Overclock.net banner
1 - 20 of 56 Posts

·
Registered
Joined
·
456 Posts
Discussion Starter · #1 ·
I would like to share my results with the forums


IHS Mod

Because of very high temps, I've modded the IHS by removing the manufacturer original thermal compound and replacing it with Artic Silver 5. This mod drops 10 to 15°C in full load




SYSTEM

Code:
Code:
FX-55 CABCE 0516WPMW core #484
DFI nF4 Ultra-D {501-2fix | 623-1 | 623-2 | 704-2BTA}
2*512 TCCD 507 | 2*256 BH5
1*Gb X800XL
Antec True Power 2.0 550W
Win XP Pro SP2 | Win 2k3
COOLING

Code:
Code:
Watercooling:
Pump Atman A-103 (1200l - 1.2m)
Radiator from Air-Conditioning {+ a Radiator BIX}
Homemade waterblock

Liquid chilling:
Compressor 1/4HP
Pecomark condenser 
HX tube inside tube 
Refrigerant: R402A
Evap temp: -20°C

Cascade:
2 * Rotary Compressors 1HP
First stage refrigerant: R290
Second stage refrigerant: R170
Evap temp: -98.5°C unloaded
OVERCLOCKING RESULTS

1) CPU Max Screen = [email protected]



2) CPU Max Stable = not yet

3) 3DMark01 SE



34658pts | 3205MHz @1.65V Chiller | BH5 267MHz 1.5-2-2-0 @3.7V | X800XL 450/590MHz | http://service.futuremark.com/compare?2k1=8930228



32551pts | 3031MHz @1.55V Watercooled | TCCD 303MHz 2.5-4-3-5 @2.8V | X800XL 410/555MHz


31884pts | 3001MHz @1.52V Watercooled | BH5 273MHz 2-2-2-5 @3.7V | X800XL 410/555MHz | http://service.futuremark.com/compare?2k1=8852131



30464pts | 3001MHz @1.46V Watercooled | TCCD 300MHz 2.5-4-3-7 @2.7V | X800XL untweaked | http://service.futuremark.com/compare?2k1=8847589



25990pts CPU Watercooled & TCCD at stock freqs | X800XL untweaked | http://service.futuremark.com/compare?2k1=8843244

4) Super PI 1M

24.234 | 3432MHz @1.65V Cascade controled temp | BH5 @264MHz 1.5-2-2-0 @3.7V



24.719 | 3414MHz @1.70V Cascade controled temp | BH5 @263MHz 1.5-2-2-0 @3.7V



25.375 | 3251MHz @1.55V Chiller | BH5 @271MHz 1.5-2-2-0 @3.7V



25.968 | 3138MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-0 @3.7V



26.188 | 3133MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-5 @3.7V



26.422s | 3113MHz @1.55V Chilled C2H5OH | BH5 283MHz 2-2-2-5 @3.7V



26.594s | 3082MHz @1.50V Watercooled | BH5 280MHz 2-2-2-5 @3.7V



26.609s | 3101MHz @1.52V Watercooled | BH5 282MHz 2-2-2-5 @3.7V



27.094s | 3080MHz @1.50V Watercooled | TCCD 308MHz 2.5-4-3-7 @2.7V



27.203s | 3050MHz @1.50V Watercooled | TCCD @305MHz 2.5-4-3-5 @2.9V



28.610s | 2944MHz @1.50V Watercooled | TCCD @310MHz 2.5-4-3-5 @2.8V



31.922s | 2612MHz @1.15V Watercooled | BH5 @201MHz 1.5-2-2-5 @2.7V



32.844s | 2405MHz @1.46V Watercooled | BH5 @267MHz 1.5-2-2-5 @3.7V



33.109s | 2405MHz @1.42V Watercooled | BH5 @267MHz 1.5-2-2-4 @3.7V



33.172s | 2406MHz @1.42V Watercooled | BH5 @267MHz 1.5-2-2-5 @3.7V



36.860s | 2156MHz @1.50V Watercooled | TCCD @308MHz 2.5-4-3-5 @2.8V



5) CPU-Z

[email protected] Cascade [email protected]°C



3517MHz @ 1.70V chiller [email protected]°C |http://valid.x86-secret.com/show_oc?id=79603

3439MHz @1.63V Chilled C2H5OH [email protected]°C | http://valid.x86-secret.com/show_oc?id=71814

3420MHz @1.55V Chilled C2H5OH [email protected]°C | http://valid.x86-secret.com/show_oc?id=70973

3300MHz @1.63V Watercooled | http://valid.x86-secret.com/show_oc.php?id=61962

3252MHz @1.50V Watercooled | http://valid.x86-secret.com/show_oc.php?id=61955

6) PiFast 4.3 10M

28.26 | 3133MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-0 @3.7V



28.83 | 3139MHz @1.42V chilled C2H5OH | BH5 @241MHz 1.5-2-2-5 @3.7V



7) OCCT

3000MHz @1.39V Chilled C2H5OH



 

·
Premium Member
Joined
·
2,871 Posts
I can tell you what. Some nice steady hands on removing the lid off that FX-55 hey ?

some nice results there..

cheers..
and be happy not mad !
 

·
Premium Member
Joined
·
13,104 Posts
Awesome work man, how come you just didnt leave the IHS off altogether? You couldve had even better temps without it. Nice job on the ocing!
 

·
Registered
Joined
·
33 Posts
Quote:


Originally Posted by bigval

Awesome work man, how come you just didnt leave the IHS off altogether? You couldve had even better temps without it. Nice job on the ocing!

I was just thinking the same thing.....
thinking that the IHS would actuall hold in a little heat...as with out it like on the older AMD chips that having a little air being able to flow slightly around the chip and having direct contact with the heatsink would allow for better cooling...
 

·
Registered
Joined
·
688 Posts
wht did you use to stick the IHS back on?
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #6 ·
Quote:


Originally Posted by xie67

I can tell you what. Some nice steady hands on removing the lid off that FX-55 hey ?

some nice results there..

cheers..
and be happy not mad !

Thx mate!

Quote:


Originally Posted by bigval

Awesome work man, how come you just didnt leave the IHS off altogether? You couldve had even better temps without it. Nice job on the ocing!


Quote:


Originally Posted by brad62249

I was just thinking the same thing.....
thinking that the IHS would actuall hold in a little heat...as with out it like on the older AMD chips that having a little air being able to flow slightly around the chip and having direct contact with the heatsink would allow for better cooling...

In fact, the improvement of leaving the IHS off in comparison to sticking it back on isn't sufficiently significant to justify the risk of broking the core while it's naked. What causes the high temps isn't really the copper IHS but mostly the manufacturer thermal compound... if you substitute it with AS, or something equivalent (I'm also considering using liquid metal; but I really don't like it...), chances are high that you'll never experience a significant difference between this mod and the simple IHS removal.
Anyway, I'm using some kind of non-traditional cooling such as homemade waterblocks and thus I need to preserve safety.

Check this figure for temps differencies between stock IHS+thermal compound and IHS+AS Ceramic (with Air cooling):


Quote:


Originally Posted by iker0

wht did you use to stick the IHS back on?

In order to stick back on the IHS, I've used some kind of silicon usually used to make joins of sealing into engines. Actual on is called Loctite Autojoin.
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #7 ·
I ran a 3DMark 01 in order to break 32K with the TCCDs

32551pts | 3031MHz @1.55V Watercooled | TCCD 303MHz 2.5-4-3-5 @2.8V | X800XL 410/555MHz
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #8 ·

·
Premium Member
Joined
·
45,674 Posts
Very Impressive
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #10 ·
Thx mate!
 
  • Rep+
Reactions: Crash

·
Registered
Joined
·
456 Posts
Discussion Starter · #12 ·
I'm very happy... thank you so much!
 

·
Banned
Joined
·
230 Posts
what do you do with the area thats not being in contact with the IHS? i know that needs cooling aswell,but i dont think just thermal paste will do it.
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #14 ·
I've used silicone to substitute the original one in order to stick back on the IHS, and only AS5 over the core, but nothing more...
 

·
Registered
Joined
·
456 Posts
Discussion Starter · #15 ·
Some new results of Pi compting

PiFast 4.3 10M

28.83 | 3139MHz @1.42V chilled C2H5OH | BH5 @241MHz 1.5-2-2-5 @3.7V



Super PI 1M

26.188 | 3133MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-5 @3.7V

 

·
Registered
Joined
·
456 Posts
Discussion Starter · #17 ·
Some new results of Pi compting. Check these timings, there are really really tight!


PiFast 4.3 10M

I saw a 28.19s but it crashed and then rebooted immediately after and the saved .txt file didn't show the time...


28.26 | 3133MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-0 @3.7V



Super PI 1M

25.968 | 3138MHz @1.42V chilled C2H5OH | BH5 @261MHz 1.5-2-2-0 @3.7V

 

·
Registered
Joined
·
456 Posts
Discussion Starter · #19 ·
Not high enough
Let me install a new pump for the liquid-chiller and I might be able to find some sufficient stability to pass Pi 32M or OCCT @3.4GHz or above...
 
  • Rep+
Reactions: dangerousHobo

·
Registered
Joined
·
456 Posts
Discussion Starter · #20 ·
It seems me that I've a kind of bad luck guys!

(1) I've been waiting since the 30 December for 24 Opterons from my hardware components distributor, (2) received the chiller broken during the first shipping, (3) suffered from bad eBayer, (4) received the chiller again but temps were awful (required me to change the WB and the pump), and (5) now after some freezes the FX55 refuses to boot again and the 4 leds of the DFI are still on
 
1 - 20 of 56 Posts
Top