Waterblocks effectiveness in terms of power dissipation

[bigben2k]

I based that on the on-line calculator's parameters. (wrong? or a limit of the calculator?)

Of course it doesn't include other ways of shooting a liquid, and the geometry and alignment of the baseplate.

[BillA]

Les
I think Ben is a masochist

ok Ben
wtf you talkin' about ?
2 - 12 is enough spread to cover an ocean
SysCrusher was referring to a 462-B, and you ?

ahh, I see your latest post
so what does a java calculator have to do with your determination of "best" ?

this is drivel
why do you need to add such tripe to your already inflated post-count ?

get serious Ben

[bigben2k]

Ok, I incorrectly assumed that the calculator's limits were meant as parameter limitations. I was wrong.

Sorry!

The other point still stands: the circular motion is a result of the opening's geometry.

[airspirit]

I think Ben's first post was on the mark 100%. This is NOT measuring the cooling power at all.

An easy way to describe this would be to take two blocks, one of high resistance, and one of low resistance. Looking at the original equations, you would raise W by increasing the resistance of the block. Raising the resistance of the block would lower the flow rate. Lowering the flow rate decreases the efficiency of the block, and you get higher temps.

Conversely, with a low resistance block, you get more flow, and more efficiency, and while W goes down, temps lower as well.

The perfect waterblock would draw heat from the die in the most efficient manner possible while providing the least amount of resistance to flow. Increasing the resistance of a block and thereby increasing the amount of time a slug of coolant remains in it will NOT lead to better temps. While I commend you on the effort, I think you are measuring a different type of force, not the cooling potential of the block.

[SysCrusher]

Originally posted by Les
News to me .

EDIT:: Flomerics calculator maximizes at ~ 2.0 -2.5D but this is not for submerged jet impingement and in no way considers cases outside the calculators limits.

Being that it's submerged completely changes things. I don't understand what you two mean with 2D - 12D. I assume that it means 2 times the diameter. If that is the case that would great for non-submerged. In the case of 12D, that is one high pressure pump?

Distance is also effected by flow, pressure and the velocity that the nozzle produces within limits of the pump being used. You want more pressure against the base plate. So with our low pressure pumps, bringing the nozzle closer to the base will give the required pressure since the jet no longer has to fight against the surrounding water. Pressure from the jet is needed to thin the boundary layer and turbulance will come from velocity.

BillA, what would be interesting is if you would find away to lower the hose barb opening so that it is 2mm above the base plate and test that 462-B (that's the one with the barbs directly ontop?) again to compare with what you have done with different barb sizes. I feel you would find great benefit to cooling, atleast with the 462-B with it's big chamber size.

[bigben2k]

Originally posted by airspirit
I think Ben's first post was on the mark 100%. This is NOT measuring the cooling power at all.

Actually, I think I was at 97.7% (1.6 from 70W)...

It is relevant to calculate the heat induced by the pump, or more specifically, the heat induced by the flow restrictions.

As Since87 demonstrated, it can be easily calculated.

[SysCrusher]

Flow restrictions shouldn't add to much if any. Unless it's a very strong head pump or the restrictions are so great it causes pump cavatation.

[murray13]

Originally posted by airspirit
....An easy way to describe this would be to take two blocks, one of high resistance, and one of low resistance. Looking at the original equations, you would raise W by increasing the resistance of the block. Raising the resistance of the block would lower the flow rate. Lowering the flow rate decreases the efficiency of the block, and you get higher temps.

Conversely, with a low resistance block, you get more flow, and more efficiency, and while W goes down, temps lower as well.

The perfect waterblock would draw heat from the die in the most efficient manner possible while providing the least amount of resistance to flow. Increasing the resistance of a block and thereby increasing the amount of time a slug of coolant remains in it will NOT lead to better temps....

I assume that you mean while using a pump that has VERY little head to work with. And results don't hold up to your statements. Cathar's WW is a 'high' (relatively) resistance block with results to challenge just about everything (low resistance) out there.

I do agree that your not measuring cooling power, rather the power used from the pump's output.

[Les]

Submerged Liquid Jet Impingement Heat Transfer(SLJIH)

The only submerged jet data with which I have played are presented http://widget.ecn.purdue.edu/~eclweb/jet_benchmark/ in the "Heat Transfer Results as Excel spreadsheet".
I do not offer a graph because of "Any unauthorised use, copying or mirroring strictly prohibited ".
However the graphs show the Heat Tranfer profile changes with conditions.
However I do ,naively, consider the "plateau profile" of the "3.1mm ID" results maybe the ones applicable to to flat bottomed wbs.Additonally in "fantasy calculations" I equate the "h" of the plateau to that calculated by Flomerics at r= ~2.5D.
Yes, it is "stretching it a bit" but ......................

[bigben2k]

Murray13 is right here.

This is where you want to look at your pump's PQ curve. If you were running under the peak efficiency, then a higher restriction would yield better results, because you would then have more total power applied by the pump.


I found out something interesting the other day, while looking up orifice plates for LiquidRules: there are two kinds of restrictions out there; those that provide a recoverable pressurre drop (venturi) and those that don't (orifice plate). Jet inpingement produces a non-recoverable pressure drop.

What that means is that although the fluid is flowing at a higher speed, it is permitted to enter an open area, in which it will turbulate: this speed cannot be recovered, it is simply a result of the opening. In a venturi, the speed is (essentially) recoverable. (Venturi is like a couple of funnels connected together at the small end).

[myv65]

Originally posted by bigben2k
I found out something interesting the other day, while looking up orifice plates for LiquidRules: there are two kinds of restrictions out there; those that provide a recoverable pressurre drop (venturi) and those that don't (orifice plate). Jet inpingement produces a non-recoverable pressure drop.

Ben, I've neither directly spoken with nor seen a picture of BillA (other than that X-Files spoof), but I gotta believe it's statements like this that make him shake his head and ride you hard. Technically there is no zero-loss device, though smooth transitions have lower losses than others. You've been making thousands of posts here and elsewhere about "theory" and you just figured this out "the other day"?

Well, at least I got a good chuckle on a Saturday morning. Thanks.

[bigben2k]

I'm glad I could liven up your morning

I just mentionned it, because I hadn't seen it anywhere, and because I never looked at it that way before. It was specifically about recoverable vs non-recoverable pressure drops. The friction losses are an add-on (where it applies).

It might not be relevant to anything, but it's there... It also relates to Radius, as I'm looking at trying to implement a jet (or jets), but my solutions always seem to end up into (mostly) recoverable PDs.:shrug:

[SysCrusher]

Originally posted by bigben2k
I'm glad I could liven up your morning

I just mentionned it, because I hadn't seen it anywhere, and because I never looked at it that way before. It was specifically about recoverable vs non-recoverable pressure drops. The friction losses are an add-on (where it applies).

It might not be relevant to anything, but it's there... It also relates to Radius, as I'm looking at trying to implement a jet (or jets), but my solutions always seem to end up into (mostly) recoverable PDs.:shrug:

Try a pump with a stronger head. I just got done trying out my new block with a single 1/16 jet. It simply was to restrictive for the pump. 1/8 seems to be the limit for my pump. 3/16 is ok with 1/4 being a little to much. So anything between 1/8 and 3/16. With multiple jets, the total orfice area of the jets has to be within that 1/8 to 3/16. Doesn't give you much to play with but your limited to the head of the pump which isn't all that great with these aquarium pumps.

I like to try a Iwaki (that the name?) with a nice head that uses about the same amount watts. When I tried the faucet trick that had about 50 psi and the same constant water temp as ambient air , the gains were great with a 1C difference between idle and load. Total 5C delta. Finding a pump like that without putting it's heat into the water to negate the gains is probably impossible.

[murray13]

Originally posted by SysCrusher
....With multiple jets, the total orfice area of the jets has to be within that 1/8 to 3/16. Doesn't give you much to play with but your limited to the head of the pump which isn't all that great with these aquarium pumps....

Not really SysCrusher. If you can only go down to a 1/8 single jet because of lack of head you will not be able to go smaller with multiple jets. Hydrodynamics rearing its ugly head again.

[SysCrusher]

Originally posted by murray13
Not really SysCrusher. If you can only go down to a 1/8 single jet because of lack of head you will not be able to go smaller with multiple jets. Hydrodynamics rearing its ugly head again.

I don't understand why. If the total pressure and velocity of the smaller jets was equal to a single jet with the same pressure and velocity, wouldn't the two be the same? I must be missing something.

I have no interest to use multiple jets. Over a wider area of heat it's great. To cool an area the size of a cpu die all is needed is one single impingement. Once one tries to use multiple jets so close together, it effects the wall flow and possibly the stagnation area.

[Les]

Originally posted by SysCrusher
I don't understand why. If the total pressure and velocity of the smaller jets was equal to a single jet with the same pressure and velocity, wouldn't the two be the same? I must be missing something.

Maybe relevant:
Some rough sums:

Based on Billa Data extracted from GIF* by Inspection and using :


* http://www.thermal-management-testin...vs.%20flow.gif

[bigben2k]

Originally posted by SysCrusher
Try a pump with a stronger head. I just got done trying out my new block with a single 1/16 jet. It simply was to restrictive for the pump. 1/8 seems to be the limit for my pump. 3/16 is ok with 1/4 being a little to much. So anything between 1/8 and 3/16. With multiple jets, the total orfice area of the jets has to be within that 1/8 to 3/16. Doesn't give you much to play with but your limited to the head of the pump which isn't all that great with these aquarium pumps.

I like to try a Iwaki (that the name?) with a nice head that uses about the same amount watts. When I tried the faucet trick that had about 50 psi and the same constant water temp as ambient air , the gains were great with a 1C difference between idle and load. Total 5C delta. Finding a pump like that without putting it's heat into the water to negate the gains is probably impossible.

I caught that, thanks.

I've got a Little Giant 2-MDQ-SC, that has a max head of 14.6 feet.

The problem I'm encountering is a dead flow spot in the block. I'm aiming for a jet, but of course because of the unusual inlet geometry, I have to tune it all just right, to maximize what my pump can do.

I ran some tests with my pump, where I cap the outlet with caps with different size openings. I was originally shooting for 3/16. Right now, I'm looking at 1/4. I also tried a 3/8 test.

None of my tests were conclusive. Part of the reason is the calculation (missing measurement?) of the pressure drop. I also suspect that the pump is behaving erraticly, because it has a design limitation (i.e. it doesn't do 0 head), or maybe I'm just completely wrong and nothing has anything to do with anything...:shrug:

See LiquidRulez thread for a bit more info.

[BillA]

Originally posted by bigben2k
. . . .
nothing has anything to do with anything...:shrug:
. . . .

only to those who cannot connect the dots
a (repeat) gift for the 3432 count poster

this is how pump 'data' is commonly presented (w/o a curve)



but if you think BACK to the pump characteristics link in pHaestus' article, you might recall a discussion of pump (impeller) efficiency
do you know what that means ?
here is more complete data for the same Hydor pumps (but 50~)



do observe how the min-max ratio changes as the pump size increases
getting a clue yet ?

Thank you Bruce at Cooltechnica for the pumps

[gone_fishin]

"Other ways of shooting a liquid"

BB2k posted the above line and I would like to expand on it.
In some experimental data I ran across there was indeed a difference in the results of jet impingement on a flat plate by changing the shape of the nozzles (circular being among the worst). Another variable to the equation. Anyone feel like googling?

[BillA]

google ? google ?
who dat ?
whey dey at man ?

what I wanna know is . . . .

[SysCrusher]

Originally posted by Les
Maybe relevant:
Some rough sums:

Les, that's true if the id stays the same. Sure, you will lose velocity and pressure the more .25 id holes you have. If you smaller holes to equal the same pressure and velocity of one .25 id hole, their the same not including geometry.

[SysCrusher]

Originally posted by gone_fishin
"Other ways of shooting a liquid"

BB2k posted the above line and I would like to expand on it.
In some experimental data I ran across there was indeed a difference in the results of jet impingement on a flat plate by changing the shape of the nozzles (circular being among the worst). Another variable to the equation. Anyone feel like googling?

Because it changes the size of the stagnation point with a circular shape being the smallest. Also a flat plate isn't as good as a concaved plate. Another one to add to the equation. The wall flow and film is the hint.

EDIT: The shape also effects the spreading rate and turbulance levels.

[SysCrusher]

Originally posted by bigben2k
I caught that, thanks.

I've got a Little Giant 2-MDQ-SC, that has a max head of 14.6 feet.

The problem I'm encountering is a dead flow spot in the block. I'm aiming for a jet, but of course because of the unusual inlet geometry, I have to tune it all just right, to maximize what my pump can do.

I ran some tests with my pump, where I cap the outlet with caps with different size openings. I was originally shooting for 3/16. Right now, I'm looking at 1/4. I also tried a 3/8 test.

None of my tests were conclusive. Part of the reason is the calculation (missing measurement?) of the pressure drop. I also suspect that the pump is behaving erraticly, because it has a design limitation (i.e. it doesn't do 0 head), or maybe I'm just completely wrong and nothing has anything to do with anything...:shrug:

See LiquidRulez thread for a bit more info.

The pump may well do a 14 foot head but try to get a 14 foot head with a 3/16 id hose. Does that pump have the required pressure to push the water now in the same distance? It's the limitation of the impeller called back pressure.

[SysCrusher]

Here's a link courtesy of myv65 that was imformitive for me about pumps.

[Les]

Originally posted by SysCrusher
Les, that's true if the id stays the same. Sure, you will lose velocity and pressure the more .25 id holes you have. If you smaller holes to equal the same pressure and velocity of one .25 id hole, their the same not including geometry.

Without accurate data would not like to add to further to my "evasion".
Have found trying to estimate dP in different ID nozzles beyond me.
Will leave the "evasion" posted - graphs quite pretty(and maybe informative

Originally posted by SysCrusher
Because it changes the size of the stagnation point with a circular shape being the smallest. Also a flat plate isn't as good as a concaved plate. Another one to add to the equation. The wall flow and film is the hint.

EDIT: The shape also effects the spreading rate and turbulance levels.

Would be most interested in any links to any of this work.All are intuitively attractive.