Go Back   Hardware Canucks > PC BUILDERS & TWEAKERS CORNER > Overclocking, Tweaking and Benchmarking

    
Reply
 
LinkBack Thread Tools Display Modes
  #191 (permalink)  
Old April 21, 2009, 03:16 PM
3 of 7's Avatar
Allstar
 
Join Date: Nov 2008
Location: Nanaimo BC
Posts: 919
Default

Quote:
Originally Posted by ICD7 View Post
3 of 7 - post 183

Your point is understood and you are correct on that I believe. Taking another look, if there was was flex it would be highly coincidental that it would be only in the copper area.

So why only in the copper area is it different? Thermal Expansion Coefficients between the copper and aluminum?

or assuming no difference in surface quality between the aluminum and copper could it be the thermal gradient across the sink under load affects how the paste adheres when it is separated? Very interesting effect,at least to me anyway.

The application looks fine to me, it may be the camera angle but the application is slightly off center hence the tendency of the paste would be to flow to the closest edge or path of least resistance.

Likely that the exposed area was not making contact with the sink/IHS separated by at least the particle thickness in that one area. Not sure whether it would make that much difference though in thermal result as the center and top section of the IHS are well covered.
It wasn't the camera angle. Even though I was very careful to place the HS on squarely, one time the pattern was definately towards the bottom and the second time it was more towards the top.
I believe the 2 screw system of tightening the HS allows it to squish the paste towards one side or the other before there is enough clamping force to pull it down squarely.
My bet is a line of paste perpendicular to the screws would do a better job.
__________________
Nature always sides with the hidden flaw
Reply With Quote
  #192 (permalink)  
Old April 21, 2009, 04:46 PM
ICD7's Avatar
Top Prospect
 
Join Date: Mar 2009
Posts: 58
Default

Quote:
Originally Posted by sswilson View Post
Has anybody looked into the effect of the Swiftech GT "Bow"?

edit: Sorry, guess I should have elaborated a bit.... I'm wondering if anybody had tried both with the bow and without the bow.... It would seem to me that the bow would prevent an even distribution of pressure across the contact point.

I think has to do with pressure - Motherboards have their limitations as to the amount of stress they can handle and I think the Intel max spec is somewhere around 80.lbs.
The stock heat sink runs about 45lbs +/- 10

I have seen workstation mounts that spec'd 90 psi to squeeze the last degree of performance.

So if you cut the contact area by 1/2 and the pressure was 50 psi to begin with you might end up double the force in the area of contact while still maintaining a reasonable load on the MB. a little bit of a balancing act there.

I have found from the forum testing real world vs Lab that below 50 psi differences between compound become kind of homogenized with less notable differences between compounds

Reply With Quote
  #193 (permalink)  
Old April 21, 2009, 04:53 PM
MpG's Avatar
MpG MpG is offline
Hall Of Fame
 
Join Date: Aug 2007
Location: Kitchener, ON
Posts: 3,143
Default

Sorry it took so long. My i7 motherboard decided to die on me, and RMA was in the US. So I've got my old rig making heat for this testing.

Relevant System Specs (for all runs):
QX9650 @ 4.00GHz (10x400), 1.450Vcore (not lapped)
Asus Formula Maximus
D-Tek Fuzion V2
Thermochill PA120.3 radiators (x3), w/ assorted 38mm fans at approx 500rpm
DDC 3.2 (x2)
120mm S-Flex "D" fan, aimed at socket area for chipset cooling
CPU temps obtained with RealTemp v3.00, uncalibrated
Water temps obtained using a DS18B20 digital sensor

Thermal Pastes Used: MX-2, IC7 Diamond, Cool Labs Liquid Pro

-Each run had at least 24hrs computer on-time, of which at least 12hrs was spent doing Small FFT's on Prime v25.7, before any measurements were taken.
-Idle measurements were taken after computer sat idle for 1/2 hour doing nothing.
-Load measurements were taken after 1 hr of Prime v25.7, set to Small FFT's.

Name:  IC7Contest.JPG
Views: 2018
Size:  58.8 KB

Thoughts/Notes:
-The very first MX-2 run had the advantage of several weeks to seat. I have yet to duplicate those temps in the short-term.
-Obviously, the CPU temps can't be lower than the water temps. I didn't calibrate RealTemp to try and fix this, but at the very least the comparisons should be apples-to-apples across the board.
-After the second IC7 run, after looking over some notes, I was curious as to whether the Fuzion V2 mount system wasn't putting out enough pressure. So, without disturbing the block, I inserted a 1/4" spacer between the backing plate and the motherboard and torqued it down again, giving it another 24hr period to (possibly) cure better. The almost 2 degrees improvement is a little too much to attribute solely to a little extra curing time, so I'm think the Fuzion V2 is a little light on contact pressure.
-While it had the best temps, CL Liquid Pro is seriously a to spread, it'll bork a motherboard if you get it anywhere except on the processor IHS, and I'll probably have to polish the stuff off the waterblock when I finally remove it. Still... definitely came out ahead in this test.

-So take these results how you want. Since the first MX-2 run wasn't 100% fair (the extra curing time), my own judgment says that the IC7 beat out the MX-2 by about 2 degrees. Long-term curing may change these results, but wasn't exactly practical to test here. On the other hand, the CLLP beats out the IC7 by another 2.5 degrees.
__________________
i7 2600K | ASUS Maximus IV GENE-Z | 580GTX | Corsair DDR3-2133
Reply With Quote
  #194 (permalink)  
Old April 21, 2009, 05:46 PM
sswilson's Avatar
Moderator
F@H
 
Join Date: Dec 2006
Location: Moncton NB
Posts: 14,481

My System Specs

Default

That's quite the split you've got on those cores MpG! How long did it initially take you to decide it was the chip and not the mount(s) when you first got the chip? :)
__________________
MSI Z87I Gaming AC / i5 4670K / 2X 4G Gskill 1866 DDR3 / XFX XTR 750 / EVGA GTX 680 SC+ 2GB / Intel DC S3700 200G / random 160G Sata HDD
Inwin 904 / Swiftech MCP655-b / Alphacool NexXxos XT45 120 Rad / 2X Scythe GT AP-15 / EK Supreme HF / Dell UltraSharp U2412M

Asrock AM1H-ITX / AM1 Athlon 5350 / 2X4G Gskill PC3-14900 / Intel 6235 Wi-Fi / 90W Targus Power Brick / 320G Seagate Momentus / Mini-Box M350 / 1X 22" Dell IPS / 1X 22" HP
Reply With Quote
  #195 (permalink)  
Old April 21, 2009, 05:57 PM
MpG's Avatar
MpG MpG is offline
Hall Of Fame
 
Join Date: Aug 2007
Location: Kitchener, ON
Posts: 3,143
Default

Urgh. Looooooooong time. Half a dozen different waterblocks/heatsinks, with more than a few days spent tipping bowed blocks towards all corners of the IHS. They actually get within about 5 degrees of each other by about 80 degrees. But yeah, when I say that the 45nm chips have lousy sensors, I know what I'm talking about.
__________________
i7 2600K | ASUS Maximus IV GENE-Z | 580GTX | Corsair DDR3-2133
Reply With Quote
  #196 (permalink)  
Old April 21, 2009, 07:51 PM
Soultribunal's Avatar
Moderator
F@H
 
Join Date: Dec 2008
Location: Mississauga
Posts: 8,065

My System Specs

Default

Quote:
Originally Posted by 3 of 7 View Post
It wasn't the camera angle. Even though I was very careful to place the HS on squarely, one time the pattern was definately towards the bottom and the second time it was more towards the top.
I believe the 2 screw system of tightening the HS allows it to squish the paste towards one side or the other before there is enough clamping force to pull it down squarely.
My bet is a line of paste perpendicular to the screws would do a better job.

Hey 3 of 7 , try the application method I did on my AMD build (I had the same damn problem with my CNPS9000 (Yours is a Zalman Too right?).

Put 5 really little dots , in a pattern like the 5 side of a dice and then plant it down square while putting the mount through.
Might work like it did for me.
Took many attempts to get the damn HSF seated right with this stuff.

ST
__________________




"We know not why he calls for us, only that when he does we must answer" - DMP 2009

"Dear Iceberg, I am sorry to hear about global warming. Karma is a bitch. Signed - Titanic"

I would rather believe and find god doesn't exist than to not believe and find that he does.

www.realhardwarereviews.com
Reply With Quote
  #197 (permalink)  
Old April 21, 2009, 07:56 PM
3 of 7's Avatar
Allstar
 
Join Date: Nov 2008
Location: Nanaimo BC
Posts: 919
Default

Quote:
Originally Posted by Soultribunal View Post
Hey 3 of 7 , try the application method I did on my AMD build (I had the same damn problem with my CNPS9000 (Yours is a Zalman Too right?).

Put 5 really little dots , in a pattern like the 5 side of a dice and then plant it down square while putting the mount through.
Might work like it did for me.
Took many attempts to get the damn HSF seated right with this stuff.

ST
I've put a Sythe Katana III on this machine and the Zalman is being retired to my dad's machine as we need a non pwm fan for his.
Having said that, even with the eneven spread that HS was still pulling down very good numbers ( even though this mobo is famous for inaccurate sensor readings.)
__________________
Nature always sides with the hidden flaw
Reply With Quote
  #198 (permalink)  
Old April 21, 2009, 08:00 PM
Soultribunal's Avatar
Moderator
F@H
 
Join Date: Dec 2008
Location: Mississauga
Posts: 8,065

My System Specs

Default

Quote:
Originally Posted by 3 of 7 View Post
I've put a Sythe Katana III on this machine and the Zalman is being retired to my dad's machine as we need a non pwm fan for his.
Having said that, even with the eneven spread that HS was still pulling down very good numbers ( even though this mobo is famous for inaccurate sensor readings.)

Okay Fair enough , Was just sharing my experience I had with this and an AMD mounting setup.

Yeah, good numbers still, I agree!

ST
__________________




"We know not why he calls for us, only that when he does we must answer" - DMP 2009

"Dear Iceberg, I am sorry to hear about global warming. Karma is a bitch. Signed - Titanic"

I would rather believe and find god doesn't exist than to not believe and find that he does.

www.realhardwarereviews.com
Reply With Quote
  #199 (permalink)  
Old April 21, 2009, 08:58 PM
3 of 7's Avatar
Allstar
 
Join Date: Nov 2008
Location: Nanaimo BC
Posts: 919
Default

Quote:
Originally Posted by Soultribunal View Post
Okay Fair enough , Was just sharing my experience I had with this and an AMD mounting setup.

Yeah, good numbers still, I agree!

ST
It does makes sense, I'll do it that way when I mount it on the old mans P4.
__________________
Nature always sides with the hidden flaw
Reply With Quote
  #200 (permalink)  
Old April 27, 2009, 09:27 PM
Hall Of Fame
F@H
 
Join Date: Jan 2007
Location: Scarborough
Posts: 1,437

My System Specs

Default

Here are my results.
CPU: E6400
MOTHERBOARD: EVGA 780I
COOLER: STOCK INTEL
CASE: OPEN TEST BENCH
AMBIENT: 72*F

PRE IC7
COMPOUND: THERMALRIGHT

IDLE: CORE#0, 41-42
CORE#1, 39

LOAD: CORE#0, 51
CORE#1, 51

IC7

IDLE: CORE#0, 41
CORE#1, 39

LOAD: CORE#0, 50
CORE#1, 48

The idle temps were taken 30 minutes after bootup. Load tests were taken after running prime95 for 30 minutes. The IC7 and Thermalright applications were as stated in the instructions. As you can see, there was little or no difference in the temps.

After the first IC7 load temp was taken, prime95 was set to run for approximately 4 hours with the cpu oc'ed to 3.0 to aid in migration and curing followed by a 1 hour pc shutdown. Load temps of 60C were achieved with a moderate voltage increase.
Upon reboot, the IC7 tests were repeated. There was no change in the idle temps, but there was a small change in the load temps. The following pics are the IC7 load tests. Since there was no difference in idle temps, I didn't think it necessary to post pics.
Name:  IC7-1.jpg
Views: 450
Size:  224.3 KB

Name:  IC7-2.jpg
Views: 420
Size:  224.9 KB
__________________


EBAY(http://myworld.ebay.ca/scott0639)
Reply With Quote
Reply


Thread Tools
Display Modes

Similar Threads
Thread Thread Starter Forum Replies Last Post
Free innovation cooling 7 carat diamond tim giveaway!! Nademon Overclocking, Tweaking and Benchmarking 218 April 1, 2009 08:10 AM
X3 720BE > X4 :) The Results Thread :) arekieh CPU's and Motherboards 4 March 3, 2009 10:08 PM
[FS] IC Diamond 7 Carat TIM Apollo4g Buy/Sell & Trade 13 February 25, 2009 05:23 AM
First Results, First OC Soultribunal Overclocking, Tweaking and Benchmarking 2 January 31, 2009 07:57 AM
oc results so far Jim CPU's and Motherboards 7 July 25, 2007 12:52 PM