New design, incorporating many ideas.

[Cathar]

Originally posted by GTA
I base this on the fact that its a lot easier to drill smaller holes than it is to mill narrower channels.

My thought is that an adaption of this design for very powerfull pumps would be for example 1mm holes, very very close together at say 2mm deep, and 0.25mm jets. Huge water velocity into the holes, and a HUGE amount of turbulence circulating over what is essentially a very thin base.

What are your thoughts on this? I'm still toying with the idea of a huge pump, 30 PSI or more, just to test, as I have found a tool hire shop that will hire out for £20 per day a four stage electrical pump that produces a 50 metre head.

Obviously noisy as hell, and enormous, but It would be nice to see what sort of effect power like that would have on a design.

There are a few factors to consider with this design.

A nice one is that as the holes/jets drop down in size, the total orifice area drops down by a quadratic factor, but the number of holes also increases quadratically, so it all balances out. I really think that the jet diameter should be half of the hole diameter, but that's just my thoughts on the matter. It's an attempt to balance volumetric flow/velocity and the need to get the jet to reach the bottom of the well.

Making the base too thin (thinner than about 3mm with all the holes) means that the base will start to flex under mounting pressure. The basic honeycombing between the holes assists in strength, but once we start getting down close to 2mm thick or less you'll have to solve the issue of base flex, potentially easily solved by having support posts, but then, how to make?

Third issue is that high velocity water will gradually eat away and erode the copper. This is not something I have a lot of experience in, but very high velocity water is often used to cut metal. Not that I think that we're anything near those levels, but over a few months of high pressure jetting will we find that the block starts to leak out of the base as the jets bore their way into the remaining metal?

Use a 50m head height pump may result in issues with frictional heating. Forcing largish volumes of water through narrow holes is going to generate heat, along with the heat that the pump itself is dumping into the water. How to get rid of that heat? For a 50m pressure head pump pushing through a mass of little holes, the amount of frictional heat generated is going to be quite large is my guess. Think of pumping up a car tyre and how hot the pump head gets. That's just with 30PSI (~20m pressure head) and a low viscosity gas (air). Think of it this way. If the pump is providing 200W of pumping power, but after passing through the jet holes it loses 100W of pumping power just forcing through those holes, then the missing 100W is being converted into heat energy.

[theetruscan]

Metal Cutting Water Jets are run at PSI levels approaching 60,000 PSI. I don't think we need to worry about this at all. A factor of 2000 is nothing to laugh at.

Wow, 60000 PSI.

EDIT: Even then, abrasive particles (i.e. Garnet) are often mixed with the water to make the cutting more effective.

[LiquidRulez]

I wouldnt use a pump that powerful.
1.....it would be complete overkill as Cathar stated in his own terms.........Erosion....Id have used the term EXPLOSION!
2.....What the h&ll are you gonna seal it with that will hold that kind of pressure??
3....WTF do you think the effect will be on your block with that amount of pressure ???.......
Even if your block will withstand that amount.......I dont think your "jet" holes in your "shower head"will last that long!
4..Where would you possibly put a pump of that size?
Youre going to connect this thing to your PC?..........
Why dont you just donate those system components to someone before they get destroyed in your experiment?
5....Are you serious????????

[Skulemate]

As theetruscan has said, I wouldn't likely worry too much about erosion of the block. However, you're going to need tubing better than regular Clearflex or Tygon (rated to 25-35psi) as well as a radiator suited to those high pressures (50m H2O ~ 70psi). The block shouldn't be a problem though... o-rings and solder joints will handle more pressure than the pump can produce... you'll have to watch the threaded connections though.

[Graystar]

Originally posted by Skulemate
However, you're going to need tubing better than regular Clearflex or Tygon (rated to 25-35psi)

Clear Braided Vinyl Tubing. Common stuff. About a dollar a foot. Rated around 160 psi.

http://www.plumbingstore.com/cleartubing.html

[GTA]

Originally posted by LiquidRulez
I wouldnt use a pump that powerful.
1.....it would be complete overkill as Cathar stated in his own terms.........Erosion....Id have used the term EXPLOSION!

He didn't say it would be overkill. It would be though

Its something I would like to try for a short period of time, just to see what happens. Thats why I mention hiring the pump rather than buying it, these things cost about £700 to buy, which is way out of my league.

2.....What the h&ll are you gonna seal it with that will hold that kind of pressure??

I'd solder it together, a decent solder joint will easily take 50 PSI. An o-ring would be pointless, yes, but I've never used one ayway, I always solder everything together, very strong joint.

3....WTF do you think the effect will be on your block with that amount of pressure ???.......
Even if your block will withstand that amount.......I dont think your "jet" holes in your "shower head"will last that long!

The effect on the block would be negligable, if it would take the pressure without leaking, then it'd be no different to running a normal pump and block together.

What would happen to my jet do you think?

4..Where would you possibly put a pump of that size?

On the floor.

Youre going to connect this thing to your PC?..........

Yeah, I'd just run the hose in through a hole underneth my PCI slots.

Why dont you just donate those system components to someone before they get destroyed in your experiment?

I honestly don't see why this should break anything, as long as everything is sealed properly. For example, the fittings used by Bladerunner would be fine for this job, they're rated up to 100 PSI or more IIRC. Failing that, some hydrolic fittings and pipes would easily handle this pump.

5....Are you serious????????

Absolutely.

Incedenty, for a radiator, I was thinking of a huge coil of copper pipe. Basically a huge drum of the stuff. Hydrolic fitting or something on each end, just run the water through it.

And I wouldn't use threaded connections, I'd find brass fittings that could take that pressure, and solder them on as well.

[Volenti]

You could use a normal rad if you place it after the water block in the cooling loop, the high pressure "zone" is only the pump to block section (including the actual pump and block themselves)

but a large res would suffice for short term testing, which is what you're doing anyway.

[GTA]

I have used my almost professional standard drawing skills to produce the diagram below, which I hope will explain things slightly more clearly.

If anyone wants to pay me to draw diagrams of their blocks to a similar high standard, please let me know. But be warned, quality like this does not come cheap



The only thing left to explain is on the side view of the nozzle, the top plate sits just above where the sloping edges connect to the inlet barb.

[jaydee]

You guys seriously underestimate O rings and NPT threaded barbs eh? What do you think Forklifts use in their hydrolics. O ring will take it just peachy if it is installed right as there shouldn't be all that much pressure on it. Most of the pressure should be taken by the two copper peices. The peices should be touching each other taking most if not all of the load. The O ring should not have to deal with much PSI. If a 30 ton Hydrolic press's NPT threaded brass barbs can handle it.......

[Skulemate]

Originally posted by jaydee116
If a 30 ton Hydrolic press's NPT threaded brass barbs can handle it.......

That's semi misleading jaydee. Telling us that the press is 30 tons doesn't give us any idea of how much pressure the hydraulic system developes...

[jaydee]

Originally posted by Skulemate
That's semi misleading jaydee. Telling us that the press is 30 tons doesn't give us any idea of how much pressure the hydraulic system developes...

Ok I guess. When I lift 15,000lbs of steel 10ft in the air that weight is being stressed somewhere. It was my theory that it was in the hydrolics as that it what makes it go up and down then from there to the ground. Guess I could be wrong though.

[Bundles]

My pnuematic air rifle charges to 3,500 psi, and it AND its charging bottle use threaded brass fittings, and they dont pop, so unless his water pump producess about 3000 psi min, it aint gonna pop. the pipes on the other hand, are a diffrent story.:shrug:

[Skulemate]

Bundles, is there an o-ring in these threaded fittings? If so then it's not really the threads that are sealing against the pressure... garden hoses do this for example.

[LiquidRulez]

Well, I dont know exactly what type of pressure this pump produces. But I can assume its quite alot considering it pumps liquid roughly 150 feet head, in what is undoubtly a >1"diam pipe....probably around 2".

Thats quite alot to expect out of a waterblock made by (no insult intended) someone who really has no idea to put one together , that will hold under the stress that its going to be under and deal with, with no leaks !........
Unless of course, GTA has skills, and is well versed in what he is proposing to build utilising this monster pump.....


Whats the point in using a pump as described, again?

Hey.... if you think you can do it, then by all means, do it. And good luck!!

[GTA]

No insult taken liquid rulez, I am new here, but have been watercooling for several years.

The block I am proposing here is not my first block by any means. I've made about 7 or 8 CPU blocks, a couple of GFX blocks, 1 NB block, a HDD block, a PSU block, and a GFX ram block.

I'm not as experienced as some here, I'd be the first to admit that, but I reckon that this is very doable.

Admittedly, I use a very simple method each time, but I find it very effective. Simple soldering, cheap to do, very good seal.


The point? actually, you've got me there

There's not really any point doing this, its not in any way a 24/7 useable cooling solution. However, I'd look at in the same way that people look at using LN2 on a chip; Worth doing, if only to see what happens. I mean, if I did it, I bet you'd like to read the results? If someone else did it, I'd definatly give it a read.

Its just a test, would it work? what would happen? and If I can do it for less than £30, then I don't see why not.

[Cathar]

Originally posted by LiquidRulez
Well, I dont know exactly what type of pressure this pump produces. But I can assume its quite alot considering it pumps liquid roughly 150 feet head, in what is undoubtly a >1"diam pipe....probably around 2".

Thats quite alot to expect out of a waterblock made by (no insult intended) someone who really has no idea to put one together , that will hold under the stress that its going to be under and deal with, with no leaks !........

A 50m head pump (~150') is supply around 75PSI pressure when dead headed (outlet end is blocked).

Threaded barb fittings sealed with various thread sealant are rated for well above 1000PSI. The thread sealant I use on my blocks is rated for 10,000PSI when cured after 24 hours. 1000PSI is its instant load rating.

O-rings can handle well over 300PSI, depending if they are implemented properly.

A waterblock handling 75PSI is quite doable. Heck, even a Maze 3 is rated for that kind of pressure I believe. I'm sure Danger Den state somewhere it's tested at 80PSI, or something like that.

[murray13]

Without more info on the pump I made a guesstimate of 140' of useable head. Which with 12 jets of .8mm (1/32") gives a velocity of about 92f/s or 28m/s or 60 mph or 100 kph. With a flow rate around 3gpm. (forgot to write it down)

[GTA]

cerwick update guys and gals ( any gals actually here? )

Anyway......

Base, finished and drilled. My camera doesn't do closeups, so this is pretty much the best i'm gonna do.



The spray generated through the nozzle ( have left this as a link, the picture has been left very large, because I didn't want to lose any of the already diminished quality )

http://www.gta45.pwp.blueyonder.co.uk/spraytest1.jpg

Pump used ( Eheim 1048 ) doesn't really give an impression of how its really going to work. I normally use a 1260 ( 4 times more flowrate and 3 times more pressure ) but I'm not unhooking that from my system to get one picture

The holes on the nozzle are lined up perfectly with the holes in the base, although I lost 4 1.5mm drillbits doing the nozzle holes. The holes in the base are 2.5mm.

Is it usual that when holes are drilled into the base, when I turn it over, I can see drill impression marks on the side that will go over the core?

hehe, maybe 0.5mm is too thin?

[Volenti]

Originally posted by GTA

Is it usual that when holes are drilled into the base, when I turn it over, I can see drill impression marks on the side that will go over the core?

hehe, maybe 0.5mm is too thin?

hehe, lap very carefully