New design idea. - Page 2

bigben2k · 03-21-2003, 10:02 AM

This would be more in line with what I'm thinking:

(The top has both a small and a large barb glued to it. The flow is then seperated later)

Gilhooley · 03-21-2003, 10:22 AM

bigben2k:

Yes that would also work. However, there is much more work with your top. Designing a leak free separator that lies firm against your jet holes is a daunting task

utabintarbo · 03-21-2003, 10:33 AM

Quote:

Originally posted by bigben2k
Going over the designs presented here (and elsewhere), notably Nicozeg, and a few #rotor style ones, I came up with a theory/idea for another block (be nice if I finished Radius first!):

In short, the idea is to combine jet inpingement, with a drilled block.

Goddammit, Ben! Get out of my mind!

I too have been looking into a "morphing" of #rotor's and cathar's ideas into something I can accomplish with a drill press and a sliding (x-y) table. Pics to come.

Bob

bigben2k · 03-21-2003, 10:36 AM

Actually, the seperator wouldn't be up on top of the block, I would opt to have it up against the opposite side of the case (unlike the monstrosity I built for Radius).

I'm going to refer you to Morphling1's excellent drawing here.

Gilhooley · 03-21-2003, 11:11 AM

You could enlarge the area around the die spot in block 2. Then the separator would have few if any advantages. Pertaining higher flow rate that is. (ooooh NYPD Blue

)

bigben2k · 03-21-2003, 11:41 AM

LOL@Bob! Didn't I include you in on my mind-reading machine?

McMaster p/n for copper tubing:
1/16: 7190K51
3/32: 7190K52

I should make a nice smooth opening for those tubes, to promote flow.

Cathar · 03-21-2003, 07:33 PM

Been thinking a lot on this, and drawing up some plans/schematics/rules-of-thumb/trigonometry.

Can be done using a square grid array.

Pick a drill bit size to do the dimples with. Let's choose 3/32" (~2.4mm)

Sides on elements of the grid should by 5/6 (exactly) of the drill bit size (or 5/64" for our 3/32" drill bit size)

This creates rough diamond shaped pins that are about 0.25 of the grid distance on a side (in this example around 5/256" or about 0.5mm on a side). These pins are going to be important as structural supports for the base-plate.

Plate thickness should be twice as thick as the drill bit size (3/16" in this case). Drill depth should be 1.5x of the drill bit size to the bottom of the drill holes.

This creates the dimples, the pins, and enough channel area between the pins for the water to flow out.

A dimpled area of 5/8" x 5/8" will be adequate for most any CPU. Need to create side outlet channels to enable middle in, dual-side outlet config.

Inlet jet plate will have jet holes to match the dimples. Jet holes should be 1/3 of the drill bit size in width (1/32")

For a 3/32" dill bit grid array over a 5/8" x 5/8" area this creates basically a 8 x 8 grid. Total jet orifice area is around 32mm^2 which is restrictive enough to create a very good jet velocity with something like an Eheim 1250 pump.

Ben, got any objections if I get my machinists to make up the above on a CNC mill?

bigben2k · 03-21-2003, 08:35 PM

No, by all means, go right ahead!

But why are you dropping the honeycomb? Convenience of calcs?

Also, you might have missed it, but I mentionned using a square ended endmill, as a drill bit, to dodge the cone shape. Would your calculations differ significantly?

Cathar · 03-21-2003, 08:38 PM

Dropping honeycomb because I'm thinking of rotating the grid by 45 degrees.

I'm actually thinking that the coned shape for the dimples is a good thing.

SysCrusher · 03-21-2003, 09:27 PM

The cone shape of the dimples is what you need. The impingement presses against the side of the angle in the dimple as it flows up that cone shape promoting an even thinner boundary layer. The reason I did my baseplate with convex.

Volenti · 03-22-2003, 08:17 AM

Ok I have a quick prototype ready to go;

BP is 4.5mm thick and the holes (bit too deep to be called dimples

) are 3.5mm deep and 3mm in dia, the jet's are 1.5mm in dia.

Some of you may recognise the top part of the block, since it's cut down from my p4 direct die watercooling block, it has multiple layers, with diagonally opposing dual in and out.

I'll hook it up to my t-bred 2100+@2.4gig tomorrow.

bigben2k · 03-22-2003, 08:53 AM

Quote:

Originally posted by SysCrusher
The cone shape of the dimples is what you need. The impingement presses against the side of the angle in the dimple as it flows up that cone shape promoting an even thinner boundary layer. The reason I did my baseplate with convex.

You may be right. The cone shape adds a significant amount of surface area, vs a square ended endmill.

LiquidRulez · 03-22-2003, 07:00 PM

BEN,
While I like the simplicity of this idea, I think you should NOT give up on Radius(or are you just getting bored?).
I think Radius has alot of potential and will be simply amazing in performance once you complete it....Not to mention, beautiful too with an LED lighting it, sitting on top of an AMD chip!

Volenti · 03-22-2003, 09:40 PM

Ok I've got the impingement block set up;

I must say I'm quite impressed, it matches my micropin block in performance

using the latest verson of MBM to record of the XP's internal diode, allthough the absolute reading is off (as usual), relitive readings seem to be accurate enough for my purposes here.

leejsmith · 03-23-2003, 04:15 AM

This is an idea i had for a water delivery system in the wb.

cut channels into a star and then if possible make an outer circle and an inner. Then drill the holes into the channles. If made correct you should be able to direct the jets at the cones in the base plate.

then just place it under the inlet barb as a middle layer.

if a little more effort and thought was put into it i am sure it will work.

bigben2k · 03-23-2003, 08:27 AM

Quote:

Originally posted by LiquidRulez
BEN,
While I like the simplicity of this idea, I think you should NOT give up on Radius(or are you just getting bored?).
I think Radius has alot of potential and will be simply amazing in performance once you complete it....Not to mention, beautiful too with an LED lighting it, sitting on top of an AMD chip!

Oh not to worry, I haven't given up! I'm just stuck in a loop, where I need a good PC to get a DXF out of SW2K3, and the block for the PC

(and funds, versus getting a car!)

Actually, while developing Radius, I've imagined many designs, including the ProCooling block, but this one is one I thought was worthy of mention.

Nice work Volenti! Glad to see my idea put to the test so quickly, and work well too! With a BP thickness of 4.5, did you consider trying out less? What pump did you use?

Volenti · 03-23-2003, 10:00 PM

Quote:

Originally posted by bigben2k

Nice work Volenti! Glad to see my idea put to the test so quickly, and work well too! With a BP thickness of 4.5, did you consider trying out less? What pump did you use?

The bottom of the jet's in the perxpex top is actually a shallow chamber, so there's no support for a thin bp, the holes in the bp penetrate down to within 1mm of the bottom though, the fact that it's performing as well as it is proves that we got something right at least (:

I'll try a couple of different thicknesses of BP and hole/dimple setups (including simple featureless bp's) though I may not be able to see much of a difference with the equipment I have...

the pump's an iwaki MD-20RZ

JFettig · 03-24-2003, 12:35 PM

volenti, Do you by chance have any pictuers of your jet in action? without the base attached to see the spray?

Jon

Volenti · 03-25-2003, 01:39 AM

Quote:

Originally posted by JFettig
volenti, Do you by chance have any pictuers of your jet in action? without the base attached to see the spray?

Jon

The camera flash freezes the action a little too well, the streams look more fluid in real life.

g.l.amour · 03-25-2003, 03:59 AM

whoaah, xplosive action there.

i still cant figure out too well how you are doing the exit channels of the water. care to give us some build drawings or something.

how far can the jet exits be from the beginning of the baseplate. atm i haev a copper stock with bars of 1cm or 1/3" thickness. if i go for a BP thickness of 3mm, then there is 7mm drilled away. can the jet effect reach over that 7mm void, or will effectiveness be lost.

thx

Volenti · 03-25-2003, 04:56 AM

Quote:

Originally posted by g.l.amour
whoaah, xplosive action there.

i still cant figure out too well how you are doing the exit channels of the water. care to give us some build drawings or something.

how far can the jet exits be from the beginning of the baseplate. atm i haev a copper stock with bars of 1cm or 1/3" thickness. if i go for a BP thickness of 3mm, then there is 7mm drilled away. can the jet effect reach over that 7mm void, or will effectiveness be lost.

thx

"reading" the image from left right is bottom layer to top layer.

the 4 pipes are 2 intakes and 2 exhausts, each the pairs diagonally oppose each other, the intake pipes only penetrate through the very top layer and "empty" into the 6mm thick cavity, feeding the 9 jet holes, the exhaust pipes (lol almost typed port, old starwars nerd) penetrate all the way down through the jet layer, and create an escape route for the water after the jet's have done their thing.

My jet's are currently traversing 3mm+3.5mm=6.5mm so I can't see an extra .5mm being an issue. You can actually use touch to roughly gauge how "powerfull" the jet's are v distance while underwater by directing the jet's at your finger or forearm (also under water)

I'm going to expand this current design to the pattern that big ben posted earlier (4 extra holes) but I probably won't be able to see the difference that it will make.

BaThMaN · 03-31-2003, 10:43 AM

guys need some comments on this...

here's what im planning to do:

here's what im "ready" to do:

sorry for the big pics and crappy corel job... hope you get what i mean...

btw, my first posteee here =)

bigben2k · 03-31-2003, 10:49 AM

Welcome to ProCooling, BaThMaN!

As I asked at OC, what size jet holes where you thinking?

BaThMaN · 03-31-2003, 11:49 AM

hmmm, let me see... got 12x12mm area, maybe just a + or ||... or none at all... im afraid the height from the canal itself would affect the "jethole"...

it's a different situation on the first layout though... I wonder how the water will exit

what puzzles me is how many holes do I need (I assume a little bit larger than CPU core area should be occupied)

as seen from Volenti's work, he just worked out a couple of dimples

edit: just reviewed Volenti's block, I thought the hole was less than 1mm deep... argh! my eyes tricked me again!

btw, will be cooling 2 AXP 1700+ tbred Bs

theetruscan · 03-31-2003, 11:50 AM

It says 2mm jets,
in a piece of copper 1/8" thick.

03-21-2003, 10:22 AM	#27
Gilhooley Cooling Neophyte Join Date: Sep 2002 Location: Sweden Posts: 16	hmm bigben2k: Yes that would also work. However, there is much more work with your top. Designing a leak free separator that lies firm against your jet holes is a daunting task

03-21-2003, 10:36 AM	#29
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Actually, the seperator wouldn't be up on top of the block, I would opt to have it up against the opposite side of the case (unlike the monstrosity I built for Radius). I'm going to refer you to Morphling1's excellent drawing here. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-21-2003, 11:11 AM	#30
Gilhooley Cooling Neophyte Join Date: Sep 2002 Location: Sweden Posts: 16	Harder You could enlarge the area around the die spot in block 2. Then the separator would have few if any advantages. Pertaining higher flow rate that is. (ooooh NYPD Blue )

03-21-2003, 11:41 AM	#31
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	LOL@Bob! Didn't I include you in on my mind-reading machine? McMaster p/n for copper tubing: 1/16: 7190K51 3/32: 7190K52 I should make a nice smooth opening for those tubes, to promote flow. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-21-2003, 08:35 PM	#33
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	No, by all means, go right ahead! But why are you dropping the honeycomb? Convenience of calcs? Also, you might have missed it, but I mentionned using a square ended endmill, as a drill bit, to dodge the cone shape. Would your calculations differ significantly? __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-21-2003, 07:33 PM	#32
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Been thinking a lot on this, and drawing up some plans/schematics/rules-of-thumb/trigonometry. Can be done using a square grid array. Pick a drill bit size to do the dimples with. Let's choose 3/32" (~2.4mm) Sides on elements of the grid should by 5/6 (exactly) of the drill bit size (or 5/64" for our 3/32" drill bit size) This creates rough diamond shaped pins that are about 0.25 of the grid distance on a side (in this example around 5/256" or about 0.5mm on a side). These pins are going to be important as structural supports for the base-plate. Plate thickness should be twice as thick as the drill bit size (3/16" in this case). Drill depth should be 1.5x of the drill bit size to the bottom of the drill holes. This creates the dimples, the pins, and enough channel area between the pins for the water to flow out. A dimpled area of 5/8" x 5/8" will be adequate for most any CPU. Need to create side outlet channels to enable middle in, dual-side outlet config. Inlet jet plate will have jet holes to match the dimples. Jet holes should be 1/3 of the drill bit size in width (1/32") For a 3/32" dill bit grid array over a 5/8" x 5/8" area this creates basically a 8 x 8 grid. Total jet orifice area is around 32mm^2 which is restrictive enough to create a very good jet velocity with something like an Eheim 1250 pump. Ben, got any objections if I get my machinists to make up the above on a CNC mill?

03-21-2003, 08:38 PM	#34
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Dropping honeycomb because I'm thinking of rotating the grid by 45 degrees. I'm actually thinking that the coned shape for the dimples is a good thing.

03-21-2003, 09:27 PM	#35
SysCrusher Cooling Savant Join Date: Dec 2002 Location: Florida Posts: 256	The cone shape of the dimples is what you need. The impingement presses against the side of the angle in the dimple as it flows up that cone shape promoting an even thinner boundary layer. The reason I did my baseplate with convex.

03-22-2003, 08:17 AM	#36
Volenti Cooling Savant Join Date: Jan 2002 Location: in a nice cool spot Posts: 427	Ok I have a quick prototype ready to go; BP is 4.5mm thick and the holes (bit too deep to be called dimples) are 3.5mm deep and 3mm in dia, the jet's are 1.5mm in dia. Some of you may recognise the top part of the block, since it's cut down from my p4 direct die watercooling block, it has multiple layers, with diagonally opposing dual in and out. I'll hook it up to my t-bred 2100+@2.4gig tomorrow. __________________ feel free to icq/msn me, I'm always willing to toss around ideas.

03-22-2003, 07:00 PM	#38
LiquidRulez Cooling Savant Join Date: Sep 2002 Location: In Hell Posts: 322	BEN, While I like the simplicity of this idea, I think you should NOT give up on Radius(or are you just getting bored?). I think Radius has alot of potential and will be simply amazing in performance once you complete it....Not to mention, beautiful too with an LED lighting it, sitting on top of an AMD chip!

03-22-2003, 09:40 PM	#39
Volenti Cooling Savant Join Date: Jan 2002 Location: in a nice cool spot Posts: 427	Ok I've got the impingement block set up; I must say I'm quite impressed, it matches my micropin block in performance using the latest verson of MBM to record of the XP's internal diode, allthough the absolute reading is off (as usual), relitive readings seem to be accurate enough for my purposes here. __________________ feel free to icq/msn me, I'm always willing to toss around ideas.

03-23-2003, 04:15 AM	#40
leejsmith Cooling Savant Join Date: Mar 2003 Location: notts uk Posts: 408	This is an idea i had for a water delivery system in the wb. cut channels into a star and then if possible make an outer circle and an inner. Then drill the holes into the channles. If made correct you should be able to direct the jets at the cones in the base plate. then just place it under the inlet barb as a middle layer. if a little more effort and thought was put into it i am sure it will work.

03-24-2003, 12:35 PM	#43
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	volenti, Do you by chance have any pictuers of your jet in action? without the base attached to see the spray? Jon

03-25-2003, 03:59 AM	#45
g.l.amour Cooling Savant Join Date: Mar 2002 Location: on da case Posts: 933	whoaah, xplosive action there. i still cant figure out too well how you are doing the exit channels of the water. care to give us some build drawings or something. how far can the jet exits be from the beginning of the baseplate. atm i haev a copper stock with bars of 1cm or 1/3" thickness. if i go for a BP thickness of 3mm, then there is 7mm drilled away. can the jet effect reach over that 7mm void, or will effectiveness be lost. thx __________________ yo soy un tiburón

03-31-2003, 10:43 AM	#47
BaThMaN Cooling Neophyte Join Date: Jan 2003 Location: Philippines Posts: 18	guys need some comments on this... here's what im planning to do: here's what im "ready" to do: sorry for the big pics and crappy corel job... hope you get what i mean... btw, my first posteee here =)

03-31-2003, 10:49 AM	#48
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Welcome to ProCooling, BaThMaN! As I asked at OC, what size jet holes where you thinking? __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

03-31-2003, 11:49 AM	#49
BaThMaN Cooling Neophyte Join Date: Jan 2003 Location: Philippines Posts: 18	hmmm, let me see... got 12x12mm area, maybe just a + or \|\|... or none at all... im afraid the height from the canal itself would affect the "jethole"... it's a different situation on the first layout though... I wonder how the water will exit what puzzles me is how many holes do I need (I assume a little bit larger than CPU core area should be occupied) as seen from Volenti's work, he just worked out a couple of dimples edit: just reviewed Volenti's block, I thought the hole was less than 1mm deep... argh! my eyes tricked me again! btw, will be cooling 2 AXP 1700+ tbred Bs Last edited by BaThMaN; 03-31-2003 at 11:54 AM.

03-31-2003, 11:50 AM	#50
theetruscan Cooling Neophyte Join Date: Jan 2003 Location: OR/CA/NY Posts: 81	It says 2mm jets, in a piece of copper 1/8" thick.

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