. I'll see if I can replicate your findings. ETA: unknown.

oh and, i also contacted a guy I know who runs CFD simulations, so we can see exactly what is happening with the HDT in terms of airflow.
 

·
Retired Reviewer
Joined
·
10,092 Posts
I don't find that my net temps vary with ambients that range from 20c to 22c on my D14. I have less data for 19c to 23c, but I'm not seeing any patterns. I'm not sure that the dependence on ambient temps can be generalized to all heatsinks. It may be true for heatsink x, but not for heatsink y.
 

·
Retired Reviewer
Joined
·
10,092 Posts
windfire alludes to data obtained with one or more heatsinks that would indicate that the net temperature -- core temp over ambient, what I call TOA -- increases with an increasing ambient temp. So his 1.5c correction factor is that a net temp measured at an ambient of 21c would be 0.5c higher than a net temp measured at 20c on the same system.

I specifically looked for that when I did my pre-test investigations before starting on my D14 work. I didn't see any evidence for it on my test rig.

I'm not saying that the change in ambient would not affect the TOA at higher temps. I'm not saying that a correction factor of 1.1c might be there. I'm also not saying that you won't find a dependence on ambient with other heatsinks. I am saying I didn't find it.

If you have varying ambient temps, you should probably make several identical runs at ambient A, and several identical runs at ambient B, more at ambient C, etc. If you see a trend toward larger TOA's at higher ambients, then you should measure it and apply the correction factor.
 

·
Programmer
Joined
·
28,680 Posts
Well, the general rule is; The higher the ambient temps, the higher the full load temps, given that during the testing, speed and voltage of the system is the same along with the same cooling and degree of airflow.
 

·
Registered
Joined
·
8,273 Posts
Discussion Starter #20
well it would make sense that the higher the ambient the higher the load, however I don't know if I would say I found that trend in mine, then again I didn't do tests at those temps, I did this all in one run with my window open and cold air flowing in.
 
1 - 20 of 250 Posts
Top