• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

kayls direct die project

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

kayl

Member
Joined
Nov 22, 2003
Location
australia
i had a go at insulating the direct die block, i used roof gutter sealant and some 17mm thick foam.
for the back i used neoprene and a coffee cup coaster as back plate, i hope to use another coaster as the top hold down
 

Attachments

  • testing 007.jpg
    testing 007.jpg
    145.6 KB · Views: 775
  • testing 010.jpg
    testing 010.jpg
    141.1 KB · Views: 661
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
did a little tunning last night and got idle -28c and after running 3dmark2001 at end of run it was -23c.
i couldnt go any higher than before when using tecs.
got to 11x223 on me locked barton at 2v.
either the chip is at the end or my power supply is no good.
i cant seem to use a vocre above 2.05v just crashes

i used about 7 foot of 0.026 cap tube and for refrigerant im using bbq r290
 

Attachments

  • testing.jpg
    testing.jpg
    142.6 KB · Views: 513
  • -48c 004.jpg
    -48c 004.jpg
    132.4 KB · Views: 506

nealric

Member
Joined
Sep 9, 2002
Location
under the floorboards
I like the insulation.
Is is silocone free?

I think you need a better chip to let that baby rip. You are probably motherboard/ram limited. Moble barton :)

You should be getting at least 2.8ghz on a mobile. I would sell the locked one.
 

ZachM

Member
Joined
Oct 11, 2003
Very cool. Sounds like a power supply problem. Above 2.05 vcore your PSU might just not be able to keep up. What PSU are you using?

Good mobile 2600s can do 2.8 with normal water cooling. Imagine with phase-change.:eek:
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
thanks nealric i used silicon to seal the foam to direct die and used dielectric grease all over the cpu, socket, back and front of insulation on board, i just hope it dosnt transfer the cold to mother board too much.

my ram is the most limiting fsb, on me last board got to 228 the epox had heaps more
only needed vdd 1.8v but that board died
few months ago, the new one has a fan on it, to do 220fsb i dont even have to increase vdd
i also have a barton 331 unlocked it does just over 2.6ghz on tecs, it run about 10c hotter though, in a few weeks i slap that in.
its a little hard to get mobiles here in australa, except for ordering in or buying on ebay, but i think once i have have had me fun with these 2 chips have to sell them both and buy a mobile and get 2.9ghz
i think when i select 2.05v it drops down to about 2v in bios
i think im goin have to do some readings with motherboard monitor to see how the vcroe goes, im currently using a 450w supply
 

nealric

Member
Joined
Sep 9, 2002
Location
under the floorboards
The reason I was asking about the silicone is because Im tryin to think how it would be possible to not use it. It makes mounting and unmounting a cpu a real chore. It also really messes with the mobo.
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
both,
i have a sensor on top of direct die
mother board monitor reads about 3 to 5 c lower
i have also takin readings at bottom near side of direct die and they are lower.
i need better contact between the sensor and the copper block though, i have a feeling it might be a little colder than i think not much though
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
nealric said:
The reason I was asking about the silicone is because Im tryin to think how it would be possible to not use it. It makes mounting and unmounting a cpu a real chore. It also really messes with the mobo.


i know that a few of the ppl at xtreme systems dont use dielectric grease at all any where. if there is no air in the socket, no condensation
they even run cascades reaching temps of -100c
a air tight socket will be safe. also not using grease reduces the amount of cold transfered to the mobo
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
today i ripped block off to check for condensation, there was 3 little beads resting on the silver sticky tape.:thumbdn:
not sure what im going to do about that and i though i did such a good job of insulating.
i also grabbed an old tv cabinent and organised the system a little
 

Attachments

  • direct die completepost.jpg
    direct die completepost.jpg
    137.9 KB · Views: 1,027
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
making the cooler is the easy part
i would have to buy a direct die block for the graphics card though
the ones i make are like half a kilo, thats why i have it laying on the bench and not mounted in a case
on my last graphics card i had a chiller cooling the gpu it died though, the card got warped out of shape because of the water block, dont want that to happen.

i also have to read up on how to oc this card, it needs software to oc it.
i also need to do some vmods to before i can do that
it would be nice though
best thing would be to use this cooler and make another direct die with a bigger compressor for better temps
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
today I managed to oc the barton to 11x230 with vcore of 1.97v
Using this chip and my last mobo I could only go to 11x223 and with a vocore of 2.1v
I pretty sure it’s the cooling factor that is enabling my xtra oc, as before the 226w tec had a hard trouble cooling the cpu.
My last epox was a rev1.x and this one is a rev3.x
I use to he able to use vcore as high as 2.2v (not that the cooling can handle it, but this rev3.x just wont boot over 2v (real 1.97v)
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
last night i put my unlocked barton in and put it to the test.
9600XT gecube graphics card clocked at 585/381 and have no vmods on it yeat, looking into that

with a clock speed of 2.7ghz managed to achiveve a 3dmarkscore 13242marks? why is it so low?
 

Attachments

  • bench 2.7ghz.jpg
    bench 2.7ghz.jpg
    38.2 KB · Views: 952
Last edited:
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
two compressors(8500BTU/6500BTU) one sucking the other as one big vacuum pump
with one compressor it use to only suck to 23Hg, now it sucks to 28Hg with is low enough to vaporize moisture out of the system.
idle at -26c top of evap and i18c mother board monitor
load top evap -21c and mother board monitor -13c with a 2.7GHZ barton

stock it does idle -26evap and -28c motherboard monitor, so that shows hows inaccurate mbm really is at -ve temps
 

Attachments

  • untitled.JPG
    untitled.JPG
    43.5 KB · Views: 927
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
i updated all the drivers (never done this before)
i use to get 13k now im getting 16k, im a little more happy with this score.
evap top temp -28c and -17c motherboard monitor, these were at the end of the run of 3dMark2001SE. i had 12 x 227 for 2724MHZ and 580.5 core and 782mem.
i also ran the computer stock speeds and evap top was -33c and motherboard monitor was -31c (idle though)
room temps 18c
 

Attachments

  • 16030 583 387 27c -17c.JPG
    16030 583 387 27c -17c.JPG
    95.1 KB · Views: 882
Last edited:
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
made a new hold down.
a shelve fram brace, just cut it at the 90 joiin and the holes matched the layout, fancy the ods ey
and this is me iposting at xtreme at 2724MHZ
 

Attachments

  • surfin internet.jpg
    surfin internet.jpg
    140.5 KB · Views: 840

Neco

Member
Joined
Nov 17, 2001
Location
Waterloo, Wisconsin, USA
Damnit, for some reason as I was scrolling the pic right above this post I thought it was a DOOM 3 screenshot.

How sad, lol

That looks pretty cool. Wish I had the $$$ to mess with stuff like this
 
OP
kayl

kayl

Member
Joined
Nov 22, 2003
Location
australia
heeh
i was playing doom3 yesterday too
ok this is what the system looks like now
im ready to charge it with gas


edit:
pic fixed
 

Attachments

  • system2.jpg
    system2.jpg
    44.5 KB · Views: 777
Last edited: