Overclock.net › Forums › Cooling › Water Cooling › How to run a proper FAN comparison?
New Posts  All Forums:Forum Nav:

How to run a proper FAN comparison? - Page 8

post #71 of 97
Thread Starter 
I'm going to test each heatsink using PK2. It has no set time (im sure every paste has a slight burn in time though), and is supposed to be the easiest to spread.

Everyone always talks about PK1, which is supposed to be 'the best' paste, from what i read (except indigo extreme which isn't a paste and that **** that melts copper and damages aluminum, the liquid pro and liquid ultra stuff which is super expensive anyways), but apparently:

PK-3 > PK-2 > PK-1 in terms of performance
PK-2 > PK-1 > PK-3 in terms of ease of spread

I'm thinking of testing all 3 pastes on both the H50 and Nh-d14 though. Due to the HDT design of the 212 I'm not sure if it would be a great idea to test all 3 pastes on it. I mean maybe, but it'd just be a big waste of paste i think and its too variable.

Anyone have any comments on any of this? I'm going to using PK2 to test all 3 heatsinks because of ease of spread and no burn-in time - i think it'd make a more consistent result (i know i said i'd test all 3 on the nh-d14 and h50, but only pk2 and stock on the hyper 212.)
Edited by Belial - 12/5/12 at 9:34pm
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
post #72 of 97
IMHO, pick one paste you like best and stick to it and rather than test more pastes, test multiple mount and remounts.

I have always liked 5 mounts when testing CPU coolers and felt that even 5 mounts isn't enough sometimes. A nice thin past should give you more consistent and repeatable results, but I think choosing a paste that is common and well liked/used is probably more important than anything and the more mounts the better.thumb.gif
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
post #73 of 97
Thread Starter 
I'm doing 2 separate tests here...

1. A comparison of PK1, PK2, and PK3. I'll do this with the H50, all on the same heatsink, although I'll consider testing all 3 on the NH-d14 as well.

2. A comparison of the Hyper 212+, H50, and Noctua NH-D14 in as close to an apples to apples performance as possible. In this case, I will attempt to use the easiest of the 3 pastes I have available to maintain consistency, ie rule out thermal paste or application or anything besides the heatsinks themselves being a factor. It seems to me that PK2 is the easiest to spread (base on prolimatech's word) and lack of burn in time (both pk2 and 3 seem to have no set time).

Hyper 212 = 2 x yate loon medium
H50 = 2 x yate loon, radiator outside the case so no intake vs outtake issues
NH-d14 = 120mm yateloon medium/stock140mm/120mm yateloon medium (i feel like the 140mm fan is such an integral part of the dual tower heatsinks, that's why I'm leaving it on, it's a clear advantage to be able to mount 3 fans that i think should be included in the test)

I'll air condition the room, ceiling fan on, something like 65*C. A problem I see with a lot of tests is they compare delta, but as ambient temperature increase, the max load temperature will increase exponentially, so a heatsink might be twice as good but it has the same delta when it's 65*Cidle - 80max vs 70&C - 80 max.

Prolimatech gave me a sample of:
1g PK1
1.5g PK2
1.5g PK3

In order to do this test. PK3 is supposed to be the best so I imagine I'll stick with that, I'll probably have to buy some because I'll probably use everything up in these tests, but I'll use whichever paste ends up testing best.
Quote:
but I think choosing a paste that is common and well liked/used is probably more important

That sounds absolutely retarded and I'm pretty sure you didn't read my posts or thread at all. Like those people who say AS5 is the best paste ever but it's like what 5 years old now and clearly beaten by every new high end paste (would love to test it if someone has a sample, arctic silver didnt' respond to email).
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
post #74 of 97
@op you might wants to do 3 tests of the same paste per cooler, like 3 different mounts. Then take the average of the 3, this is because its really difficult to get a perfect mount so you should try the same paste more than one to get a better avg temp per cooler:thumb:
post #75 of 97
Quote:
Originally Posted by Belial View Post

That sounds absolutely retarded and I'm pretty sure you didn't read my posts or thread at all. Like those people who say AS5 is the best paste ever but it's like what 5 years old now and clearly beaten by every new high end paste (would love to test it if someone has a sample, arctic silver didnt' respond to email).

My point is with any testing you need to isolate the variables and focus on one thing at a time. No I didn't read all the details, I assumed the intent was heatsink comparison, but the concept still applies. I wasn't trying to say use AS5, I was simply saying choose a test variable and make sure you spend enough time on multiple mounts.

If you really want to do both heatsinks and thermal pastes, go for it, but don't forget the most critical piece and that is completing at least 5 remounts per combination. A single mount test is only good for maybe 2-3 degrees of accuracy at best, too few numbers of mounts is a rookie mistake and creates useless data. Unless you do at least five mounts and retests of the same setup, it's just not worth bothering because the differences between pastes is usually less than the amount of mounting error. Back to my earlier recommendation, but 5 remounts with 2 heatsinks over 4 pastes would add up to 40 tests. I see people far to often think too big and never finish. Baby steps...

You are also going to log temperatures of ambient and core using some sort of accurate logging tool right? Another rookie mistake is to try to benchmark without logging the data. Ambients vary a lot more than people understand and unless you log the ambient temperature AND the core temps simultaneously, you probably have a 2-3C ambient error. A one mount test looking at a wall thermometer is no more accurate than probably 5 degrees or worse. That's how people come to the forums and say "retarded" things like they got a 10+C improvement from changing out a block or other item when in reality the block may have made a 1C difference and the other 9C was testing error.

I have used a Crystalfontz CFA-633 and used digital one wire sensors which allows upwards of 20 sensors you can log which have a 0.2C accuracy. An aquaero 5 is an easier way to do it, but the sensors are only 1C accurate and you are limited to 8 I think To do it right you really need about four sensors per fan section located right at the inlet of the fan. If you don't have a logging tool, you should at least do some sort of manual logging every so often and make sure to use something with a resolution better than 1C.

But why do anything that increases accuracy of the test to the point of making it repeatable and valid that other people can use (a proper benchmark comparison), that would be retarded..tongue.gif
Edited by Martinm210 - 12/6/12 at 4:37pm
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
post #76 of 97
Thread Starter 
Quote:
Originally Posted by NomNomNom View Post

@op you might wants to do 3 tests of the same paste per cooler, like 3 different mounts. Then take the average of the 3, this is because its really difficult to get a perfect mount so you should try the same paste more than one to get a better avg temp per cooler:thumb:

I'm not sure if i have enough paste for that. Is this for the Thermal Paste test, or for the Heatsink test, that your talking about?

The hyper 212 with its HDT design i think would make it too inconsistent to compare the different pastes on, or the same paste over and over. I mean you're talking 1*C per mounting anyways, at most.
Quote:
If you really want to do both heatsinks and thermal pastes, go for it, but don't forget the most critical piece and that is completing at least 5 remounts per combination. A single mount test is only good for maybe 2-3 degrees of accuracy at best, too few numbers of mounts is a rookie mistake and creates useless data. Unless you do at least five mounts and retests of the same setup, it's just not worth bothering because the differences between pastes is usually less than the amount of mounting error. Back to my earlier recommendation, but 5 remounts with 2 heatsinks over 4 pastes would add up to 40 tests. I see people far to often think too big and never finish. Baby steps...

I do not have near enough paste for that lol. What I want to get out of this test is PK1 vs PK2 vs PK3 vs Stock (i dont even have stock heatpaste, but ill use peanut butter or whatever tomshardware tested really well with that was goofy like that, or maybe the nh-d14 will come with paste or something).

Also the heatsinks should all be rather different, so hopefully the test will basically show is the H50/NH-D14 a significant step up from the 212/h50.

I will make sure to do all testing with ambients at identical temps with A/C. Like set A/C to 65*C or whatever, always test there. I'm aware of the expontential impact of differing ambient temps on the load temp. I mean that's the best I can really do.

I have an NZXT sentry 2 fan controller, one the wires is set up just for ambient/case temps, but frankly it varies just sitting there a good 3-4*C. I'm not really sure what I could do.

I would love to make the test as accurate as possible but there's only so many resources i have. I mean I'm the guy who's using sampled paste, and got the H50 for $27 and NH-D14 for $43 (and going to sell one off).
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
post #77 of 97
I meant for the thermal paste test
post #78 of 97
Sounds good man..thumb.gif Honestly we all get a little too anal in watercooling testing and benchmarking because we're all a bunch of computer geeks and that's to be expected..hehe

Heck I think Vapor spent a good 6 months working on his TIM round up which I think eventually did him into permanent testing retirement, he went to the extreme X10 in testing not only multiple mounts, but had three test rigs to evaluate three levels of contact. Check it out if you haven't already:
http://skinneelabs.com/2011-thermal-paste-review-comparison/2/

He must have litterally logged hudreds of hours benching TIM.


Most importantly do what you can and have fun with it.thumb.gif
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
3930K ASrock X79 Ext7 Evga GTX570 CORSAIR Vengeance 16GB 
Hard DriveOptical DriveOSMonitor
Phoenix SSD+2TB Seagate drives ASUS DVD-RW combo Windows 7 28" Hanns-G HZ281 
KeyboardPowerCaseMouse
MS Natural Corsair TX650 Danger Den Torture Rack Logitech MX518 
Mouse Pad
Desk 
  hide details  
Reply
post #79 of 97
As Martin said, make sure, first of all, that's your benchmark analysis remains fun for you to do. When it becomes a chore, stop it before you get burnt out.

Now onto the benchmarking methodology. I'll start by stating that I'm a compensation consultant, and that my job is to define, sell and make sense of market compensation data for my clients, so I have some experience about benchmarking methodologies. My advice is to
1. Define your testing methodologie beforehand, and stick to it. As MArtin said, the objective is to test 1 variable at a time, by making sure other influencing factors are constants. This usually translates into the methodology page that everybody skips when reading review, but this is key to obtain consistancy.
2. Define your acceptable data sample. To put it simply, don't shoot for reliable sampling, as it would mean running all tests about 20-25 times before calculating median results (so that each indivdiaul results weights approx. 4-5% of the stat, thus neutraliziing outlying results). Report median results rather than average results, as the average is skewed by extreme observations.

Good luck with your testing!
post #80 of 97
Thread Starter 
ibt seems a bit too random on my temps, i noticed it causes waves, so im going to use custom small fft fft length 8min 8 max. I think I'll just use max temp... since with small fft custom 8 min/max, temps really dont seem to vary much like in ibt or p95, the max is same as the average/median during the test.

I dont think I can do multiple remounts, but i'll do as i stated in op/later posts to make it as close to apples to apples (use identical fans, h50 outside the case.)

If I have TIM left over, I will do a thermal paste test (all 3 pastes) on the NH-D14 cooler, and if I still have paste leftover, I'll do it on the Hyper 212. That would also help the heatsink comparisons because then I'll have a benchmark of each heatsink on pk1, pk2, and pk3.
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
Build In Progress
(17 items)
 
   
CPUMotherboardGraphicsRAM
Haswell i7-4770K 4.8GHz@1.452v/1.97vRIN Gigabyte Z87X-UD3H Older Gigabyte 7950 (H60 Mod) 2x4GB Gskill 2400CL11 Hynix CFRs 
Hard DriveCoolingOSMonitor
Samsung 830 128gb Corsair H110 w/4x 140mm Yate Loon High Blues W7 x64 Ultimate Yarrgh Matey Dell U2312HM Matte IPS 1080p 
MonitorMonitorKeyboardPower
Sanyo DP19640 18.5" 768p KDS Rad-7xp 1024p CM Storm Quickfire Rapids Brown Rosewill Capstone 750 
CaseMouseMouse PadAudio
Bitfenix Shinobi Modded CM Storm Spawn Steelseries Diablo 3 2 set CD Speaker + Big Sub 
Other
Bitfenix Recon Fan Controller Moded 
CPUMotherboardGraphicsRAM
Delidded i7-3770K 5GHz@1.499 Gigabyte Z77X-UD5H 1.1 Sapphire Dual-X 7950 3L Boost 4x2GB Mushkin Ridgeback 2200mhz 8/11/8/27 1.75v... 
Hard DriveCoolingCoolingCooling
Intel X25-M 80GB SSD NH-D14 SE2011 Noctua 120/140/140 Fans 3 x Yate Loon Mediums (Petras) 2 x NZXT Havik 140mm Fans 
OSMonitorMonitorKeyboard
Hax'd W7 Ultimate 64-bit Sanyo DP19640 19'' KDS Radius Rad-7xp 17'' Ducky 1087 MX Brown 10keyless 
PowerCaseMouseMouse Pad
Rosewill Capstone 550w Modular NZXT Source 210 w/Modded Window Razer Deathadder 3.5G Steelseries Blizzcon D3 
AudioAudioOtherOther
Phillips FW-C250 2 Speaker Set w/3 CD Changer JBL Sub 6 Subwoofer NZXT Sleeved LED Strip 1m Blue NZXT Sentry 2 Fan Controller 
Other
Colonial Blue Paracord Sleeving 
CPUMotherboardGraphicsRAM
Phenom ii X4 955C2, $31, 5 broken pins, works! ... Biostar A770e3 6.3 Red&Black MSI GTX 460 768mb Twin Frozr II 2 x 2GB Kingston HyperX 1333 OC to 1348CL7@1.65v 
Hard DriveCoolingCoolingCooling
Intel X25-M G2 80GB SSD NH-D14 SE2011 6 x Case Fans (4 x YL-M, NZXT case fan, CM Hype... Jerryrigged Chipset-Northbridge 80mm HSF 
CoolingOSMonitorKeyboard
Jerryrigged VRM Heatsink (sawed off stock AM2 h... Hax'd W7 Ultimate 64bit Sanyo DP19640 19"1360x768 Ducky Cherry Browns 10keyless 
PowerCaseMouseMouse Pad
Antec Earthwatts 430D NZXT Gamma Steelseries Kinzu Optical Steelseries Blizzcon Diablo 3 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › How to run a proper FAN comparison?