eva2000

Intel Core 2 Duo E6600 + Stock HSF (920D) / TT Big Typhoon / Corsair Hydrocool 200EX

Decided to setup my 2nd Intel D975XBX Badaxe rev304 board to see if it allows me a higher max fsb than my 1st board. Instead of using water cooling, decied to grab my Intel Pentium D 920D stock hsf and pluck a 80x25mm Sunon 49cfm fan on it to try air cooling. I also installed my 2nd Thermaltake Extreme Spirit II chipset cooler as you can see below
Full pics here

System:

• Intel Core 2 Duo E6600 Conroe QPGE Step 5 / B1
• Stock Intel Pentium D 920D heatsink + Sunon 80x25mm 49cfm fan
• Intel D975XBX Badaxe rev 304 1334 modded bios / no volt mods
• 2x 512MB OCZ PC5400 Platinum Extreme XTC 3-3-3-8
• 256MB BFG 7800GTX stock hsf
• 74GB WD Raptor 8MB
• FDD
• 650W Silverstone SLI psu

Temperatures:

I know badaxe board temps are off bty +15C to +22C, but still with stock hsf air, it's hot compared to Swiftech G4 Storm watercooling. How hot you ask ?

@2400Mhz at 9x266fsb stock at 1.275v vcore (1.242v)
room temp = 22.5C

Swiftech G4 Storm water cooling
CPU idle = 31-35C bios, 30-33C coretemp windows
CPU load = 36-39C coretemp windows (stressprime Orthos edition)

vs

Stock 920D heatsink + 80x25mm Sunon 49cfm fan
CPU idle = 52-56C bios, 40-43C coretemp windows
CPU load = 74-75C coretemp windows (stressprime Orthos edition) - even if it's off by +22C, that's 53C load at stock on air hehe.



That's like a difference of +10C idle and +39C load hotter on stock hsf air than water!

Guess need to get a better heatsink but am limited due to TT Extreme Spirit II chipset cooler position, I will need a heatsink that is no more than 100mm x 100mm or has a clearance of 60mm above cpu socket to accomodate chipset cooler. Only heatsink i know that can do that so far is Tuniq 120 Tower as you can see from Coolaler's thread at http://www.xtremesystems.org/forums/...d.php?t=107212

Right now don't have any LGA775 heatsinks that i can use for this 2nd rig yet

eva2000

Just took the stock heatsink off to see if i can put a thermal probe under. Looks like the spread of AS5 on IHS shows there isn't enough pressure from the stock heatsink mounting mechanism (920D).

Put the thermal probe under and boot up to bios idle temps of 99-101C

Shutt off pc ASAP, and remove stock heatsink and the probe (although very thin) seems to reduce the mounting pressure even more, only the tip of the dab of AS5 touched the stock heatsink's base!

Whether the stock conroe heatsink's mounting pressure is different not sure ???

I did lap the E6600's IHS till the copper showed, so maybe that's why ? No probs with Swiftech G4 Storm mounting, maybe due to spring tension system ?

eva2000

Had to get a better heatsink for my 2nd rig, so got a Thermaltake Big Typhoon http://www.thermaltake.com/product/C...4/cl-p0114.asp but changed the stock 120x25mm 0.3A 54cfm 3pin 1.87mmH20 fan for a Thermaltake 120x35mm 0.4A 81cfm fan 4pin 4.71mmH20 fan from http://www.thermaltake.com/product/D...2003/a2003.asp

It's a huge heatsink weighing in at 813 grams but simple hand weighing guestimate doesn't seem like 813g to me shrugs



Full pictures at http://i4memory.com/reviewimages/coo...gtyphoon/html/

Just finished installing the Big Typhoon, I had to use the smaller H clip for mounting and instead of using the 4 include nuts, I grabbed the spare screw in stand offs off my Thermaltake Extreme Spirit II chipset coolers, so i could finger tigten the mount instead of using pliers for the nuts.





Full pics at http://i4memory.com/reviewimages/coo...axe_installed/

Load into bios at stock 2400Mhz 9x266fsb at 1.242v vcore showed much better temps. Same goes for in windows

Thermaltake Big Typhoon + TT 120x25mm 81cfm fan
CPU bios idle = 38-40C idle in bios (that's 5-7C hotter than H20 cooling but a full 12-18C cooler than the stock 920D heatsink!)
CPU windows idle = 32-33C (that's 20-24C cooler than stock 920D heatsink and comparable to H20 cooling!)
CPU load = 47-48C dual stressprime small FFTs load (a full 25+ C cooler than stock 920D heatsink and only 9-12C hotter than H20 cooling)

Big Typhoon Idle


Big Typhoon Load


Looks like this 2nd Bad axe board is better than my 1st, with 1067fsb strap this board seems to be able to do 430-432fsb compared to my 1st Bad axe board which maxes at 421-423fsb at 1067fsb strap



Lucky I didn't sell this 2nd Intel D975XBX Bad Axe board, the higher fsb limit on this board allowed me to easily beat my previous best 32M time of 13min 38.062s which was on water cooling.

How's this for Big Typhoon air



And higher and faster



So far with Big Typhoon and TT 81cfm 4.71mmH20 fan 3809mhz at 1.5v but badaxe temp must be off 83C under dual stressprime 'Orthos' load LOL



That failed after ~15mins heh

Dropping down to 3750mhz to see how it goes.





Just decided to drop to 8x400fsb for a run with ram @400mhz 4-4-3-5 at 2.2v

eva2000

The Thermaltake Big Typhoon was the best investment I made. Why ? Because it's LGA775 mounting gear is perfect for my old and trusty Corsair Hydrocool 200EX unit which uses puny 1/4" OD tubing but still packs the cooling power to beat the Big Typhoon heatsink

The pictures:

First, a pic of the Hydrocool 200EX stock waterblock with 1/4" OD tubing and what I intend to do with Big Typhoon mounting gear for LGA775 socket. Yes 6mm inner diameter tubing .. puny but still effective for this



Problem though is the waterblock has a thermal probe for displaying temps on Hydrocool 200EX and it's soldered to the top of the block making mounting uneven. So what to do ? Yup, Thermaltake again to the rescue with copper BGA ramsinks



Much more even mounting surface now:



Corsair Hydrocool 200EX lives on thanks to Thermaltake Big Typhoon and BGA copper ramsinks



Full pics at http://i4memory.com/reviewimages/coo...X_LGA775/html/

How's it performing ?

Corsair Hydrocool 200EX with TT Big Typhoon LGA775 mount
CPU bios idle = 31-32C idle in bios (that's 5-8C cooler than Big Typhoon)
CPU windows idle = 24-25C (that's 7-8C cooler than Big Typhoon)
CPU load = 36-39C (that's a full 8-11C cooler than Big Typhoon)

@3750Mhz at 1.5v
Big Typhoon idle = 45-48C room temp = 24.5C
Big Typhoon load = 82-83C room temp = 24.5C
vs
Hydrocool 200EX idle = 34-36C room temp = 21.5C
Hydrocool 200EX load = 65-66C room temp = 21.5C

So far nice improvement.. managed to increase my clock from Big Typhoon @3865mhz to Hydrocool 200EX @3878mhz and cut 2.703 seconds of my personal best 32M time

eva2000

Wow i changed power supplies from 650W Silverstone SLI to my 1KW PCP&C SLI psu on this Hydrocool 200EX cooled E6600 rig and it seems my Gskill PC6400PHU2 2GB HZ pair #1 is clocking higher than it did on my other E6600/Intel D975XBX rig at 6GB Gskill F2-PC6400PHU2 2GB HZ Micron D9GMH + Intel Core 2 Duo E6600. On other rig at 2.2v vdimm with 4-4-4-12 timing the max for pair #1 was 478mhz, but on this rig so far been able to do 486mhz and i haven't tried higher yet!

I wonder if it's the better psu or the better Intel D975XBX Badaxe motherboard on this 2nd rig that is the reason for higher mem clocks ?

Here's pair #1 at 486mhz 4-4-4-12 at 2.2v



Also seems to have managed my highest cpu clocked passed 32M runs since changing power supplies. Not my fastest times but that can be worked on later

@3882mhz


@3886mhz


from same windows session above two 32M times

No probs for pair #1 to complete 32M at 490mhz 4-4-4-12 at 2.2v, that's 12mhz higher than with other badaxe/700W OCZ GameXStream system at same timings and vdimm!

CPUZ Validation http://valid.x86-secret.com/show_oc.php?id=109066



And here's dual 32M at 490mhz 4-4-4-12 at 2.2v

Digi cam shot at half way mark


near the end


Complete @3533Mhz at 9x392fsb with memory @490mhz 4-4-4-12 at 2.2v

eva2000

Anyway giving Super Pi a rest while i raid the ORB with FM approved 6.5 drivers

1st submission for 3dmark06 @729/859 stock hsf air for X1900CF/X1900XTX combo



Compare Url = http://service.futuremark.com/compare?3dm06=411731

Not too often will I see myself in FM Hall of Fame Top 5

All stock hsf for HIS X1900CF @732/859 + Gecube X1900XTX @729/859

Compare Url = http://service.futuremark.com/compare?3dm06=411744





3dmark05 HOF #4 too now as well http://www.futuremark.com/community/halloffame/

3dmark05 = 22,160
http://service.futuremark.com/compare?3dm05=2138103



Very cool night for 3dmark benching on stock hsf. Room temp ~19.8C (around 10C cooler than normal, coldest in over 18+ months in Brisbane Australia! ).

Both master & slave @732/864 with 6.5 official catalyst drivers.

3dmark06 = 12,599
http://service.futuremark.com/compare?3dm06=412452



That would put me in #3 spot on 3dmark06 Hall of Fame for now

Here's 3dmark2001 with and without LOD

Without LOD = 67,193
With LOD = 68,798

eva2000

Some more stability tests, Hydrocool 200EX holding it's own although room temps have risen nearly 5+ C since last heh.

@9x400fsb = 3600mhz at 1.418v straight from bios to boot into windows, ran dual stressprime large FFTs for 2 1/2hrs then while still running, ran 3dmark06 with X1900 Crossfire at stock 625/725

idle



load

eva2000

Guess what arrived today and yesterday ?

Yesterday, my 2nd hand Corsair Cool kit arrived (thanks Cashy ), with single 120mm fan radiator, Swiftech MCP350 pump and AMD64/s478 waterblock and res etc.

Today my Radiical Chipset waterblock http://www.radiical.com.au/products/Chipset.asp, 3/8" steel hose clamps and 3/8" Clearflex 1/16" wall tubing arrived.

My plan is to take the 120mm single fan rad and MCP350 pump out of the Corsair Cool kit and pair it with my Radiical Chipset waterblock and Dangerden fill port. This will make a nice compact dedicated MCH/northbridge cooling loop for my Core 2 Duo adventures

The parts http://i4memory.com/reviewimages/coo...corsaircool/1/




Leak testing with my Kayl custom 100W psu powering the pump @13.2v http://i4memory.com/reviewimages/coo...cool/leaktest/



It's going to replace my Thermaltake Extreme Spirit II chipset air heatsink



Anyone say overkill


Just finished installing the Radiical chipset waterblock on my Intel D975XBX http://i4memory.com/reviewimages/coo...ool/installed/



Wish me good luck




Looks like in windows i've hit both the limit for cpu/cpu cooling at 1.55v vcore and the MCH dependent max fsb

For cpu/cpu cooling hit limit at 9x433-436fsb = 3898-3924mhz for 3dmark06 cpu test.

For MCH dependent fsb max at 8x443fsb.

Improvement nonetheless considering i couldn't even boot with 9x433fsb before!

3dmark06 with X1900CF at stock speeds



Interesting does better MCH cooling give faster super pi times ?

With chipset water cooled, I ran 32M with crossfire + bloated ATI CCC installed, then ran 32M without crossfire and ATI CCC removed @9x400fsb 4-4-3-8 and compared it to my best times with Thermaltake Extreme Spirit II chipset air heatsink and interesting results.

TT Extreme Spirit II chipset air
32M untweaked = 14min 24.750s
32M tweaked = 14min 21.531s

vs

Radiical chipset water cooled
32M untweaked crossfire enabled = 14min 30.843s
32M untweaked no crossfire = 14min 19.422s - that is my fastest 32M time to date with 9x400fsb 4-4-3-8!

New 32M personal best 7.407s faster than previous best



My fastest 1M times to date

@3914Mhz - 9x434.9fsb


@3938Mhz - 9x437.6fsb


@3962Mhz - 9x440.3fsb


Hmm looks like i might be still MCH limited on 1067fsb strap due to the tight 1067fsb strap latencies.

Before
1067fsb strap = 430fsb max
1333fsb strap = 430fsb max, 433fsb will no boot

Radiical chipset water
1067fsb strap = 433fsb, 434fsb even at 8x 434 will fail 32M
1333fsb strap = 440fsb at 8x multiplier

Lacking MCH volts probably at 1067fsb strap. Still an improvment i guess.


Edit: possibly made a big boo boo in my assumptions. I did all tests at 4-4-3-8 timings at 2.2v vdimm. Turns out this pair of Gskill PC6400PHU2 HZ's (1st pair i have) maxes out at 432-433mhz 4-4-3 at 2.2v vdimm and 490mhz 4-4-4-12 at 2.2v. I made the general assumption that since my 3rd pair of Gskill PC6400PHU2 HZ's i have on my other Bad axe/E6600 system did 473mhz 4-4-3-8 at 2.2v that 432mhz would be easy for my 1st pair of memory

Interesting differences
E6600/D975XBX Badaxe 1334 bios with 700W OCZ GameXStream
pair #1 memory = 478mhz 4-4-4-12 at 2.2v max but didn't test 4-4-3 max
pair #3 memory = 478mhz 4-4-4-12 at 2.2v max and 473mhz 4-4-3-8 at 2.2v max

vs

E6600/D975XBX Badaxe 1334 bios with 1KW PCP&C SLI
pair #1 memory = 490mhz 4-4-4-12 at 2.2v max (+12mhz higher than other rig)
pair #3 memory = not tested on this system yet

So have to retest my max fsb with 4-4-4-12 timings now heh

Update: looks like the 432mhz 4-4-3-8 at 2.2v limit was a bad mount for the Radiical chipset waterblock. I did a remount with fresh application of Ceramique thermal paste and now no probs testing up to 460mhz 4-4-3-8 at 2.2v so far. More tests to come





....
Okay can confirm, with 975x motherboards at least the key to both high fsb and memory clocks and tighter memory timings is all in the NB/MCH (memory controller hub).

With Thermaltake Extreme Spirit II chipset air heatsink, my 1st pair of 2x1GB Gskill PC6400PHU2 HZ maxed out at 490mhz 4-4-4-12 at 2.2v



Now with Radiical chipset waterblock cooling my D975XBX Bad axe MCH/NB, the 1st mount, the same 1st pair of memory maxed out at 476mhz 4-4-4-12 and 433mhz 4-4-3-8 at 2.2v.

So decided to remount the Radiical chipset waterblock with fresh layer of Ceramique thermal paste. Guess what can do 490mhz 4-4-4-12 at 2.2v again. But best of all I can now do 488mhz 4-4-3-5 at 2.2v with very tight tRFC value of 25!



With better MCH cooling seems my memory is flying even at 2.2v vdimm

Here's what clocks it took to break 15mins