Ultimate waterblock

[Fixittt]

Ben when I get done with the current project in hand, shoot me the design, Ill see what I can get done.

[bigben2k]

Originally posted by gone_fishin
Don't know how you came to that conclusion by what I said

:shrug:

Can't wait to see it when it is done

You wrote:
"A restriction is built into the base of my EQ design ..."

I assumed that it meant that you can't try out different restrictions. Am I wrong?

[gone_fishin]

Originally posted by bigben2k
You wrote:
"A restriction is built into the base of my EQ design ..."

I assumed that it meant that you can't try out different restrictions. Am I wrong?

Well they have been tried in the various prototypes I made. And although my testing is not real accurate for sure, I settled on a 4mm cylinder down to the 1mm thick bp. A trade off again, narrowing the cylinder means reducing the area of this bp sectional thickness and widening it reduces the area of the geometry over the core and reduces the velocity.

[MadDogMe]

By MadDogMe
"A block alone does not make or break a system 'regardless of pump and rad and tube', never has never will, that is my opinion"
end quote

Strongly vague opinion if you ask me. Regardless of pump, rad, tube. So given any combo of these, the waterblock cannot break the system? Go get yourself the BlueCooling's BTMS block and live with it

Don't be pedantic!!(how do you spell forcetious???) the emphasis was on the make .

I think what Cathar is trying to point out, is that he is using a restriction as a nozzle directly over strategic areas in his block. This lowers the overall flow of the loop but increases heat transfer because of the extreme velocity directed at various points of heat transfer, for a beneficial tradeoff. He is not using a restrictive barb nor do I.

I understand this, but 'restrictive barb' does'nt automaticaly equal 'restriction as a nozzle directly over strategic areas'.

I in fact do use a 5/8" OD barb as the inlet on my design.

Thank you!!...
************************************************** *

I think a thin base, Inlet over core, 'Radial water path' is my ultimate block design... (with dual parralel 'OCPC' The Deep rads with 14mm barbs, and a 1060 pump[or better equivalent]. with 12mm inner diam' silicone, and 14mm equal Y's)
I think calculating bore size is important to gain the most out of the 'restriction' aspect of the block.(in relevance to tube size )

Would it be possible to 'chart' progress at varios flow rate/to/speeds?, to find a velosity 'sweetspot'?. Use the 'universal' 12mm~inner tube size and chart the various 'optimum mm~squared channel size' to 'X' flow rate.
would this produce a clear graph?, enabling you to calculate at a glance?...
Or is that so much B#llsh#t?, it's early here! ...

[LiquidRulez]

Originally posted by MadDogMe


I think a thin base, Inlet over core, 'Radial water path' is my ultimate block design...

what ,in your opinion, is a thin baseplate?
4 or 5 mm?? what?

[MadDogMe]

I have no idea what would be best , i don't pretend too. but at a guess I'd say my design could'nt go below two millimetre, unless I put 'fins' in a clocklike pattern arond the core that meet the next layer, to give it strenght, stop it bowing. does that settle your query?.

[bigben2k]

Originally posted by Fixittt
Ben when I get done with the current project in hand, shoot me the design, Ill see what I can get done.

Thanks Fixittt, I'll definitely keep you in mind! I'll try to finish my tweaking out asap. In the mean time, maybe you can give me an overview of your limitations/parameters, since this thing might require some small cuts.

[gone_fishin]

Originally posted by MadDogMe
Don't be pedantic!!(how do you spell forcetious???) the emphasis was on the make . I understand this, but 'restrictive barb' does'nt automaticaly equal 'restriction as a nozzle directly over strategic areas'. Thank you!!...
************************************************** *

I think a thin base, Inlet over core, 'Radial water path' is my ultimate block design... (with dual parralel 'OCPC' The Deep rads with 14mm barbs, and a 1060 pump[or better equivalent]. with 12mm inner diam' silicone, and 14mm equal Y's)
I think calculating bore size is important to gain the most out of the 'restriction' aspect of the block.(in relevance to tube size )

Would it be possible to 'chart' progress at varios flow rate/to/speeds?, to find a velosity 'sweetspot'?. Use the 'universal' 12mm~inner tube size and chart the various 'optimum mm~squared channel size' to 'X' flow rate.
would this produce a clear graph?, enabling you to calculate at a glance?...
Or is that so much B#llsh#t?, it's early here! ...

As for overall flowrate, a number of engineers have posted up various graph work to show the benefits of added flow to conduction. A system balance is still needed, a bad example would be a powerful pump with 1/4" restrictions throughout the loop. Too much heat generated by the larger pump and the system chokes it off. Now having high flowrate and pressure of say 5gpm which accelerates in the block alone will give the possibillity of the benefit of fluid velocity even though flow is restricted to a smaller degree.
I am still testing on my block with possible ways to run it. I do not have the equipment to give highly accurate results but if a benefit in one way over another is repeatable by me then I will move in that direction. My latest test (if you followed the design in my other thread) will be to choke off the upper layer of holes in the block. This will undoubtedly reduce overall flow but at the same time increase velocity in the remaining passageways. Will a benefit be readable by me? I don't know yet so the fun continues.

[Can O' Beans]

This is my theory & planned system:

1. High flow block, a lot of surface area and a decent amount of turbulence at the metal surface, just enough to break up any laminar flow - self designed.

2. Nice, big heatercore with either 2 120mm in & out, or 1-2 of the 24v Comair Rotrons(most likely) at 12v.

3. All tubing & barbs will be at least 1/2" - no hard 90deg turns.

4. 700gph Danner Mag Drive Pump (Pondmaster), NOT sitting in a res...

Now with the first 3 designed with greater system efficiency in mind, the heat from the pump can be a setback. I plan on removing most of the pump heat from the water system itself. It will be a very compact internal closed loop system.

Before this gets fully implemented, I'm designing a WC system monitoring curcuit using a PIC MCU and I will be using the digital thermometers for more accurate readings. This might take a bit as I'm fighting bad computer addiction & slothic(new word!) lazzyness and I have to actually LEARN PIC programming But I plan on progress pics and temp measurements - flow is easily adjusted. I would like actual flow readings, etc.. but don't know how to implement that (cheaply).

[MadDogMe]

To realy get the most out of velocity, don't you have to design the block to a given flow rate?, a lesser flow and the velocity just would'nt be there would it?, or it would be deminished.
This was more what I meant when I said about tuning what you have, designing the block for a given system...

[Cathar]

The biggest barrier to improving water-block performance in my mind is the base-plate thickness.

This is the one reason why almost every good block on the market performs about the same (within 2C) as another.

The copper is the primary source of thermal resistance. What you do on top of the base-plate thickness (channels, dimples, etc) has minimal effect if the copper base-plate is 3mm or more in thickness.

You want to optimise a design that works well with very thin base-plates (1mm or even less) if you want to have any hope of pushing the envelope.

It's still possible to do this properly with even rather mediocre pumps (at least 1gpm actual flow rate though).

[MadDogMe]

It's true ain't it?. the way I see it is you want to remove the heat from the CPU as it's occuring, you don't want the heat being 'stored' in the blocks mass, and cooling heat that was produced 5seconds/whatever ago...
You could always have the core area with an ultra thin layer, and let the rest be abit thicker for strenght/integrety.
How small/large an area do you think you could get away with being 1mm or less?, without the need for fins or pins reaching from the base to the top plate for support?. 5mm around the edge of the CPU's footprint?...

PS. is there a way of milling a slope/gradient?, without steps?...

PPS. why don't more people cool direct die?, I'd have thought jetting water onto the die directly would be the ultimate???, perhaps make a Die~sized mini plate of Cu to stick somehow to the die, very thin(thinner than a block would hold up to) but with ridges/channels/whatever to increase surface area...

[myv65]

The thickness that Cathar uses requires fins above that push against a cover. Without that, the baseplate would bow under the load resulting in zero contact over much of the die.

As for direct-die cooling, you'd find it far inferior to pretty much any water block. The die area is simply too small to allow enough convection under the most concentrated of flow streams. Let's also not forget that at some point even pure water will result in material erosion (not to mention being "unfriendly" to the CPU and motherboard). This isn't a big deal when you've got millimeters of metal to play with.

Yeah, you could make direct die cooling work with very high flow velocity (or very low coolant temperature), but for a given velocity/temperature you'll do better with a block.

Cathar's block works well because it has a very thin (low delta-T) baseplate that connects to sufficient surface area (the fins). Most blocks use a thicker baseplate to transfer heat both directly through the baseplate but also radially toward the edges. If you can connect enough surface area to the region directly over the die while controlling deflection, you'll get very good results. This is what Cathar has done and I personally find it pretty impressive.

[godsdice]

Cathar- do you use a shim? I expect using one would allow for more pressure than the AMD specifications as well as less bending of the base plate if there is a possibility of that?

[Cathar]

I don't believe in shims at all.

The base-plate does not bend. The tops of the fins push against the bottom of the top plate and brace the base.

I put a LOT of pressure on my CPU dies, probably around 50lbs for my AthlonXP's if not more. No shim is necessary.

[MadDogMe]

Please don't think I was crapping on Cathars block when I said I did't think 4DegC was a big difference to the H20 'genre', it was meant towards H20 as a whole, not blocks(Cathars especialy) to a specific. I can understand how much time effort must of gone into those gains, but it's exactly that time & effort that says to me there's no 'leaps'& bounds' left to be made:shrug:. I think every deg is gonna be a struggle, and maybe 7-8DegC over manufacturers temps today being the absolute limit IMO.
I did'nt say it was'nt a worthy cause though ...

Can you tell me how thin you can go without supports?.

Shims should'nt actualy touch if the blocks on right/levell should they?, it's more to stop it tilting on fitting and pressing~chipping an edge off the CPU?. so if you're careful they're not needed. it was only the big awkward copper HSF's that warranted them I think...

[Cathar]

The absolute thinnest you could make the base-plate without any bracing is about 1.5mm, but really it depends on what you have above the base-plate (open?, channelled?, spikes? etc). Anything that gives structural bracing in both the X-Y (horizontal) dimensions could get away with 1mm thin, bracing in 1 dimension I'd say 1.5mm and no bracing (open-pool) 2mm. Thing is it'd be very hard to get a 2mm open-pool block to perform well as open pool blocks really rely on the lateral heat spread of a thicker base to widen the convective cooling area.