just to get some responses....

Albigger · 11-05-2002, 10:08 AM

here's my creation. All criticisms/suggestions definately welcome, as I am currently redesigning and want to make a second version out of copper...

My current block:

specs:
material: aluminum
base thickness: 1/4" (between bottom of water channel and cpu die)
inside surface finish: sandblasted
base finish: flycut milled (soon to be lapped)
radiator: 6" x 7 3/4" x 2" heater core, 5/8" ins and outs
pump: 700gph danner mag drive

Performance:
Duron 750 @ 1020 MHz, 1.98 vcore
Ambient temp ~26C, 34C (respectively)
Cpu load temp (in socket) ~ 33C, 40C (respectively)
flow rate: ~ 6.5 gpm through block only (guessing ~ 4gpm with components - didn't test though)

relative size:

base:

'hinged' open:

finished:

testing:

I need to get another XP chip to test this on, and I would like to mill down the base a little and lap it as well.

gmat · 11-05-2002, 10:42 AM

Wow there's a lot of work here, kudos. That base is awesome.
The o-ring is really a nice touch for a "hobbyist block" !

How did you get the central shape on the base (slanted 'columns') ??
Ah and the texture looks "grainy".. How come ?

Albigger · 11-05-2002, 03:05 PM

it looks grainy (the base, inside) because I sandblasted it - both to rough up the surface and to try to get rid of minor burrs after i was done milliing.

I made the 'half-bowl' shape in the center by plunging a 5/8 or 9/16 (don't remember) ball mill into the center, to half its depth, BEFOR i milled all the individual pillars.

gmat · 11-05-2002, 05:54 PM

Nice. I like that. It follows the 3D shape of heat propagation around the core i saw in various CFD modellings.
Those pillars will be a challenge with a copper block ! Good luck it's awesome.

bigben2k · 11-05-2002, 06:12 PM

Quote:

Originally posted by gmat
Nice. I like that. It follows the 3D shape of heat propagation around the core i saw in various CFD modellings.

Except for one detail: the fins aren't centered over the core, they're centered over the socket.

Albigger · 11-05-2002, 06:26 PM

Quote:

Originally posted by bigben2k
Except for one detail: the fins aren't centered over the core, they're centered over the socket.

well, they are pretty close to being centered over the core actually about + or - .200 also depending on what (AMD) core you are talking about

but I was looking at this in my revision...

how much do you think it hurts cooling ability??

Fixittt · 11-05-2002, 08:40 PM

Im impressed, was that done on a manual? It looks like it.

Very nice work indeed.

Alot of drilling and tapping for all those bolts!!!!

What is the base thickness? It looks a bit thick. The sand blasting is a nice touch.

Albigger · 11-05-2002, 09:17 PM

Yes it was all done on a manual hand-crank Bridgeport machine.

1/8" pillars with 1/8" channels in between. The base is definately (in my and many others' opinions) a bit thick, at .250" right now. I'd like to mill it down to around at least to .125" or thinner, though I might just wait till I make the second version to make this change.

and i probably (almost definately) didn't need that many bolts, but i didn't feel like having it leak and aluminum is easy enough to tap.

thanks for your comments (the creator of the ever-so-love spiral block

)

Have people sandblasted the inside of other blocks? does it help performance?

Fixittt · 11-05-2002, 10:05 PM

yeah the base is to thick. (For me anyways)

Sandblasting doesnt hurt any at all, and yes if pondered about, it can be explained why it would help performace. By roughing up the bottom, you have very effectivly increased surface area. Now a good harsh media blasting would prolly double the amount of surface area on any given block.

The machining is very nice, its a good looking part if nothing else.

I could sit here and count the different operations, and prolly follow step by step how you machined it. Very nice indeed.

One thing I would have done, just to make it different, was when it was time to machine the pillars, I would have made a set of soft jaws for the vise and rotated the block. Instead of semetiracal (Sorry spelling sucks) squares, then would look more like diamonds. Basicly they would be turned sideways.

Good luck on the copper. Use coolant!

MadDogMe · 11-06-2002, 02:57 AM

If you're prepared to put that much work into a block you'll get there in Cu...
Bit paranoid on the top retention were'nt you though?

...
PS. it was mentioned in the Lemon block thread by Brad, how abouit staggering the pins so the water does'nt take a straight path to the outlets?...
I don't think the temps tally at all. 26C and 33C load temp(7C rise), but 36C only giving a 4C rise!! (40C). it should be 10C at least!...

bigben2k · 11-06-2002, 08:37 AM

Quote:

Originally posted by Albigger
well, they are pretty close to being centered over the core actually about + or - .200 also depending on what (AMD) core you are talking about

but I was looking at this in my revision...

how much do you think it hurts cooling ability??

No it probably won't, simply because a fin that high would only be a benefit at low flow rates, i.e. < 1/2 gpm, at which point the performance would suck, regardless.

Nice work though. I've been contemplating sandblasting, and I have to look into "bead blasting", just because I don't know what it is, precisely. The sandblasted surface is probably more flow restrictive, with a Hazen-Williams factor of 80 or less, where bead-blasting, in my rough estimate, might give a factor of 100.

See page 2 of this PDF click me!

Albigger · 11-06-2002, 11:29 AM

Quote:

Originally posted by Fixittt
yeah the base is to thick. (For me anyways)

yeah, definately (at least from everything i've been reading, and others' real world temps with thinner bases seem to be better) not considering peltier use of course.

Quote:

Originally posted by Fixittt
One thing I would have done, just to make it different, was when it was time to machine the pillars, I would have made a set of soft jaws for the vise and rotated the block. Instead of semetiracal (Sorry spelling sucks) squares, then would look more like diamonds. Basicly they would be turned sideways.

Good luck on the copper. Use coolant!

I will definately have to use coolant. Thanks.

This would be interesting - about having the pillars oriented 45 degrees from where they are now? correct? or maybe you wouldn't have to do squares at all, maybe that is what you are saying, a 'true' diamond or rhombus shaped pillar. It would be interesting to compare side by side the results, all else equal.

Quote:

Originally posted by MadDogMe
If you're prepared to put that much work into a block you'll get there in Cu...

good to hear

Quote:

Originally posted by MadDogMe
Bit paranoid on the top retention were'nt you though? ...

of course.....

couldn't have any leaks now could i?

Quote:

Originally posted by MadDogMe
PS. it was mentioned in the Lemon block thread by Brad, how abouit staggering the pins so the water does'nt take a straight path to the outlets?...

that would be A LOT of machining work, i think. Way more than I am prepared to give at this point. But if flowrate isn't compromised too much with that method it would definately provide more hard water to block contact. Has ANYONE tried this yet???

Quote:

Originally posted by MadDogMe
I don't think the temps tally at all. 26C and 33C load temp(7C rise), but 36C only giving a 4C rise!! (40C). it should be 10C at least!...

I quoted 34C to 40C, a 6C rise which is pretty consistent with the 7C rise at lower ambient temps. I consider my temps very stable/predictable since I've been using this block, and I would think the extra 1C increase at lower ambient temps is due to my 'testing' method - open window and put a box fan blowing at my case. But my radiator is in the very top so it was likely drawing in warmer air than the mobo monitor was reading, so that could be it...

and why do you say that it should be 10C at least??? I'm not understanding this? (don't forget I'm running a crappy duron and Not an XP chip currently, they don't put out as many watts...)

at idle my cpu temps are at or below my system temp. and i kicked down the voltage and overclock the other day to test temps:
Duron 750 @ 800 1.68vcore
system: 26C cpu at load: 29C

Man I can't wait to get another chip so I know how well this block 'actually' performs. It seems good for me now though. where can I find wattage charts for a duron @ 1gig @ ~2.0vcore, and say an xp1600 @ ~ 1800 @ 2.0vcore? If I knew that I could predict temps possibly? or not?

All that being said - I'm not sure how much I can trust the mobo temp sensors.
#1 - in socket and not on die
#2 - temps seem unusually stable now that I went watercooling, and I can't decide if that's a result of going H20 or if the sensors are messed somehow.

- but i do plan to get a rad shack temp sensor so i can independently monitor ambient temps, water temps...

Quote:

Originally posted by bigben2k
No it probably won't, simply because a fin that high would only be a benefit at low flow rates, i.e. < 1/2 gpm, at which point the performance would suck, regardless.

yeah, my block was designed to be a (very) high flow block, so I wouldn't hope to be anywhere near 1/2 gpm, at which perf. would suck as you say. Do you think even when going to a thinner base thickness that not being directly centered will hurt more? Enough to be noticeable?

Quote:

Originally posted by bigben2k
Nice work though. I've been contemplating sandblasting, and I have to look into "bead blasting", just because I don't know what it is, precisely. The sandblasted surface is probably more flow restrictive, with a Hazen-Williams factor of 80 or less, where bead-blasting, in my rough estimate, might give a factor of 100.

See page 2 of this PDF click me! [/b]

interesting PDF there, thanks for the link. In my revision I would like to use a small ball mill to create many dimples in the bottom of the block. This would increase surface area around 57% as opposed to a 'flat' milled base.

I will also have to look into media blasting. what do most here use for roughing up the inside of the block? Anything? I could swipe some 80 or 60 grit paper in between the pillars, but I wasn't sure if this would be better/worse than sand blasting....

bigben2k · 11-06-2002, 11:53 AM

Quote:

Originally posted by Albigger

yeah, my block was designed to be a (very) high flow block, so I wouldn't hope to be anywhere near 1/2 gpm, at which perf. would suck as you say. Do you think even when going to a thinner base thickness that not being directly centered will hurt more? Enough to be noticeable?

interesting PDF there, thanks for the link. In my revision I would like to use a small ball mill to create many dimples in the bottom of the block. This would increase surface area around 57% as opposed to a 'flat' milled base.

I will also have to look into media blasting. what do most here use for roughing up the inside of the block? Anything? I could swipe some 80 or 60 grit paper in between the pillars, but I wasn't sure if this would be better/worse than sand blasting....

You're welcome.

For a higher flow, you'd be much better off with a thinner baseplate. At .125 (1/8 inch, aka 3.2 mm), that's good. In my design, I'm going for 2 mm, but I may drop it to 1, for a flow rate that will be in the order of 4 to 6 gpm (240 to 360 gph).

The effect of the off-center would be more noticeable, but again, because of the fin height, you probably won't see any. It would perform better if you dropped the fin height, because it would increase the flow speed.

I think most people use whatever is available. Sandblasting is pretty common. I 've thought about using an acid, but the surface would be way too irregular, and the block would probably perform much worse. Sandpaper is good, but it's awfully inconvenient.

Albigger · 11-06-2002, 12:13 PM

yeah, good points about surface finish, and i'll probably stick with sandblasting because it's available to me free and it works and not too much of a pain.

my new design incorporates lower fin heights (only two levels, one 1/4" and next 1/2", and I may even lower those). (for reference the current block is 1/4, 1/2, and 3/4 inch long fins)

I was thinking of going around .090 in. base for my next block, (copper) which is about 2.3mm. But then I want to put dimples in the base (basically half of a sphere), effectively dropping portions of the base thickness down to about .058 in. or 1.5 mm. I think this would be MUCH better than what I have now, as 1/4 is much too thick. I also expect higher flow rates with my next revision, hopefully something like 5gpm (300 gph) through ALL components when complete.

I will post some design pics in a few days maybe when I have more time to get all your help and thoughts on it...

bigben2k · 11-06-2002, 12:43 PM

Actually, I might need YOUR help!

Dropping the baseplate thickness lets a structural integrity issue creep up. In other words, the block might collapse?!?

I'll have to give my mechanical engineer cousin a buzz...

Blackeagle · 11-06-2002, 03:09 PM

Nice job ! You clearly spent a good deal of time with the building of your block and it shows. Best of luck when you go wth a copper version.

What is the max head of your Danner? Model # ?

Albigger · 11-06-2002, 03:21 PM

for a hold down i was thinking plexi top that would apply pressure only in the center.
EDIT: i was in a hurry before, might not have been clear what I meant. i mean a thick plexi plate with holes for four mounting screws and clearand holes for the barbs, and some mechanism that pushes on the wb in the middle (over the cpu) so as not to bend the base. I will also be using a shim with my cpu, so as to help prevent this possibly.

about the pumps - look here
http://www.aquatictech.com/pumps.html

max height for model 7 is 13 feet. 120 gph at 10 feet

at anywhere from zero to two feet ( which is about where i'm at) it quotes 550 to 700 gph.

jaydee · 11-09-2002, 07:00 PM

Quote:

Originally posted by MadDogMe

PS. it was mentioned in the Lemon block thread by Brad, how abouit staggering the pins so the water does'nt take a straight path to the outlets?...

The Lemon Block is mine as most of you know and I been trying to come up with a staggered solutionI found one workable one but that would mean spreading the pins farther apart. At 1/8" already they are to far apart IMO as I cannot afford the price a a smaller endmill at the moment. I thin it could be done with a 1/16" and ALOT more patients than I have. I think it would be easier trying to make Bens radial concept block.

Albigger · 11-15-2002, 04:16 PM

Got my xp 1600 installed about 2 hours ago guys!!!

1600 AGOIA 0226 "Z"

clocks to about 1780 stable so far, not much past that....

not bad, but i still wish i had the AGOIA "Y" that did 1840 on air...

anyhow, didn' t change a thing from my duron setup, just put this processor in. Maybe tightened down the bolts harder (I tightened them pretty darn good...)

Here's a screenie for ya:

That's 1.95 vcore (reported in MBM) ----- 1.98 vcore (reported in BIOS)

at 10.5 x 169 = 1775 (1777 reported)

Man these temps are making me want to get my copper block done even faster!!!

Though i know my air to the rad is at least a couple degrees cooler than my system temp is, i just have no way to stabilize temps in here that well.....

11-05-2002, 08:40 PM	#7
Fixittt CNC Beyatch Join Date: Aug 2001 Location: Tulsa Spell it backwards Posts: 721	Im impressed, was that done on a manual? It looks like it. Very nice work indeed. Alot of drilling and tapping for all those bolts!!!! What is the base thickness? It looks a bit thick. The sand blasting is a nice touch. __________________ Creator of the Spir@l Block Longest post ever http://forums.procooling.com/vbb/showthread.php?s=&postid=43808#post43808

11-05-2002, 10:05 PM	#9
Fixittt CNC Beyatch Join Date: Aug 2001 Location: Tulsa Spell it backwards Posts: 721	yeah the base is to thick. (For me anyways) Sandblasting doesnt hurt any at all, and yes if pondered about, it can be explained why it would help performace. By roughing up the bottom, you have very effectivly increased surface area. Now a good harsh media blasting would prolly double the amount of surface area on any given block. The machining is very nice, its a good looking part if nothing else. I could sit here and count the different operations, and prolly follow step by step how you machined it. Very nice indeed. One thing I would have done, just to make it different, was when it was time to machine the pillars, I would have made a set of soft jaws for the vise and rotated the block. Instead of semetiracal (Sorry spelling sucks) squares, then would look more like diamonds. Basicly they would be turned sideways. Good luck on the copper. Use coolant! __________________ Creator of the Spir@l Block Longest post ever http://forums.procooling.com/vbb/showthread.php?s=&postid=43808#post43808

11-06-2002, 02:57 AM	#10
MadDogMe Thermophile Join Date: Aug 2002 Location: Just shut up ;) ... Posts: 1,068	If you're prepared to put that much work into a block you'll get there in Cu... Bit paranoid on the top retention were'nt you though? ... PS. it was mentioned in the Lemon block thread by Brad, how abouit staggering the pins so the water does'nt take a straight path to the outlets?... I don't think the temps tally at all. 26C and 33C load temp(7C rise), but 36C only giving a 4C rise!! (40C). it should be 10C at least!... Last edited by MadDogMe; 11-06-2002 at 03:07 AM.

11-06-2002, 12:43 PM	#15
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Actually, I might need YOUR help! Dropping the baseplate thickness lets a structural integrity issue creep up. In other words, the block might collapse?!? I'll have to give my mechanical engineer cousin a buzz... __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

11-06-2002, 03:21 PM	#17
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	for a hold down i was thinking plexi top that would apply pressure only in the center. EDIT: i was in a hurry before, might not have been clear what I meant. i mean a thick plexi plate with holes for four mounting screws and clearand holes for the barbs, and some mechanism that pushes on the wb in the middle (over the cpu) so as not to bend the base. I will also be using a shim with my cpu, so as to help prevent this possibly. about the pumps - look here http://www.aquatictech.com/pumps.html max height for model 7 is 13 feet. 120 gph at 10 feet at anywhere from zero to two feet ( which is about where i'm at) it quotes 550 to 700 gph. Last edited by Albigger; 11-07-2002 at 01:49 AM.

11-05-2002, 10:08 AM	#1
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	just to get some responses.... here's my creation. All criticisms/suggestions definately welcome, as I am currently redesigning and want to make a second version out of copper... My current block: specs: material: aluminum base thickness: 1/4" (between bottom of water channel and cpu die) inside surface finish: sandblasted base finish: flycut milled (soon to be lapped) radiator: 6" x 7 3/4" x 2" heater core, 5/8" ins and outs pump: 700gph danner mag drive Performance: Duron 750 @ 1020 MHz, 1.98 vcore Ambient temp ~26C, 34C (respectively) Cpu load temp (in socket) ~ 33C, 40C (respectively) flow rate: ~ 6.5 gpm through block only (guessing ~ 4gpm with components - didn't test though) relative size: base: 'hinged' open: finished: testing: I need to get another XP chip to test this on, and I would like to mill down the base a little and lap it as well.

11-05-2002, 10:42 AM	#2
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Wow there's a lot of work here, kudos. That base is awesome. The o-ring is really a nice touch for a "hobbyist block" ! How did you get the central shape on the base (slanted 'columns') ?? Ah and the texture looks "grainy".. How come ?

11-05-2002, 03:05 PM	#3
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	it looks grainy (the base, inside) because I sandblasted it - both to rough up the surface and to try to get rid of minor burrs after i was done milliing. I made the 'half-bowl' shape in the center by plunging a 5/8 or 9/16 (don't remember) ball mill into the center, to half its depth, BEFOR i milled all the individual pillars.

11-05-2002, 05:54 PM	#4
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Nice. I like that. It follows the 3D shape of heat propagation around the core i saw in various CFD modellings. Those pillars will be a challenge with a copper block ! Good luck it's awesome.

11-05-2002, 09:17 PM	#8
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	Yes it was all done on a manual hand-crank Bridgeport machine. 1/8" pillars with 1/8" channels in between. The base is definately (in my and many others' opinions) a bit thick, at .250" right now. I'd like to mill it down to around at least to .125" or thinner, though I might just wait till I make the second version to make this change. and i probably (almost definately) didn't need that many bolts, but i didn't feel like having it leak and aluminum is easy enough to tap. thanks for your comments (the creator of the ever-so-love spiral block ) Have people sandblasted the inside of other blocks? does it help performance?

11-06-2002, 12:13 PM	#14
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	yeah, good points about surface finish, and i'll probably stick with sandblasting because it's available to me free and it works and not too much of a pain. my new design incorporates lower fin heights (only two levels, one 1/4" and next 1/2", and I may even lower those). (for reference the current block is 1/4, 1/2, and 3/4 inch long fins) I was thinking of going around .090 in. base for my next block, (copper) which is about 2.3mm. But then I want to put dimples in the base (basically half of a sphere), effectively dropping portions of the base thickness down to about .058 in. or 1.5 mm. I think this would be MUCH better than what I have now, as 1/4 is much too thick. I also expect higher flow rates with my next revision, hopefully something like 5gpm (300 gph) through ALL components when complete. I will post some design pics in a few days maybe when I have more time to get all your help and thoughts on it...

11-06-2002, 03:09 PM	#16
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	Nice job ! You clearly spent a good deal of time with the building of your block and it shows. Best of luck when you go wth a copper version. What is the max head of your Danner? Model # ?

11-15-2002, 04:16 PM	#19
Albigger Cooling Savant Join Date: Sep 2002 Location: ohio Posts: 140	Got my xp 1600 installed about 2 hours ago guys!!! 1600 AGOIA 0226 "Z" clocks to about 1780 stable so far, not much past that.... not bad, but i still wish i had the AGOIA "Y" that did 1840 on air... anyhow, didn' t change a thing from my duron setup, just put this processor in. Maybe tightened down the bolts harder (I tightened them pretty darn good...) Here's a screenie for ya: That's 1.95 vcore (reported in MBM) ----- 1.98 vcore (reported in BIOS) at 10.5 x 169 = 1775 (1777 reported) Man these temps are making me want to get my copper block done even faster!!! Though i know my air to the rad is at least a couple degrees cooler than my system temp is, i just have no way to stabilize temps in here that well.....

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