Block weight

WinFlex · Jan 23, 2003

I just designed my newest waterblock in solidworks. The mass calculation returned an astonishing weight of 1.3lbs as the total weight. Is this too heavy, even tho I am using the four holes around the socket to mount the block?
Also, I am unclear as to what the exact specifications are regarding those four holes... what diameter do they have and how far appart are they?

I appreciate your feedback,

Felix

bigben2k · Jan 23, 2003

That's on the heavy side. AMD specs call for no more than about 300 grams: you're at about twice that, but it shouldn't be a problem.

The hole specs can be found in the sticky, I think...

Mikeonatrike · Jan 23, 2003

I was wondering the same thing when i picked up the block of copper i made my wateblock out of, but i did a little test on an old 486 motherboard i had. And the socket held 4 pounds of copper no problem

well it was bending a little but it shows that 1 pound prolly wont hurt anything. But ya thats just me

NeoMoses · Jan 23, 2003

many High Performance Heatsinks are well over 600 grams and I haven't heard any serious problems. 1.3 lb is pretty heavy, but should be fine. Is there any of the copper that could be replaced with aluminum? Anodized aluminum tops have helped me keep the weight of my blocks very low while making them just as leak-proof as copper. (No cracking like platic tops.)

I use #6 screws for mounting though the 4 holes, which are centered on the corners of a 36x66mm rectangle. I can show you a pic if you want.

BillA · Jan 23, 2003

a heavy wb, plus the weight of the hoses (if not externally supported - which they never seem to be) makes the task of applying a uniform compressive load across the TIM joint hugely difficult

light wbs are better, always

be cool

JFettig · Jan 23, 2003

I would agree, make it as light as possible, I dont see any reason you need the block that massive? can you show us?

WinFlex · Jan 23, 2003

Thanks alot for all the feedback. As of now, the block is in the design stage, so there is no actual prototype. I will post some picks for you guys tomorrow (the drawings are on my computer at work).
NeoMoses: As far as the top goes, the block will have a 1/2 inch lexan top. Also, are the bolts you use metric M6 (ISO) or US6 (ANSI)?
So, again, I will have some picks up tomorrow. The block itself is a 3.2x2 inch piece of 3/4 copper110 plate. It has a spiral type cut-out that is 1/2 inch deep, leaving the optimal 1/4 inch base thickness (a little more than 6mm for you metric people)...
I'll put the pics up first thing tomorrow!

Jessfm · Jan 23, 2003

One of my design goals has always to eliminate the excess mass of a waterblock, as BillA rightly points out the ture problems that can make the very differnce.
Keep it as light as possible, and yet stick to your design goals.
Wasted material only serves to hinder, not help you.

JFettig · Jan 23, 2003

winflex, 1/4inch isnt the optimal base, People thought that for some reason before, but thinner is usually better. it kinda depends on the design, if you take a look at geminicools waterblocks, they use a .05inch base on the low flow spirals, cathar uses like a .03inch base, I ususally shoot a .07-.1inch base.
Also, the channel depth isnt always needed to be that deep, its almost useless.....you want high velocity in the block.

Jon

Les56 · Jan 23, 2003

Jessfm said:
One of my design goals has always to eliminate the excess mass of a waterblock, as BillA rightly points out the ture problems that can make the very differnce.
Keep it as light as possible, and yet stick to your design goals.
Wasted material only serves to hinder, not help you.

Hi Jess,
Been a long time since we spoke.May bell you sometime - phone number still the same?

Possibly for wbs with a large bp area extra weight( in bp thickness) may not be wasted:-

JFettig · Jan 23, 2003

nice graph and refrenece, btw, whats 'h[bp to Water] W/m*m*c' exactly mean? is that something like water velocity?

Jon

Les56 · Jan 23, 2003

JFettig
"h[bp to Water] W/m*m*c" is a calculated "heat transfer coeff "(h) between the water and a plane in the bp . The plane is 0.1mm distant from any internal wb furniture and parallel to the the heat source die surface. Maybe coould be better defined but ................

WinFlex · Jan 24, 2003

yes, can someone please explain the significance of the x-axis in the graph

but yes, if anybody has more data regarding cooling efficiency related to base plate thickness and or area, it would me much appreciated!

Cathar · Jan 24, 2003

One thing to note about block weight is that a low centre of gravity also helps to reduce the torsional twisting effect of a block sitting sideways. A block can weigh twice as much as another, but if it's very low and flat, it can actually try to "twist" less. Standard gate/hinge theory from high school physics.

BillA has it though. Worry about a block weight of 200-600g with a CoG that's 6-10mm from a CPU die, is nothing compared to what a whole stack of tubing hanging sideways off a block can provide in terms of rotational torque trying to pull the block away from the CPU.

Arguing about block weight that sits low and close to the CPU is like arguing about whether or not it's going to rain when you're about to be hit by a 100 foot tidal wave in comparison to the torque load that most people have with tubing hanging off their blocks.

JFettig · Jan 24, 2003

Les56 said:
JFettig
"h[bp to Water] W/m*m*c" is a calculated "heat transfer coeff "(h) between the water and a plane in the bp . The plane is 0.1mm distant from any internal wb furniture and parallel to the the heat source die surface. Maybe coould be better defined but ................

... that really didnt make too much sence, what does all these big numbers on the x axis mean? and that little explination didnt really tell me much.

Jon

WinFlex · Jan 24, 2003

Pics

OK. Here are the pics! The block's total weight is roughly 1.3 pounds = 591grams. Regarding torque about the TIM, it should be low since the block's CoG is close to the socket!
Base is 3/4 in copper 110 plate, channel is 1/2 in deep. Pattern on bottom creates turbulence... base is 1/4 in thick.

WinFlex · Jan 24, 2003

http://winflexers.tripod.com/b1.jpg
http://winflexers.tripod.com/b2.jpg
http://winflexers.tripod.com/b3.jpg

Just type the url's in your address bar, it should work. All those lame free webhosting companies won't allow external linking!
Sorry, this is the best webspace I could find in such a hurry

Hoot · Jan 24, 2003

Man, that is a big chunk of copper!
Are you going to be cooling a big peltier with it? Otherwise the mass, IMHO, is way overkill and I'm big on over-engineering.

Hoot

WinFlex · Jan 24, 2003

yes, I was thinking about switching to a 1/2 inch plate design, then a 220W pelt, then a 1/4 inch coldplate...
uh, can somebody also please verify that the mounting holes for the block are 6mm holes in a 36mm x 66mm rectangle?

Hoot · Jan 24, 2003

36mm X 66mm looks good, but I'm not certain on the motherboard hole diameter. Some say 4mm with board clearance at 6mm and some just say 6mm. Using 6mm holes on your block should cover both.

Hoot

Block weight

Member

Member

Member

Member

choke man

Hey! I showered! Senior

Member

Registered

Hey! I showered! Senior

Member

Hey! I showered! Senior

Member

Member

Senior Member

Hey! I showered! Senior

Member

Member

Inactive Moderator

Member

Inactive Moderator

Similar threads