Re-Design of the Jet Block.

[bigben2k]

Originally posted by deepblue
another point about the pipes. get them to dip as much into the cup as you dare. the ideal distance between cup-button and pipe-bottom should be somewhere around 1mm - so if you're cup is 4mm deep, make sure the pipe reaches about ~3mm into the cup.

I disagree.

Cathar posted a graph of the area affected by a jet of this nature, and 1mm certainly isn't it. There is a need for some distance, to maximize the jet effect.

[jaydee]

Is there a part number from McMaster?
This is the smallest I can find:

Form
Round Tubes

Material
Alloy 260 Brass

Outside Diameter
1/16"

Wall Thickness
.014"

Length
1'

Package Quantity
15

Outside Diameter Tolerance
± .002"

Specifications
ASTM B36

If my calculations are correct the ID is .0485" or 1.2319mm. Pack of 15 1' pieces is a little excessive though. Might work out alright with 1/8" wide holes.

[Gulp35]

oops double post (my first ) look below

[Gulp35]

BB2k where is this graph that cathar made?

what lee has been saying it all good. you want to calculate out the best heat transfer situation.

these are the dimentions for the jets that I am planning on using for my block.

Jet ID. --- 1/32" - .79375mm.
Cup width - 3/32" - 2.38125mm.
distance from jet to base of cup
1/8" to 5/32" - 3.175mm. to 3.96875mm.

I'm going for around 50 jets.
but for around 19 jets 1/16" would be the easyist(sp?)

Lee what do you mean by "the water will be sucked out"?

JayDee116- I found alot of tiny tubes at my local hobby supply store.

[jaydee]

Originally posted by Gulp35


JayDee116- I found alot of tiny tubes at my local hobby supply store.

There is 1 hobby supply store in this city and they do not carry anything resembling tubes of anytype let alone tiny brass or copper ones. Already called and asked. Small city, only 300,000 people here. I would much rather order things anyway. I don't have time to drive around for hours looking for and picking up stuff these days. Would rather it be delived to me at work and ready for use when I get home. I used to burn more time looking for stuff around this city than actually working on anything.

[bigben2k]

Originally posted by Gulp35
BB2k where is this graph that cathar made?

...

JayDee116- I found alot of tiny tubes at my local hobby supply store.

I think it's in Lee's thread.
http://forums.procooling.com/vbb/sho...&threadid=6032


JD, yeah, that's it, 1/16 tubes.

From the "CPU backside cooler" thread:

Copper tubes:
P/N 8967K51
1 PKG = $9.98
Package content: 20 tubes, 1 foot length each.
OD: 1/16" (.0625", or 1.59 mm)
ID: .0345" (0.88 mm)
wall: .014" (0.36 mm)

[leejsmith]

Originally posted by bigben2k
I think it's in Lee's thread.
http://forums.procooling.com/vbb/sho...&threadid=6032


JD, yeah, that's it, 1/16 tubes.

From the "CPU backside cooler" thread:

is this the one you wanted ben




gulp. this is from ben i had 1.4mmID tubes with 2.5mm cups at the time.

Ok, i ran the numbers, and there is indeed a problem: the ratio of the hydraulic equivalent of the outlet area, to the tube ID, is about 2:1, which means that the coolant is indeed being sucked out.

I'd recomend either using a smaller tube, or a bigger cup. If you follow Cathar's graph (above) you'll have some good guidelines, for those dimensions/proportions.

this is the link for hydrolic calculations

http://www.lmnoeng.com/PipeDuct.htm

[ozzy7750]

ok, i need a little bit more clarification on this 'sucking out' issue.

is the ratio i need to work out the outlet area:inlet area, or hydraulic diameter outlet:hydraulic diameter inlet?

if it is the latter, how do i calculate the hydraulic diameter of the outlet?

i have 3.5mm cups, with 3/32 tubes (2.38mm od, 1.67mm id)

[PoLicE]

Wouwe great block leejsmith. Can you compare the performancewith a WW in your system?

[bigben2k]

Originally posted by ozzy7750
ok, i need a little bit more clarification on this 'sucking out' issue.

is the ratio i need to work out the outlet area:inlet area, or hydraulic diameter outlet:hydraulic diameter inlet?

if it is the latter, how do i calculate the hydraulic diameter of the outlet?

i have 3.5mm cups, with 3/32 tubes (2.38mm od, 1.67mm id)

I had to dig pretty deep, to get some answers to some questions, including that one.

As has been pointed out, the hydraulic diameter can be calculated using the formulae for an annular duct, as found here:
http://www.lmnoeng.com/PipeDuct.htm

What it boils down to, is that the hydraulic diameter of the annular duct is equal to the gap between the outside of the jet tube, and the cup wall.

So... if that gap is smaller than the jet tube's ID, then the flow speed is higher as the coolant exits the cup, and that's the "sucking" effect that I'm talking about, which would significantly reduce the jet effect.


As for the distance of the jet's outlet, from the cup's bottom, I believe that there was a reference made to where it needs to be somewhere between 3 to 5 times the jet's ID. I could be off though.
[edit] I found it.

By Cathar
Jets develop fully over about a 4d distance. Less than this and what tends to occur is a concentrated "mash" of turbulent water striking the base.

Also, while the jet travelling down will lose a small amount of power due to the distance travelled, it also becomes more turbulent (or so it is hypothesised in scientific papers) and this leads to greater thermal efficiency. Again, search through google. There are many studies of the jet distance vs efficiency effect, and every single scientific paper I read found that getting less than 4d in height results in a performance drop off, hitting a peak plateau between 4-5d (different papers differ on the exact best distance) and then dropping off after that.

Also, some references here:
http://forums.procooling.com/vbb/sho...=&postid=81386

[/edit]

JD: I found a pic of the baseplate to the "Cascade", and I'm reposting it here, because I needed to show you something:



There's a fine balance in the gap, between the cups. If it's too small, then there's not enough heat rravelling up the fin (aka gap). If the gap is too large, then the mass is too large, and accumulates heat. Cathar describes it, in his thread:
http://forums.procooling.com/vbb/sho...&threadid=6666

[leejsmith]

Originally posted by PoLicE
Wouwe great block leejsmith. Can you compare the performancewith a WW in your system?

at the moment i can only compare againts a maze 3 but i do have a cascade on order.

[MadDogMe]

Have you finished with the M3 yet lee?, or do you have some more designs/testing in the pipeline?...

I though the wall thickness was for structural strenght and the secondary impingment was 'incidental'?...

[bigben2k]

Originally posted by MadDogMe
I though the wall thickness was for structural strenght and the secondary impingment was 'incidental'?...

Far from it.

The second inpingement is what allows the baseplate to have any kind of structure, with the remaining copper there. The "incidental" really, is that if you use a 3/8" plate to start off with, you are left with a good structure that should remain integral, under the clamping pressure.

These remaining bits of copper act as fins, and the second inpingement allows it to function optimally, with minimal thermal resistance.

[leejsmith]

Originally posted by MadDogMe
Have you finished with the M3 yet lee?, or do you have some more designs/testing in the pipeline?...

I though the wall thickness was for structural strenght and the secondary impingment was 'incidental'?...

as i have a block that is 2C better than the maze 3 after many tests i can use that to compare againts so i will send it back asap.

thanks MDM.

btw i have some 3mm copper if you would like some.

[Cathar]

Originally posted by MadDogMe
I though the wall thickness was for structural strenght and the secondary impingment was 'incidental'?...

They really are symbiotically linked. Quite a few things at play on that one.

You want the structural strength, true, and the cup walls give you that, so they are a necessary evil. The cups walls are also there to contain the off-flow from the jets to stop the jets from interfering with each other, so again they are necessary. The off-flow from the primary jet can be used as a "secondary" impingement. This secondary impingement is deliberate and not incidental because the cup width can be such that the impingement's efficiency is reduced. Since we have to have the walls there, we may as well make use of them in the conduction element of the cooling of the block to allow heat somewhere to go in extremely low flow scenarios otherwise the design will suck at low flow when the impingement effects start to lose their power.

[Zymrgy]

hmmm...just a thought here, but has anyone thought of what might happen if one used square/retangular pockets instead of round holes?

[jaydee]

Originally posted by Zymrgy
hmmm...just a thought here, but has anyone thought of what might happen if one used square/retangular pockets instead of round holes?

How do you make a square hole that small?

[Zymrgy]

originally posted by jaydee116
How do you make a square hole that small?

Assuming that your holes are 1/8" in diameter it really should not be that bad. You can use one of your 1mm endmills...this would leave a .020" radius in the corners & a .085" flat....like so...

[jaydee]

Originally posted by Zymrgy
Assuming that your holes are 1/8" in diameter it really should not be that bad. You can use one of your 1mm endmills...this would leave a .020" radius in the corners & a .085" flat....like so...

Interesting idea. I have no clue if it would be better or not. If I ever get around to building this I might try it.

[Zymrgy]

the whole square pocket/hole idea is based on something Cathar said...

originally posted by Cathar
They really are symbiotically linked. Quite a few things at play on that one.

You want the structural strength, true, and the cup walls give you that, so they are a necessary evil. The cups walls are also there to contain the off-flow from the jets to stop the jets from interfering with each other, so again they are necessary.

now what I am thinking here is that by having many round holes in a pattern, the wall thickness between the holes are far from being a constant...that is it changes depending on what angle you are looking at. By having a pattern of squares, it is much easier to minimize the wall thickness of the cups....to more of a constant.

Disadvantage is is that it makes it more complex to make, but not overly so....especially for a non-production block.

[Cathar]

Originally posted by Zymrgy
By having a pattern of squares, it is much easier to minimize the wall thickness of the cups....to more of a constant.

How so?

What impact do you think that the current 0.25mm variation from the thinnest to thickest sections of copper walls between cups will have on performance?

Do you think that a non-circular chamber will significantly alter the secondary impingement in a negative fashion at the corners of the square, resulting in reduced efficiency where it's most needed (at the thicker wall sections)?

[Cathar]

If you wanted to make the wall thickness more uniform, you'd concentrate on a hexagon shaped cup, not a square or a circle, however, one now has to consider what the impact will be on jet impingement efficiency now that the cup is no longer the same shape as the jet.

[MadDogMe]

Since we have to have the walls there, we may as well make use of them in the conduction element of the cooling of the block to allow heat somewhere to go in extremely low flow scenarios otherwise the design will suck at low flow when the impingement effects start to lose their power.

But that's incedental not designed for right? 0.25 is the thinest you could safely go I'd imagine(Cu IS soft! ). I thought about using brass and having 0.1 walls, 1 or 1.5mm cups to make the 'secondary' heatpath negligent?. Drilling straight through the brass and using copper 'foil' (0.2mm) for the bottom (you can buy it in varying thin gauges from RC/hobby shops I believe). But I don't know how 'fine' you could go without it cutting through the Cu when clamped?...

PS. I'd use surgical steel syringes for the jets. Not that I could ever make this (to 0.1mm accuracy) but it's good to dream ...
_______________________________

Lee, Id love some 3mm copper. Cutting that 10mm stuff across the middle is a bitch!!

[jaydee]

Well I changed course again . I know Bigben might start to .

I have been tweaking the laser at work to try and get it back to engraving like new and have come as close as I can get being the boss doesn't want to spend a dime on getting replacement parts it desperatly needs...

Anyway I made a new desing today. It has 11 jets. The holes will be drilled with a 1/8" endmill into 1/4" thick copper. I will have to measure the jets once I get home with my micrometer as they are smaller than what they were drawn to be because the laser melts away a little more material than the design is drawn to. The holes I did a pilot hole with the laser and am going to attempt to drill them out with a drill bit.

Also this is a single inlet and SINGLE outlet design. I threw this together because I am trying to route 3 computers onto one loop and having 3 barbs on one block makes it a bitch. I will leave the Lemon Block though as the last block in the loop because it is the best block and because I can easily rought the dual outlets into the resivore.

My capture card is messed up again so it looks like I will have to reload the drivers and what not before I can put up anymore pics. Hopefully this weekend I can find sometime.

[bigben2k]

Originally posted by jaydee116
Well I changed course again . I know Bigben might start to .

My prerogative.