New waterblock

WinFlex · 03-11-2003, 08:45 AM

You may remember my initial design ideas a couple of weeks back. Here is a link with pics over at OC forums... scroll to the last post... my SN there is also WinFlex... enjoy, feel free to give feedback!

http://forum.oc-forums.com/vb/showth...8&pagenumber=2

Oh, also, If you want to read the full development thread, here it is:
http://forum.oc-forums.com/vb/showth...hreadid=171243

WinFlex · 03-11-2003, 10:52 AM

Uh, I believe this is worth checking out doodz

bigben2k · 03-11-2003, 11:26 AM

I've been following it for some time. In fact, I asked you (very early) what your theory/idea was behind it, but never got quite a clear answer. Going over it all last night, I see that the center pattern is a flow director.

All in all, I think you spent too much of your effort on the flow, and not enough on the thermal aspect, but in any case, I'm sure that it'll perform nicely.

WinFlex · 03-11-2003, 11:28 AM

Uh, correct me if i'm wrong, but isn't flow a decisive parameter in almost all thermal transfer equations..? could you explain further what you mean? Keep in mind this is desined for a pelt.

bigben2k · 03-11-2003, 12:01 PM

Yes, flow is important, but if you don't understand WHY it's important, then you're going around the issue. It's not as simple as "get it in and out, as easily/fast as possible".

Anytime the water turns or twists, you create a small area where the flow is turbulent. Otherwise the flow is laminar.

Turbulent flow makes it possible for the water to conduct heat better than in a laminar flow design.

In your block, your turbulent points are, strangely enough, where you least expect them: at the bottom of the "ramps", at that wall, coming down from the previous ramp (if it makes it that far).

You also have turbulent points at all those bends, way outside of where the core sits: those only restrict flow, they don't contribute to cooling.

I'm going to take a guess: your block will match a DD Maze 3's performance.

WinFlex · 03-11-2003, 12:22 PM

uhm... ok. i guess all them darn fluid dynamics classes I took didn't help much. let's see what you wrote again: there is turbulence over the core, u admit that... good. then, when using a pelt, you want even heat transfer throughout the block, right. that's why there is evenly distributed turbulent bends on the outer diameter and a relatively thick base plate...
saying this block will perform like a maze3 is garbage, it already is 7 degrees K cooler than my old maze3 during preliminary testing, even at higher than normal voltages.
also, regarding flow. I modeled my whole system, and the actual bottleneck with over 80 percent of the preassure drop is my D-tek customs radiator, not the block

Voila
also, I kept "machinability" in mind when designing this block... u seemingly haven't done that in your fantastic design! Getting that milled on an EDM is going to set you back hundreds, if not thousands... My block took about 30 minutes milling time on a computer controlled 3 axis machine...

bigben2k · 03-11-2003, 12:32 PM

Oh yes... I keep forgetting about the Pelt!

If you're getting 7 degrees better than the Maze3, then aren't you beating WW?!?

If the design isn't very restrictive (which I didn't expect it to be), then all the best! You can find a bigger heatercore in the database: it should be less restrictive.

In my design, machinability was kept in mind, but pushed "up there"

WinFlex · 03-11-2003, 12:37 PM

WW? I also think this is mighty high. I am trying to isolate any other variables, but except for the thermal paste, I changed nothing... I went from arctic alumina to AS3, which usually only accounts for 1 or 2 degrees K...
Perhaps a difference in mounting preassure is also a factor, although I am using the exact same hardware... perhaps I tightened the screws more?

theetruscan · 03-11-2003, 03:47 PM

That is a beautiful block, thanks for sharing.

I'm confused by ben though.

What does a pelt blocks performance (Winflex's beauty) have to do with the performance of a CPU block (the LRWW block)?

And a question.

I don't understand the nature of pelts well. Will a better block show more performance increase with a "bigger" (higher temp differential) pelt, or vice versa, or about the same?

Again, thanks for sharing man, that's beautiful.

pippin88 · 03-11-2003, 08:41 PM

BB fogot that this was a block to cool pelts, and though it was just a straight water block for CPUs, and his observations in the first post apply to this.

It is a beautifull block though.

WinFlex · 03-11-2003, 09:00 PM

Quote:

I don't understand the nature of pelts well. Will a better block show more performance increase with a "bigger" (higher temp differential) pelt, or vice versa, or about the same?

I believe all TEC's actually have the same max temperature difference. The delta T between the cold side and warm side can not exceed 68 degrees celsius, if I remember correctly...

winewood · 03-11-2003, 11:42 PM

winflex.. that blocks sexy...
I bet you would have great performance on a pelt cooling. very large coverage of heat gathering area.

03-11-2003, 08:45 AM	#1
WinFlex Cooling Neophyte Join Date: Feb 2003 Location: Boston, MA, USA Posts: 51	New waterblock You may remember my initial design ideas a couple of weeks back. Here is a link with pics over at OC forums... scroll to the last post... my SN there is also WinFlex... enjoy, feel free to give feedback! http://forum.oc-forums.com/vb/showth...8&pagenumber=2 Oh, also, If you want to read the full development thread, here it is: http://forum.oc-forums.com/vb/showth...hreadid=171243

03-11-2003, 11:26 AM	#3
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	I've been following it for some time. In fact, I asked you (very early) what your theory/idea was behind it, but never got quite a clear answer. Going over it all last night, I see that the center pattern is a flow director. All in all, I think you spent too much of your effort on the flow, and not enough on the thermal aspect, but in any case, I'm sure that it'll perform nicely. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-11-2003, 12:01 PM	#5
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Yes, flow is important, but if you don't understand WHY it's important, then you're going around the issue. It's not as simple as "get it in and out, as easily/fast as possible". Anytime the water turns or twists, you create a small area where the flow is turbulent. Otherwise the flow is laminar. Turbulent flow makes it possible for the water to conduct heat better than in a laminar flow design. In your block, your turbulent points are, strangely enough, where you least expect them: at the bottom of the "ramps", at that wall, coming down from the previous ramp (if it makes it that far). You also have turbulent points at all those bends, way outside of where the core sits: those only restrict flow, they don't contribute to cooling. I'm going to take a guess: your block will match a DD Maze 3's performance. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-11-2003, 12:22 PM	#6
WinFlex Cooling Neophyte Join Date: Feb 2003 Location: Boston, MA, USA Posts: 51	uhm... ok. i guess all them darn fluid dynamics classes I took didn't help much. let's see what you wrote again: there is turbulence over the core, u admit that... good. then, when using a pelt, you want even heat transfer throughout the block, right. that's why there is evenly distributed turbulent bends on the outer diameter and a relatively thick base plate... saying this block will perform like a maze3 is garbage, it already is 7 degrees K cooler than my old maze3 during preliminary testing, even at higher than normal voltages. also, regarding flow. I modeled my whole system, and the actual bottleneck with over 80 percent of the preassure drop is my D-tek customs radiator, not the block Voila also, I kept "machinability" in mind when designing this block... u seemingly haven't done that in your fantastic design! Getting that milled on an EDM is going to set you back hundreds, if not thousands... My block took about 30 minutes milling time on a computer controlled 3 axis machine... Last edited by WinFlex; 03-11-2003 at 12:29 PM.

03-11-2003, 12:32 PM	#7
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Oh yes... I keep forgetting about the Pelt! If you're getting 7 degrees better than the Maze3, then aren't you beating WW?!? If the design isn't very restrictive (which I didn't expect it to be), then all the best! You can find a bigger heatercore in the database: it should be less restrictive. In my design, machinability was kept in mind, but pushed "up there" __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-11-2003, 10:52 AM	#2
WinFlex Cooling Neophyte Join Date: Feb 2003 Location: Boston, MA, USA Posts: 51	Uh, I believe this is worth checking out doodz

03-11-2003, 11:28 AM	#4
WinFlex Cooling Neophyte Join Date: Feb 2003 Location: Boston, MA, USA Posts: 51	Uh, correct me if i'm wrong, but isn't flow a decisive parameter in almost all thermal transfer equations..? could you explain further what you mean? Keep in mind this is desined for a pelt.

03-11-2003, 12:37 PM	#8
WinFlex Cooling Neophyte Join Date: Feb 2003 Location: Boston, MA, USA Posts: 51	WW? I also think this is mighty high. I am trying to isolate any other variables, but except for the thermal paste, I changed nothing... I went from arctic alumina to AS3, which usually only accounts for 1 or 2 degrees K... Perhaps a difference in mounting preassure is also a factor, although I am using the exact same hardware... perhaps I tightened the screws more?

03-11-2003, 03:47 PM	#9
theetruscan Cooling Neophyte Join Date: Jan 2003 Location: OR/CA/NY Posts: 81	That is a beautiful block, thanks for sharing. I'm confused by ben though. What does a pelt blocks performance (Winflex's beauty) have to do with the performance of a CPU block (the LRWW block)? And a question. I don't understand the nature of pelts well. Will a better block show more performance increase with a "bigger" (higher temp differential) pelt, or vice versa, or about the same? Again, thanks for sharing man, that's beautiful.

03-11-2003, 08:41 PM	#10
pippin88 Cooling Savant Join Date: Oct 2002 Location: Australia Posts: 285	BB fogot that this was a block to cool pelts, and though it was just a straight water block for CPUs, and his observations in the first post apply to this. It is a beautifull block though.

03-11-2003, 11:42 PM	#12
winewood Cooling Savant Join Date: Nov 2002 Location: in my chair Posts: 574	winflex.. that blocks sexy... I bet you would have great performance on a pelt cooling. very large coverage of heat gathering area. __________________ -winewood-

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