Direct Die Cooling Project. - Page 2

maxim · 08-22-2003, 01:23 AM

Arcturius, do you mean that your core wore out exposing the actual chip? and it still works? are you using just distilled water with nothing added?

MadDogMe · 08-22-2003, 04:03 AM

I'm not getting any pics in your third post JD, are they on a different server than the second posts pic?. How about a link to a relevant thread somewhere (pimprigs?) as I missed the original 'round'

, shame cause I like a good bitchfight as well as the next dude, (as long as it is'nt onesided abuse) it's good clean fun

...

What is the score with AMD's 'not being waterproof'? I can't make out if people mean it's the 'organic substrate' (which is obvious) or the silicon core wafers themselves? or the edges where the wafers are 'stacked'? (which could be sealed as aforementioned)...

PS, welcome back!

...

jaydee · 08-22-2003, 07:52 AM

Quote:

Originally posted by MadDogMe
I'm not getting any pics in your third post JD, are they on a different server than the second posts pic?. How about a link to a relevant thread somewhere (pimprigs?) as I missed the original 'round' , shame cause I like a good bitchfight as well as the next dude, (as long as it is'nt onesided abuse) it's good clean fun ...

What is the score with AMD's 'not being waterproof'? I can't make out if people mean it's the 'organic substrate' (which is obvious) or the silicon core wafers themselves? or the edges where the wafers are 'stacked'? (which could be sealed as aforementioned)...

PS, welcome back! ...

The pics should be there now. My host has had some problems lately. Good thing is their main site is on the same server my site is on so they are quick to fix it.

Non of it is really water proof but the biggest issue is where the body of the cpu meets the die. The peices are not bonded together and they use somekind of sealant around that. If you look close you will see a greenish sealant that looks as if it was poured on over the die and it seaps onto the body of the CPU.

As for the the thread I pulled those quotes from it was all one sided on their site. I never said a word in that thread nor will I ever think about joining that rediculous site. The guy was just bent out of shape because he got questioned on his testing methods after he posted his article here that had rediculous temps with no water or air temps shown and he claims his overclock is the determining factor for how good it is. He started a thread there to hear is ignorant followers praise him for something they don't really understand. This alone is why I stopped posting my projects at other sites. Here I know most people can understand and really appriciate the work I do for the most part.

Arcturius: Interesting that the die is visably eroding. Just proves that sooner or later it will wear through. Well I woke up a freaking 5am this morning for some reason. It is 5:50am now and I still need to test the Duron to make sure it works so I will do that now before I got to work. If it does work I will be testing out what my Lemon Block can do and then the maze 4 and then the direct die block. The reason for the other two blocks is I need a comparision on the same rig and I am making an article about this whole project for my website. I want to cover as many bases as I can before posting an article.

jaydee · 08-22-2003, 08:22 AM

Duron works! typing this on the very computer it is working on. Run nice and cold to with this Lemon Block. Going to see if my pencil trick is holding up and see if I can OC this thing before going to work.

jaydee · 08-22-2003, 08:52 AM

Well overclocking with the multi is a no go. I guess I need a better pencil! Will work on that after work. I had it up to 1250mhz stable and 1300mhz not so stable at one time on the older Epox 8K7A. Once I get the multi unlocked it should go up to those levels again. I got it at 1150mhz at 115FSB right now. Once the multi is unlocked I can drop the multi down the 7.5X at 133FSB to get 1000mhz and will work my way on up to 1300 or so.

MMZ_TimeLord · 08-22-2003, 10:48 AM

Good looking DDC block there Jaydee... I especially like your reliefs for the various components on the top of the CPU.

As for the hardheaded and blind responses from that "other" forum, I would chalk that up to ignorance. They obviously are content to use "133t" tricks such as putting their radiator outside for colder air to cool their water as opposed to actually making a water block/DDC block that is an excellent performer.

I have yet to see a shred of evidence that shows that they "pwnd" you. All they show is that they have what they BELIEVE is an excellent direct die cooling setup that "must" outperform anything out there because they can "0C the 5h1t" out of it.

I personally don't give a damn if they think you are right or not. I've been very impressed with your creative work so far as well as your very strict proceedures for testing. Such as... keeping the room tempretures level and "normal". I can only imagine the variances that guy gets with the radiator hanging outside.

:shrug:

Keep up the good work Jaydee... "Nuthin' but luv for yah!"

jaydee · 08-22-2003, 09:00 PM

Thanks MMZ!

Bad news though. I got home, found a good pencil, unlocked the Duron. Slapped on an air cooler I had laying around (because it is a pain to mount the water block just to find it still needs unlocking), Set it to 1333mhz and it was stable in windows air cooled!

SO, I took the air cooler off and was tightening down the Lemon Block and CRACK!. Chipped the corner and killed it. I knew I should have put those 4 pads back on first.... Damn... Well this puts a halt to this project. The block is designed for a Duron1gig-1.3gig but they are like $35 new at the egg. Not sure I want to buy one just for this. Would almost rather wait a while untill I can afford a new XP1700+ and build a new Direct Die block for it. goingto check e-bay for cheap Durons....

jaydee · 08-22-2003, 10:24 PM

Code:

CPU AMD|DURON 1.3GHZ %  1    $35.00 
        
  Subtotal $35.00 
  Tax $.00 
Shipping and Handling charge
 .00 
Total $35.00 
 
 
Step 5 of 5

Well this will make it more interesting anyway. I looked on ebay and they were MORE expensive than the egg BEFORE shipping cost.

$35 shipped from newegg.... Should be here the end of next week....

Phant0m51 · 08-22-2003, 11:11 PM

Now wait a minute, you were air cooling your CPU without those pads on there? Does it help performance taking off the pads? I've never heard of this before...

t00lb0x · 08-23-2003, 01:01 AM

Yeah, take off pads and put on thermal paste.

jaydee · 08-23-2003, 10:27 AM

Quote:

Originally posted by Phant0m51
Now wait a minute, you were air cooling your CPU without those pads on there? Does it help performance taking off the pads? I've never heard of this before...

No. no. no. I had to take the pads off to build the Direct Die block as it sits flat on the CPU body. Leave the pads on.

bretb · 08-23-2003, 10:59 AM

JayDee you have mail.

jaydee · 08-23-2003, 11:12 AM

Quote:

Originally posted by bretb
JayDee you have mail.

Thanks, just replied.

JFettig · 08-24-2003, 01:31 AM

hey jaydee, thanx for reminding me about my 2 duron 850s I have

I gotta make a direct die block for one of those sometime.

One thing you could do is cut holes for the pads and not take them off. Thats one thing I think I would do.

So far, it looks good. I really didnt look at the pics that thouroughly. though. That duron have a squair core? Mine dont....

One thing that you have to remember is the offset in the socket.

Also, in order to get your cpu out, your gonna need to take the waterblock off. The cam lever will get in the way. Thats always a pain in the butt. The way I solved that on my old direct die block was I just made a 2x2 block and a hold down.

Jon

jaydee · 08-24-2003, 09:56 AM

Quote:

Originally posted by JFettig
hey jaydee, thanx for reminding me about my 2 duron 850s I have

I gotta make a direct die block for one of those sometime.

One thing you could do is cut holes for the pads and not take them off. Thats one thing I think I would do.

So far, it looks good. I really didnt look at the pics that thouroughly. though. That duron have a squair core? Mine dont....

One thing that you have to remember is the offset in the socket.

Also, in order to get your cpu out, your gonna need to take the waterblock off. The cam lever will get in the way. Thats always a pain in the butt. The way I solved that on my old direct die block was I just made a 2x2 block and a hold down.

Jon

The block fits perfectly. I used a dead 8K7A to line everything up when building it. Note the first pic it is mounted to the CPU and ready to go, juist needs the sealing. The core isn't square, it is rectangular but it is rotated 90 degrees compared to the newer XP's and a little larger. I am going to put the block together, put the CPU in the socket, silicon the top to the CPU and put the block on using the 4 holes to apply the pressure for the seal on top o the die. It is a little over an 1" thick so it is plenty stout.

Volenti · 08-25-2003, 08:05 AM

Regarding the core erosion issue, after my p4 died after 4 months of direct die (exact cause unkown since other extreme mobo mods we're in place as well) link it had...marks on the core that corosponded to the impact points of the jet's, but they we're so faint that a wipe over the core removed them, mabie a deposit build up of some kind.

I have plans in place for a T-bred direct die exprimental block (quite similar to your approach actually) that also will try and address the porus core sealant issue.

ZaVkE · 08-25-2003, 08:18 AM

About those jets:

Will more jets get you lower temps?
Otherwise it could be possible to make them more narrow and have more of them, hitting more surface with their (smaller) beams.

This is a large forum, and the word jet is used a lot here

So i couldn't find anything that tasted which is better and why ...

jaydee · 08-25-2003, 09:05 AM

Quote:

Originally posted by ZaVkE
About those jets:

Will more jets get you lower temps?
Otherwise it could be possible to make them more narrow and have more of them, hitting more surface with their (smaller) beams.

This is a large forum, and the word jet is used a lot here
So i couldn't find anything that tasted which is better and why ...

The Die is pretty small. Not much room to add more. Out of my capability at this point anyway.

jaydee · 08-26-2003, 11:56 AM

I just can't get over this. Been laughing my ass off all morning.

Quote:

Guys....I bought (company name with-held) which holds the patent rights to direct-die waterblocks?

I'm unsure as to what to do here - love your work, but kinda makes my $$$$ investment meaningless?

Please contact me concerning this - I hate lawyers.

Cathar got an e-mail stating that yet Cathar doesn't even have a direct-die block.

For some reason the guy is of the opinion that he has all rights for direct die cooling. All he has is rights for the designs his company has made and patented. Hope he comes across this thread.

Anyway no progress yet. CPU is still in route. Should be here tomorrow or Thursday.

Also bretb was kind enough to send a T-Bird 1200 for this. Once that gets here it will be direct die project #2! I can't wait to get this block up and running. Giving me a whole new reason to carry on in this water cooling game.

Once I get back on my feet money wise I am going to build a water chiller and run cold water through the direct block.

jaydee · 08-26-2003, 01:16 PM

LOL http://arstechnica.infopop.net/OpenT...285#8620911285 LOL

t00lb0x · 08-26-2003, 01:28 PM

Quote:

Originally posted by jaydee116
I just can't get over this. Been laughing my ass off all morning.

Cathar got an e-mail stating that yet Cathar doesn't even have a direct-die block.

For some reason the guy is of the opinion that he has all rights for direct die cooling. All he has is rights for the designs his company has made and patented. Hope he comes across this thread.

Anyway no progress yet. CPU is still in route. Should be here tomorrow or Thursday.

Also bretb was kind enough to send a T-Bird 1200 for this. Once that gets here it will be direct die project #2! I can't wait to get this block up and running. Giving me a whole new reason to carry on in this water cooling game.

Once I get back on my feet money wise I am going to build a water chiller and run cold water through the direct block.

WoW cold water through a direct die block. JD sometimes the method behind hte madman is just crazy (makes sense, NO!) You'd probably get incredible temps. And that'll be a good expiriment. Can't wait to see the results.

Althornin · 08-26-2003, 07:49 PM

I still think lack of surface area will make a DD block perform worse than the "best" regular WB....
The only question is, will it perform enough worse to overcome the fact that it has no crappy TIM (and thus lose) or will it win out...and will a better TIM (like that carbon black stuff) cause DD to lose again?

Cathar · 08-26-2003, 08:17 PM

Good work jaydee.

If you're interested, I'll see if I can dig up a reject Cascade middle plate. Some come off the machine with a few tubes split - if more than 2 are split then I reject the plate, however I may be able to find one where the damage is restricted to some outer tubes and should be easy enough to work around.

All you need to do then is make up some templates that block off the top entrances to the jets that don't touch the CPU die, and basically you can re-use the one plate across a variety of dies by doing this.

Let me know and I'll find a middle and top plate to send your way if possible. Just make up a bottom plate to stand the jets away from the CPU die and to seal to the CPU and you're in business.

BTW, I have tried the Cascade with 1C water (ice blocks in an esky) and picked up an extra 100MHz out of the CPU. I was running my XP2800+ Barton at over 2800MHz without worries.

BrianW · 08-26-2003, 08:26 PM

Whats an esky? A portable insulated storage box, which we call a cooler here in the states?

BrianW

jaydee · 08-26-2003, 08:36 PM

Yes, Cathar I would be interested in that. I think I could make good use of that on the TBird core as if I remember right they are fairly large. I can cap off the uneeded pipes.

I think the biggest gain this might have over a standard block is the +-8C gain from having no TIM joint. If I can get the volocity high enough through the jets it should work well. Water shouldn't have any problem cooling the die as long as it is moving fairly quick.

Anyway I will test it against my Lemon Block and the Maze 4. And then I will let it run for a few months 24/7 or until something breaks.

I got my test bench cleared and ready to go when the CPU gets here.

08-22-2003, 10:48 AM	#31
MMZ_TimeLord Cooling Savant Join Date: Apr 2003 Location: San Mateo, CA, USA, Earth Posts: 433	Good looking DDC block there Jaydee... I especially like your reliefs for the various components on the top of the CPU. As for the hardheaded and blind responses from that "other" forum, I would chalk that up to ignorance. They obviously are content to use "133t" tricks such as putting their radiator outside for colder air to cool their water as opposed to actually making a water block/DDC block that is an excellent performer. I have yet to see a shred of evidence that shows that they "pwnd" you. All they show is that they have what they BELIEVE is an excellent direct die cooling setup that "must" outperform anything out there because they can "0C the 5h1t" out of it. I personally don't give a damn if they think you are right or not. I've been very impressed with your creative work so far as well as your very strict proceedures for testing. Such as... keeping the room tempretures level and "normal". I can only imagine the variances that guy gets with the radiator hanging outside. :shrug: Keep up the good work Jaydee... "Nuthin' but luv for yah!" __________________ MMZ>TimeLord "Oooooooooh... that's gonna leave a mark!"

08-22-2003, 10:24 PM	#33
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Code: CPU AMD\|DURON 1.3GHZ % 1 $35.00 Subtotal $35.00 Tax $.00 Shipping and Handling charge .00 Total $35.00 Step 5 of 5 Well this will make it more interesting anyway. I looked on ebay and they were MORE expensive than the egg BEFORE shipping cost. $35 shipped from newegg.... Should be here the end of next week....

08-23-2003, 01:01 AM	#35
t00lb0x Cooling Savant Join Date: May 2003 Location: In a box Posts: 221	Yeah, take off pads and put on thermal paste. __________________ "If you can't dazzle them with brilliance, baffle them with bullshit"

08-25-2003, 08:05 AM	#41
Volenti Cooling Savant Join Date: Jan 2002 Location: in a nice cool spot Posts: 427	Regarding the core erosion issue, after my p4 died after 4 months of direct die (exact cause unkown since other extreme mobo mods we're in place as well) link it had...marks on the core that corosponded to the impact points of the jet's, but they we're so faint that a wipe over the core removed them, mabie a deposit build up of some kind. I have plans in place for a T-bred direct die exprimental block (quite similar to your approach actually) that also will try and address the porus core sealant issue. __________________ feel free to icq/msn me, I'm always willing to toss around ideas.

08-25-2003, 08:18 AM	#42
ZaVkE Cooling Neophyte Join Date: Jun 2002 Location: Belgium Posts: 83	About those jets: Will more jets get you lower temps? Otherwise it could be possible to make them more narrow and have more of them, hitting more surface with their (smaller) beams. This is a large forum, and the word jet is used a lot here So i couldn't find anything that tasted which is better and why ... __________________ signatures are for wussies

08-22-2003, 01:23 AM	#26
maxim Cooling Savant Join Date: Jan 2003 Location: VA Tech Posts: 111	Arcturius, do you mean that your core wore out exposing the actual chip? and it still works? are you using just distilled water with nothing added?

08-22-2003, 04:03 AM	#27
MadDogMe Thermophile Join Date: Aug 2002 Location: Just shut up ;) ... Posts: 1,068	I'm not getting any pics in your third post JD, are they on a different server than the second posts pic?. How about a link to a relevant thread somewhere (pimprigs?) as I missed the original 'round' , shame cause I like a good bitchfight as well as the next dude, (as long as it is'nt onesided abuse) it's good clean fun ... What is the score with AMD's 'not being waterproof'? I can't make out if people mean it's the 'organic substrate' (which is obvious) or the silicon core wafers themselves? or the edges where the wafers are 'stacked'? (which could be sealed as aforementioned)... PS, welcome back! ...

08-22-2003, 08:22 AM	#29
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Duron works! typing this on the very computer it is working on. Run nice and cold to with this Lemon Block. Going to see if my pencil trick is holding up and see if I can OC this thing before going to work.

08-22-2003, 08:52 AM	#30
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Well overclocking with the multi is a no go. I guess I need a better pencil! Will work on that after work. I had it up to 1250mhz stable and 1300mhz not so stable at one time on the older Epox 8K7A. Once I get the multi unlocked it should go up to those levels again. I got it at 1150mhz at 115FSB right now. Once the multi is unlocked I can drop the multi down the 7.5X at 133FSB to get 1000mhz and will work my way on up to 1300 or so.

08-22-2003, 09:00 PM	#32
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Thanks MMZ! Bad news though. I got home, found a good pencil, unlocked the Duron. Slapped on an air cooler I had laying around (because it is a pain to mount the water block just to find it still needs unlocking), Set it to 1333mhz and it was stable in windows air cooled! SO, I took the air cooler off and was tightening down the Lemon Block and CRACK!. Chipped the corner and killed it. I knew I should have put those 4 pads back on first.... Damn... Well this puts a halt to this project. The block is designed for a Duron1gig-1.3gig but they are like $35 new at the egg. Not sure I want to buy one just for this. Would almost rather wait a while untill I can afford a new XP1700+ and build a new Direct Die block for it. goingto check e-bay for cheap Durons....

08-22-2003, 11:11 PM	#34
Phant0m51 Cooling Savant Join Date: May 2003 Location: Utah Posts: 160	Now wait a minute, you were air cooling your CPU without those pads on there? Does it help performance taking off the pads? I've never heard of this before...

08-23-2003, 10:59 AM	#37
bretb Cooling Neophyte Join Date: May 2002 Location: S. F. Bay Area Posts: 8	JayDee you have mail.

08-24-2003, 01:31 AM	#39
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	hey jaydee, thanx for reminding me about my 2 duron 850s I have I gotta make a direct die block for one of those sometime. One thing you could do is cut holes for the pads and not take them off. Thats one thing I think I would do. So far, it looks good. I really didnt look at the pics that thouroughly. though. That duron have a squair core? Mine dont.... One thing that you have to remember is the offset in the socket. Also, in order to get your cpu out, your gonna need to take the waterblock off. The cam lever will get in the way. Thats always a pain in the butt. The way I solved that on my old direct die block was I just made a 2x2 block and a hold down. Jon

08-26-2003, 01:16 PM	#45
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	LOL http://arstechnica.infopop.net/OpenT...285#8620911285 LOL

08-26-2003, 07:49 PM	#47
Althornin Cooling Savant Join Date: Dec 2002 Location: Houston, TX, USA Posts: 221	I still think lack of surface area will make a DD block perform worse than the "best" regular WB.... The only question is, will it perform enough worse to overcome the fact that it has no crappy TIM (and thus lose) or will it win out...and will a better TIM (like that carbon black stuff) cause DD to lose again?

08-26-2003, 08:17 PM	#48
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Good work jaydee. If you're interested, I'll see if I can dig up a reject Cascade middle plate. Some come off the machine with a few tubes split - if more than 2 are split then I reject the plate, however I may be able to find one where the damage is restricted to some outer tubes and should be easy enough to work around. All you need to do then is make up some templates that block off the top entrances to the jets that don't touch the CPU die, and basically you can re-use the one plate across a variety of dies by doing this. Let me know and I'll find a middle and top plate to send your way if possible. Just make up a bottom plate to stand the jets away from the CPU die and to seal to the CPU and you're in business. BTW, I have tried the Cascade with 1C water (ice blocks in an esky) and picked up an extra 100MHz out of the CPU. I was running my XP2800+ Barton at over 2800MHz without worries.

08-26-2003, 08:26 PM	#49
BrianW Cooling Savant Join Date: Jun 2001 Location: Desert City in California Posts: 631	Whats an esky? A portable insulated storage box, which we call a cooler here in the states? BrianW __________________ Water Cooled Inwin Q500 (Dual Rads: Rad1 = DTEK Pro Core \| Rad2 = Blick Ice Estreme, Hydor L30, Dangerden Maze2, Bay Res Typhoon Reservoir, 1/2 " DD Tygon Thick Wall Hose). Flow: Res, Pump, CPU watervlock, Y into both rads, both rads into res independently. Athlon XP 1800+ (@ 1731 - 150mhz fsb.), on a Asus A7N266-c, and a Radeon 9000 *waiting for RMA'd Saphire 9800 ultra from Newegg)

08-26-2003, 08:36 PM	#50
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Yes, Cathar I would be interested in that. I think I could make good use of that on the TBird core as if I remember right they are fairly large. I can cap off the uneeded pipes. I think the biggest gain this might have over a standard block is the +-8C gain from having no TIM joint. If I can get the volocity high enough through the jets it should work well. Water shouldn't have any problem cooling the die as long as it is moving fairly quick. Anyway I will test it against my Lemon Block and the Maze 4. And then I will let it run for a few months 24/7 or until something breaks. I got my test bench cleared and ready to go when the CPU gets here.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)