Lemon Block Cu 2 (development thread) - Page 2

jaydee · 10-30-2003, 06:50 PM

Quote:

Originally posted by Blackeagle
How's about this, "The Edge Block"

For the edges of the pins done with care to increase surface area of the pins edges.

Also liked the idea of having the central pin tapered to a point, with dimples around the pins.

Will be very interesting to see how much of a performance increase this will provide in a design that should still provide really high flow rates due to the lower number of pins. In the limited area just above the core this design should offer a large amount of surface area for heat removal.

While you may have posted the pump you intend for use in the 3 system folding set up I can't seem to find it. What pump will power your Tri-Folder Jaydee? Or will that choice wait for a final design of the CPU blocks?

I am planning on using a VIA 2600.
Max GPH = 740
Max head = 7ft
GPH at 4ft = 310
46watts per hour

I am also considering 2 VIA 1300's. If one fails there will still be one, but also double the chance of a failure....

Blackeagle · 10-30-2003, 07:25 PM

Another nice tweak Jaydee.

A idea I suggested in your other thread would perhaps add to things.

USe a copper reducer like you used before. Only instead of going from small to larger, have it going from 1/2" to 3/8" or even 1/4" to increase the waters velocity just as it's entering the blocks pin area. With a pointed center pin this would send a fast flow of water aginst the inside edge of all the perimiter pins for really good cooling aginst those rough sided pins as well as a dimple pattern between the pins.

And a reduction in the inlet size would be desireable as well because the inflowing water will have some of the water out at the very edge of the CPU just as it comes in or even off the edge of the CPU a bit. The water at the CPU's edge and outside the die area will contribute almost nothing to cooling of the die if the full sized inlet is used.

EDIT:
And what would be the velocity increase of a 3/8" as compared to a 1/2" be? 80% higher velocity? And a 1/2" outlet for minimum resistence can still be used.

jaydee · 10-30-2003, 11:09 PM

...

jaydee · 10-31-2003, 12:23 PM

I got motivated last night and decided to bring the mill into work today being I got nothing to do here today.

This block is being milled as I type this. SLOOOOOOLY but surely. If the end mill holdes up it should be done tonight sometime... 2 hrs already and not even a 1/8" in. Going in 3/8" so hopefully it will make it by the end of the work day, milling anyway. Still don;t have a roughing endmill but i can do that anytime. Kinda want to test the block before I do that and after to see what difference (if any) it makes.

The base is 1/2" thick overall, the pins with be 3/8" high leaving 1/8" for the end base thickness. Which might be to much... I want the base to be pretty thin so the heat doesn't spread out. i want the heat to go strit up the pins. I think this is going ot work pretty good....

JFettig · 10-31-2003, 03:12 PM

oh wow jaydee, thats SLOW! whadya say when I get my CNC running we have a race? heck I would race you with my hand wheels!

That last link of that one block took less than an hour to mill those channels manually.

How many IPM and what depth are you cutting at?

Jon

winewood · 10-31-2003, 03:19 PM

The dimples really shouldn't help at all. I think I can tell ya why. There is no easy way to get water in and out into those holes. Its more of a sight trick. Do this test.. (because its easy) Mill the block without them. Then test it. Get a good base temp. The take it off, and cut the holes in it. I think you should find they are not necessary and no performace gain.
If anything they remove the copper from the hottest part and restrict the heat leaving the die area. Since they can't dissipate the heat easily it actually may hurt. IMHO. <--- (notice the imho disclaimer)
Easy way to prove me wrong there.. you should jump on it.

zoson · 10-31-2003, 03:33 PM

What's funny and ironic about all of this, is I've been describing a block just like this in the pro/chat for over a year now. I even have some hand drawn sketches. I just hope my other idea doesn't get thought of elsewhere before I get a chance to do it...
-Zoson

winewood · 10-31-2003, 03:39 PM

Well frankly most ideas are already been used and acted on. The difference in a performer and an non -performer is the tweaks and small things that people overlook. If you looked at Jaydee's Kingpin block <- (nice name) you could attempt to replicate it, but with small variables come out with way different numbers.

I would like to see your block zoson, then have someone compare and figure out why or what makes them different performance wise.

jaydee · 10-31-2003, 04:28 PM

Quote:

Originally posted by JFettig
oh wow jaydee, thats SLOW! whadya say when I get my CNC running we have a race? heck I would race you with my hand wheels!

That last link of that one block took less than an hour to mill those channels manually.

How many IPM and what depth are you cutting at?

Jon

Most of the time was burned do to not optomizing to GNC code. I could do it in maybe 2 hours with better tooling. The only 1/8" endmill I had to use was the one I made the Lemon Block Cu with. It was already worn pretty good. It made it though. Next endmills I order are going to be 3 flute carbide instead of 2 flute. I was talking to a machinest the other day and he said use 3 flute for 1/16" and 1/8" in copper because they don't clogg up and they are much better ballanced.

I was going 1/16" each pass except the first 2 passes where 1/32" . Also 3/8" deep was a little deeper than the endmill was ment to go, but it worked.

If I can this down to an hour I will start selling them!

Got to get the holes figured out now...

jaydee · 10-31-2003, 04:51 PM

Also note I only got 2.5 hrs into the milling so far. at 1/16" it cut like a hot knife through butter. If I use a 3 flute carbide and optomize the tool path I maybe able to cut the time in half.

MMZ_TimeLord · 11-01-2003, 02:10 AM

Jaydee,

Nice work... looks good. Yeah, if you optimize the tool path you will wear your endmills less too...

JFettig · 11-01-2003, 07:58 AM

Jaydee, Here is another suggestion for you. Try using a 1/4" BP leave a 1/16" base.

Jon

jaydee · 11-01-2003, 09:26 AM

Quote:

Originally posted by JFettig
Jaydee, Here is another suggestion for you. Try using a 1/4" BP leave a 1/16" base.

Jon

The 3 for this project will all be 1/2" because that's what I have left to use right now.

In the future though I will try 3/8" and use a 1/16" base. The pins need to be 3 times higher than they are wide from my experience. I don't think I can get away with 1/4" but I might try it anyway. Once these 3 blocks are done and I feel they perform well I will round up more materials and do more experimentation on different sizes.

I got to take off for the weekend. Will post back tomorrow with more pics of the block. Should have one completed by then.

Blackeagle · 11-02-2003, 04:02 PM

Quote:

Originally posted by winewood
The dimples really shouldn't help at all. I think I can tell ya why. There is no easy way to get water in and out into those holes. Its more of a sight trick. Do this test.. (because its easy) Mill the block without them. Then test it. Get a good base temp. The take it off, and cut the holes in it. I think you should find they are not necessary and no performace gain.
If anything they remove the copper from the hottest part and restrict the heat leaving the die area. Since they can't dissipate the heat easily it actually may hurt. IMHO. <--- (notice the imho disclaimer)
Easy way to prove me wrong there.. you should jump on it.

Clearly it's up to Jaydee if he wishes to do before and after testing of the dimples.

But there is no problem getting water in and out. While it's true they are made with a drill and drill press they are not intended to be holes in any sense. "Dimples" are very shallow depressions in the bases surface made with only the sharply tapered tip of the drill.

And with the dimples surrounding the central pin which is tapered to a point, the water is directed at the dimples from right above and when it hits them just splashes right out, spreading on accross the base and out between other pins.

While the dimples do increase turbulence they also increase surface area for the water stream from the inlet to impact. Both increase cooling.

The roughing mill bit will do the same for thesides of the pins, greatly increase surface area.

While I don't need proofs that the dimples and roughing mill bit will improve performance it would be nice, if Jaydee has the time, to see a before and after in temps.

JFettig · 11-02-2003, 06:29 PM

oh yeah, about the pumps, why use dual crappy pumps when you can use one good pump? I would say just get a danner mag#3 and you would be good to go. not worry about failure either. and spend the same or just a little more for much better.

Jon

jaydee · 11-02-2003, 07:29 PM

Quote:

Originally posted by JFettig
and spend the same or just a little more for much better.

Jon

I don't like Danner's for PC water cooling. To big, to hot, to expensive for what you get. I would go with several other brands first.

Also I can get 2 VIA 2600's for one danner of equel specs. Hell I can get a VIA 4900 (1321GPH 14ft. head) for less than a Danner 500GPH.

http://www.drsfostersmith.com/produc...=6&pCatId=8164

I also might just use my Hydrothruster 500. The flow through these blocks is going to be excellent. Another reason I went 3/8 deep".

I got 2 blocks complete and will post pics here whenever I get the cam working..... One plexy topped and one all copper.

jaydee · 11-02-2003, 07:30 PM

Quote:

Originally posted by Blackeagle

While I don't need proofs that the dimples and roughing mill bit will improve performance it would be nice, if Jaydee has the time, to see a before and after in temps.

I will be running this experiment in time. I am not going to do it with these blocks as I just want to get this project done and max performance isn't an issue.

In 2 weeks I will be able to get back on the mill and run some more stuff out. That will give me time to ge some 3 flute endmills and more copper, and that roughing endmill.

jaydee · 11-02-2003, 08:31 PM

Pics of the two I got done so far.

MMZ_TimeLord · 11-02-2003, 08:47 PM

NICELY DONE!!!

How do you calculate the channel for an O-Ring again? I just made my channel last time and filled it with a silicone sealant. I'd rather have a true O-Ring.

jaydee · 11-02-2003, 08:49 PM

Quote:

Originally posted by MMZ_TimeLord
NICELY DONE!!!

How do you calculate the channel for an O-Ring again? I just made my channel last time and filled it with a silicone sealant. I'd rather have a true O-Ring.

I just made the groove to fit around the channel and made my own O-ring. I got 10ft of O-Ring material for like $3. Just use super glue to attach both sides.

jaydee · 11-02-2003, 09:07 PM

Quote:

Originally posted by jaydee116
I just made the groove to fit around the channel and made my own O-ring. I got 10ft of O-Ring material for like $3. Just use super glue to attach both sides.

Or did you mean the depth? I just went 2/3" the width of the O-Ring material.

Blackeagle · 11-02-2003, 10:20 PM

Nice work Jaydee, and some really nice close up pics of the finished work to.

Will be interesting enough for now to see how this version compares to past versions.

(We really do need a larger selection of smilies, at least a thumbs up or even better a cheers smilie for really nice work)

jaydee · 11-03-2003, 11:54 AM

Quote:

Originally posted by Blackeagle
Nice work Jaydee, and some really nice close up pics of the finished work to.

Will be interesting enough for now to see how this version compares to past versions.

(We really do need a larger selection of smilies, at least a thumbs up or even better a cheers smilie for really nice work)

This weekend I will be shutting down the D.F. Farm and getting everything layed out in the Box I am building for the 3 comp project. While that is going on I will re-setup the test bench and run some tests on the Lemon Block Cu, the Maze 4, and the King Pin with plexy top and the copper top. The Copper top has a slightly larger inlet so it might make a difference. I bored the plastic barbs out to 7/16". As close as I could get to 1/2" without them lossing their strenght and collapsing.

I will be ordering some more copper and tooling tonight aswell. The weekend after this one I will be milling a couple more King Pin bllocks and something along the lines of the pic attached.

satanicoo · 11-03-2003, 05:13 PM

WOOOOOOOOOW!

Man, i love those glass-looking blocks...

JFettig · 11-03-2003, 09:51 PM

ok jaydee,
Whats bigger? 1 danner or 2 vias? (vias of course)
what runs cooler? 1 danner or 2 vias? or even 1 via...(via of course)
Whats louder? 1 danner or 2 vias?(even 1 via is)
Whats more reliable? 1 danner or 2 vias?(you get what you pay for)

Jon

10-30-2003, 07:25 PM	#27
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	Another nice tweak Jaydee. A idea I suggested in your other thread would perhaps add to things. USe a copper reducer like you used before. Only instead of going from small to larger, have it going from 1/2" to 3/8" or even 1/4" to increase the waters velocity just as it's entering the blocks pin area. With a pointed center pin this would send a fast flow of water aginst the inside edge of all the perimiter pins for really good cooling aginst those rough sided pins as well as a dimple pattern between the pins. And a reduction in the inlet size would be desireable as well because the inflowing water will have some of the water out at the very edge of the CPU just as it comes in or even off the edge of the CPU a bit. The water at the CPU's edge and outside the die area will contribute almost nothing to cooling of the die if the full sized inlet is used. EDIT: And what would be the velocity increase of a 3/8" as compared to a 1/2" be? 80% higher velocity? And a 1/2" outlet for minimum resistence can still be used. Last edited by Blackeagle; 10-30-2003 at 07:34 PM.

10-31-2003, 03:19 PM	#31
winewood Cooling Savant Join Date: Nov 2002 Location: in my chair Posts: 574	The dimples really shouldn't help at all. I think I can tell ya why. There is no easy way to get water in and out into those holes. Its more of a sight trick. Do this test.. (because its easy) Mill the block without them. Then test it. Get a good base temp. The take it off, and cut the holes in it. I think you should find they are not necessary and no performace gain. If anything they remove the copper from the hottest part and restrict the heat leaving the die area. Since they can't dissipate the heat easily it actually may hurt. IMHO. <--- (notice the imho disclaimer) Easy way to prove me wrong there.. you should jump on it. __________________ -winewood-

10-31-2003, 03:33 PM	#32
zoson Cooling Savant Join Date: Nov 2001 Location: X Posts: 204	What's funny and ironic about all of this, is I've been describing a block just like this in the pro/chat for over a year now. I even have some hand drawn sketches. I just hope my other idea doesn't get thought of elsewhere before I get a chance to do it... -Zoson __________________ Core i7 930 @ 4254Hz 22x193 \| ASUS R3E 901 BIOS \| 12GB Mushkin Blackline 3x4GB @ 1547MHz 7-8-7-24-1N \| Lian Li PC-A10B 2x EVGA GTX275 SLI @ 720/1266/1566 \| 2x 80GB Intel X25-M G2 SSD Raid 0 \| 3x 1TB Samsung Spinpoint 7.2K Raid 5 \| Corsair 950TX H2O \| Enzotech Sapphire Rev.A \| EK-FB RE3 \| Swiftech - 2x MCW60/GTX275 Unisink, MCP655b, MCRes Micro.v2 \| 3x HWLabs SR1-120 あなたの神への弓 Time isn't wasted, when you're gettin' wasted.

10-31-2003, 03:39 PM	#33
winewood Cooling Savant Join Date: Nov 2002 Location: in my chair Posts: 574	Well frankly most ideas are already been used and acted on. The difference in a performer and an non -performer is the tweaks and small things that people overlook. If you looked at Jaydee's Kingpin block <- (nice name) you could attempt to replicate it, but with small variables come out with way different numbers. I would like to see your block zoson, then have someone compare and figure out why or what makes them different performance wise. __________________ -winewood-

11-01-2003, 02:10 AM	#36
MMZ_TimeLord Cooling Savant Join Date: Apr 2003 Location: San Mateo, CA, USA, Earth Posts: 433	Jaydee, Nice work... looks good. Yeah, if you optimize the tool path you will wear your endmills less too... __________________ MMZ>TimeLord "Oooooooooh... that's gonna leave a mark!"

10-31-2003, 12:23 PM	#29
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	I got motivated last night and decided to bring the mill into work today being I got nothing to do here today. This block is being milled as I type this. SLOOOOOOLY but surely. If the end mill holdes up it should be done tonight sometime... 2 hrs already and not even a 1/8" in. Going in 3/8" so hopefully it will make it by the end of the work day, milling anyway. Still don;t have a roughing endmill but i can do that anytime. Kinda want to test the block before I do that and after to see what difference (if any) it makes. The base is 1/2" thick overall, the pins with be 3/8" high leaving 1/8" for the end base thickness. Which might be to much... I want the base to be pretty thin so the heat doesn't spread out. i want the heat to go strit up the pins. I think this is going ot work pretty good....

10-31-2003, 03:12 PM	#30
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	oh wow jaydee, thats SLOW! whadya say when I get my CNC running we have a race? heck I would race you with my hand wheels! That last link of that one block took less than an hour to mill those channels manually. How many IPM and what depth are you cutting at? Jon

10-31-2003, 04:51 PM	#35
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Also note I only got 2.5 hrs into the milling so far. at 1/16" it cut like a hot knife through butter. If I use a 3 flute carbide and optomize the tool path I maybe able to cut the time in half.

11-01-2003, 07:58 AM	#37
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	Jaydee, Here is another suggestion for you. Try using a 1/4" BP leave a 1/16" base. Jon

11-02-2003, 06:29 PM	#40
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	oh yeah, about the pumps, why use dual crappy pumps when you can use one good pump? I would say just get a danner mag#3 and you would be good to go. not worry about failure either. and spend the same or just a little more for much better. Jon

11-02-2003, 08:31 PM	#43
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Pics of the two I got done so far.

11-02-2003, 08:47 PM	#44
MMZ_TimeLord Cooling Savant Join Date: Apr 2003 Location: San Mateo, CA, USA, Earth Posts: 433	NICELY DONE!!! How do you calculate the channel for an O-Ring again? I just made my channel last time and filled it with a silicone sealant. I'd rather have a true O-Ring. __________________ MMZ>TimeLord "Oooooooooh... that's gonna leave a mark!"

11-02-2003, 10:20 PM	#47
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	Nice work Jaydee, and some really nice close up pics of the finished work to. Will be interesting enough for now to see how this version compares to past versions. (We really do need a larger selection of smilies, at least a thumbs up or even better a cheers smilie for really nice work)

11-03-2003, 05:13 PM	#49
satanicoo Cooling Savant Join Date: Nov 2002 Location: portugal Posts: 635	WOOOOOOOOOW! Man, i love those glass-looking blocks...

11-03-2003, 09:51 PM	#50
JFettig Cooling Savant Join Date: Feb 2003 Location: Willmar MN/Fargo ND Posts: 504	ok jaydee, Whats bigger? 1 danner or 2 vias? (vias of course) what runs cooler? 1 danner or 2 vias? or even 1 via...(via of course) Whats louder? 1 danner or 2 vias?(even 1 via is) Whats more reliable? 1 danner or 2 vias?(you get what you pay for) Jon

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