Overclock.net › Forums › Cooling › Water Cooling › BlockHeads: The Thread
New Posts  All Forums:Forum Nav:

BlockHeads: The Thread - Page 62

post #611 of 753
Hmm Herc, I think you shocked everyone into silence! That's a pretty wild design. Good luck!

I'm getting ready to start my first machine shop class, and am thinking of what blocks I can make with what manual mills they have, and would like your input on design.

I figure I'll make a base, and a top, hopefully out of acrylic. Just 2 barbs, with a straight path inbetween in and out. I think the easiest might be straight rows milled into the base, leaving parallel fins along the in/out path. (whitewater like) I'm unclear on what's the best design for the pins/fins (I've seen straight fins, really mini fins, diamond cut pins, dimples, serpentine pathways.)

What design do you think i should attempt for my first block? I want to make a GPU block or NB block first I think. (NB is prob simpler!)

Since alum is cheaper than copper, it will probably be made of alum, at least as a prototype.

I've just gotten the 90 day solidworks, so I'm closer to making my own design and being able to show you images, but I'm not there yet!
Edited by ModMinded - 7/29/09 at 1:19am
Main Rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
QX9650 EVGA 750i FTW EVGA GTX 275 FTW 2 x 2GB OCZ Reaper PC2 8500 
Hard DriveOSMonitorKeyboard
500 GB Maxtor XP 32/Vista64/Win7 beta64 19" Viewsonic logitech basic 
PowerCaseMouseMouse Pad
Zalman ZM-1000HP ThermalTake Element S MX Laser desk 
  hide details  
Reply
Main Rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
QX9650 EVGA 750i FTW EVGA GTX 275 FTW 2 x 2GB OCZ Reaper PC2 8500 
Hard DriveOSMonitorKeyboard
500 GB Maxtor XP 32/Vista64/Win7 beta64 19" Viewsonic logitech basic 
PowerCaseMouseMouse Pad
Zalman ZM-1000HP ThermalTake Element S MX Laser desk 
  hide details  
Reply
post #612 of 753
yeah, aluminum is great for prototyping and short tests. you shouldnt need more than an hour maybe 2 to test each block... corrosion isnt much of a problem in such a short timespan. once you have decided the best design, you remake it in copper.
post #613 of 753
Quote:
I like redundancy and always planned to go two pumps. I think lots of people do the same as well. However, I think I will go two complete loops into one cpu block. MAYBE something like this
Something like that would probably make a better TEC block then just a WC'ing block. It would be restrictive but if it's fed off 2 18W DDC that shouldn't be a problem. The only thing I might change is having so many mid-plates, simpler single mid-plate might work better, plus less chance of leaks between the layers.

I've wanted a dual in block fed from 2 pumps for a while now, hey if nothing else it would be different...
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH 
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620 
Case
Open Bench 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
E8600 4545Mhz*9 w/1.29V DFI DK P45 XFX 8800 GS D9GKX & D9GMH 
Hard DriveOptical DriveOSPower
2x Seagate 160 Perp. RD-1 2x Asus Vista Premium Corsair HX 620 
Case
Open Bench 
  hide details  
Reply
post #614 of 753
Wow, I just spent 2 hours reading this thread and drooling...especially the brass and copper blocks. I'd like to make a GPU full cover block either full copper or copper base and brass top.

Can anyone tell me what's important when designing a GPU block? I believe low restriction, high flow is more important. Ek blocks seems to support this idea looking at their milling patterns.

Also what's the best way to seal the block? I've seen solder, gasket, and o-rings used but not sure how to size the o-rings. Thanks!
post #615 of 753
Quote:
Originally Posted by Retro View Post
Wow, I just spent 2 hours reading this thread and drooling...especially the brass and copper blocks. I'd like to make a GPU full cover block either full copper or copper base and brass top.

Can anyone tell me what's important when designing a GPU block? I believe low restriction, high flow is more important. Ek blocks seems to support this idea looking at their milling patterns.

Also what's the best way to seal the block? I've seen solder, gasket, and o-rings used but not sure how to size the o-rings. Thanks!
EK? not with the supreme in any case...

you'll need to design it with highest possible flow (low restriction) yet, highest possible turbulence. (so you'll be trading one for the other, but different types of designs have a more of either than others).

then you also gotta keep in mind WHERE you want that turbulence, and where you could do with less.
post #616 of 753
Thank you for the reply.

If I understand correctly, the highest turbulence should be across the GPU itself. This is why the EK-FC260 has the fins across the GPU area and a unrestricted flow across the other components to be cooled. The fins increase the water pressure and turbulence. Am I even close?

This was the block I was looking at when I was saying the EK blocks were lower resistance and higher flow.

As found herehttp://ekwaterblocks.com/shop/produc...b71bbe8a5cb0af
post #617 of 753
Quote:
Originally Posted by Retro View Post
Thank you for the reply.

If I understand correctly, the highest turbulence should be across the GPU itself. This is why the EK-FC260 has the fins across the GPU area and a unrestricted flow across the other components to be cooled. The fins increase the water pressure and turbulence. Am I even close?

This was the block I was looking at when I was saying the EK blocks were lower resistance and higher flow.

As found herehttp://ekwaterblocks.com/shop/produc...b71bbe8a5cb0af
yeah, pretty close. would like to add to that, the increased surface area also helps. now, EK's design is decent, but not nearly the most efficient... as you can see the channels are slim, if they were widened, they could instead be made to give even more turbulence instead without losing flow. pins for example, instead of channels would most likely do that... kinda like Swiftechs GT(X) and MCW60.
post #618 of 753
Quote:
Originally Posted by ira-k View Post
Something like that would probably make a better TEC block then just a WC'ing block. It would be restrictive but if it's fed off 2 18W DDC that shouldn't be a problem. The only thing I might change is having so many mid-plates, simpler single mid-plate might work better, plus less chance of leaks between the layers.

I've wanted a dual in block fed from 2 pumps for a while now, hey if nothing else it would be different...
That's not just two loops going into the block, the block itself also takes the loops and keeps it seperate. Not sure I can make a single mid-plate do the job of both the upper/lower 2 mid-plates in the pic. However, I doubt I will do it that way, will probably just have the loops become one within the CPU block, which will make it simpler. Also, I am building up, I want a really thick block. Thought I was gonna have a bunch of .5" thick plates to get the thickness I want, but got this 2" thick copper plate off ebay for $50!!

It's 2" thick x 7" x 6". Thats a .375" thick copper bar next to it


Not sure what grade copper it is, diffinently not 110, maybe 182, seems harder and also a lighter color...under the tarnish. Anyway, the block will be built up to be wide enough and tall enough to put a 62mm TEC on all four sides. The biggest problem is if I actually have enough space for the hotside TEC waterblocks that will be built around the stack, but I haven't gotten that far in planning yet.


Quote:
Also what's the best way to seal the block? I've seen solder, gasket, and o-rings used but not sure how to size the o-rings. Thanks
I'd go with o-rings, provided I can actually mill the grooves for it to work. I think it's very reliable and reusable and it allows the two surfaces to contact each other. Been looking into getting a spool of really soft copper wire and using it as a crushable one time use metalic o-ring.

Fullcover GPU block? Good luck with that, seems very complicated to me. Two or three, maybe more?? different heightts to deal with and how many screw holes that needs to be lined up exactly? I just went through trying to mill one end of my itx mobos block height down so that it will sit level on my CPU and NB....and that was painfull enough. I'd say try a CPU block first, lot less complexities to worry about. I am regretting trying to make a CPU/NB block as my first ever block, first time ever using a manual mill. Have already made @$20 worth of scrap copper plate lol. Probably woulda been easier to cut a 2" square, a screw at each corner, do the gasket thing, a simple fin patterm and then tap a couple fitting holes. Lot less to go wrong. Instead, I did a weird shape, two surfaces to cool at different heights, a couple screw holes that need to line up exactly...but I don't think that is anywahere as complex to mill as a fullcover gpu block...with a manual mill anyway.


Copper: I started out buying through Onlinemetals, then Mcmsters-carr. Mcmaster's is less expensive usually, but I didn't like that they didn't give me the shipping cost....which can be very bad since I live in Hawaii. However, I read that Mcmasters is very cheap for shipping compared to any store that ships UPS, so I ordered some 4"x36" bars of copper in .1875", .5" and one of brass in .1875". Both Onlinemetal and Mcmsters use UPS, which is expensive to ship...to Hawaii. My Mcmaster order cost me $120 in shipping alone. Now I found all sorts of copper plate deals on ebay. Even bought a .75" thick x 3" x 24" bar from Onlinemetals ebay store since it was fairly cheap considering the thickness. Put in the bid for that 2" thick plate I got and they shipped it USPS, so it only cost me $10 to ship. If it had gone through UPS, it'd have cost me probably $50 to ship...that thing is heavy! Found an ebay seller with a bunch of different thicknesses and sizes for copper..that shipped all uing USPS. Bought out most of his stock of .25" and .5" 4" x XX" bars...since mot of them were around $8 for the bar and even less to ship. Didn't say what grade copper it is other then "99.9%"

Saw Ira link this ebay seller's stroe for acrylic and delrin

http://stores.shop.ebay.com/The-Grea...__W0QQ_armrsZ1


really good prices and cheap USPS shipping. Ordered some .5" thick Polycarbonate sheets from him. Was cheaper (even after shipping) then going to Home Depot and getting thinner and probably not as good stuff.
post #619 of 753
Quote:
pins for example, instead of channels would most likely do that
Something like this or are you thinking denser pins? Inlet is in the center. I added the divider to distribute the flow across the pins.



This is what I'm trying to cool.




Thanks for the info on the o-rings, Herc. On the difficulty of the GPU block, I'm trying to keep it simple as I'm new at this. I also am fortunate to have access to a machine shop with CNC machine and a laser cutter. The folks who make their living driving them made the mistake of saying that if I can draw it then they can build it.

Nice chunk of copper there. I'll be very curious to see what happens with that one. I've also had good luck with off cuts on ebay.

1/4", 1/2", 1/8", 1/4"

Edited by Retro - 8/2/09 at 7:26am
post #620 of 753
i'd twist the pins 45 degrees, so they "split" the flow. maybe more and smaller but the 45 degree thing is a little more important.. you still get the turbulence, but less restriction.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › BlockHeads: The Thread