Enzotech-LUNA CPU waterblock

[Zehnsucht]

.......... polynomial .......... ??

............ top secret

You do know that the expression y=x^2+x+m is a polynomial?

EDIT:
Well obviously you didn't. But now you know.

And you really can't let excel generate the curve for you. You must let the measured up points create the curve, not the other way around.

[Vapor]

You do know that the expression y=x^2+x+m is a polynomial?

EDIT:
Well obviously you didn't. But now you know.

I think he was showing that a polynomial trendline looks like with those three data points (and how it's not the trendline he's using).

Still, I don't think three data points is a very solid basis for creating a trendline with a massive amount of extrapolation.

[Zehnsucht]

I think he was showing that a polynomial trendline looks like with those three data points (and how it's not the trendline he's using).

Oh yes, you are correct. I didn't catch that. Obviously I need sleep.



I still agree with Vapor though.

[AndreaBZ]

Another example... with five points

The equation that I use (very similar to a hyperbolic function):



polynomial trendline (Excel):


The first curve, to me , describe correctly the physical phenomenon.........
follows the points and approaching asymptotically the X axis.
Is very precise for flow rate 100/800 lph........... and also > 800 lph.
Likely for flow rate < 50? lph is less accurate (wrong).

The second curve simply follows the data, right or wrong.

************************************************** **********************************


equation that I use :



if k = 1 and n = 1 ...............



which is the equation of hyperbole "equilatera"

[Zehnsucht]

During the lunch break I calculated the function used. It may not be accurate outside the interval [100 800]

f(x) = 0.100248 * x^-0.1 + m, where m is 0.0819907 and 0.0975 (the latter I just made an educated guesstimate).

Max deviation I get is between the aX^n numbers and my calculated is 0.00018014, which is good enough in my book.

Can I has a gold starr?

[AndreaBZ]

During the lunch break I calculated the function used. It may not be accurate outside the interval [100 800]

f(x) = 0.100248 * x^-0.1 + m, where m is 0.0819907 and 0.0975 (the latter I just made an educated guesstimate).

Max deviation I get is between the aX^n numbers and my calculated is 0.00018014, which is good enough in my book.

Can I has a gold starr?

equation that I use :



if k = < 1 and n = < 0 ...............
The equation becomes

Y = k * X^n

that is the formula used by Excel
The same formula used by Martin........for the radiators
Delta temp. vs RPM (Air rate)

"...........The first curve, to me , describe correctly the physical phenomenon.........
follows the points and approaching asymptotically the X axis.
Is very precise for flow rate 100/800 lph........... and also > 800 lph.
Likely for flow rate < 50? lph is less accurate (wrong).
............. "
See above

[NaeKuh]

All I know is, the block works pretty good at keeping the CPU cooler. Just put it on my test bench last night, and I'm quite surprised.

Very nice block,

andyc

the block uses cathars pitted design from his storm.

i wouldnt be suprised.

[Waterlogged]

All I know is, the block works pretty good at keeping the CPU cooler. Just put it on my test bench last night, and I'm quite surprised.

Very nice block,

andyc

Andy, please tell me your doing a proper posted review of this block? It would be nice to actually see how it performs compared to the V2, GTZ and Supreme and not find some kind of egregious error in the results....then maybe I can shut up and this topic can die.

[Waterlogged]

I'll be glad to let you and other's know the results direct if you would like. Just received my CF/Dallas Matrix temp monitoring HW/SW via UPS today, so I'll have the new bench set up proper ready for some serious testing tonight.

<img snip>

Should have some prelim stuff by the end of the weekend,

andyc

Take your time Andy, and PM me the final results.
Are you planning on testing all the nozzles?

[AndreaBZ]

I'll be glad to let you and other's know the results direct if you would like. Just received my CF/Dallas Matrix temp monitoring HW/SW via UPS today, so I'll have the new bench set up proper ready for some serious testing tonight.

Should have some prelim stuff by the end of the weekend,

andyc

temperature difference (Core CPU - H2O)
at different flow rate? ......... three or four ?

[Vapor]

What's the major difference between the latest rev, and the original Luna when it was first released?

Accommodation and inclusion of nozzles I think

i.e., Rev. A comes with nozzles and works with nozzles while the original can't even fit/take the nozzles.

Looking forward to your results

My CF/Dallas probes should be here fairly shortly too

[Vapor]

What does that mean, please clarify????

andyc

I think he wants data in a few varieties, or is wondering if you're going to do it.

1) CPU temps minus water-in temps.

2) #1 as it changes with a decrease/increase in flowrates.

[Waterlogged]

Not sure really. As a matter of fact, I'm not even sure what revision of the block I have, or if it take nozzles.

One things for sure though, it's working pretty good just the way it is. Once I see how well this one measures up against a Supreme, I might just buy the latest version.

What's the major difference between the latest rev, and the original Luna when it was first released?

andyc

Not sure myself. All the images I've seen of both rev's show them both to be very similar internally. It'd be nice if Enzo-Tech(Dav) could magically re-appear and explain the differences.

[Vapor]

equation that I use :



if k = < 1 and n = < 0 ...............
The equation becomes

Y = k * X^n

that is the formula used by Excel
The same formula used by Martin........for the radiators
Delta temp. vs RPM (Air rate)

"...........The first curve, to me , describe correctly the physical phenomenon.........
follows the points and approaching asymptotically the X axis.
Is very precise for flow rate 100/800 lph........... and also > 800 lph.
Likely for flow rate < 50? lph is less accurate (wrong).
............. "
See above

A trendline with so few datapoints (and tightly packed datapoints in the grand scheme of things) really shouldn't be used for so much extrapolation. What's worse is that the endpoints of the actual data have the largest error compared to the trendline--does not bode well for the accuracy of the extrapolation.

I've played a lot with c/w vs. fan RPM for aircoolers and power trendlines do not extrapolate well at all. They look the nicest and are certainly the lesser of 6 evils, but the power trendlines do not play well with cooling functions when it comes to extrapolation (long story short--they lack accurate vertical asymptotic behavior as flow approaches 0 and also don't give due justice to the diminishing returns behavior of flow in waterblocks).

The interpolative strength of a trendline is viable and great for estimators, but I'd really really be careful with so much extrapolation :-/ You're doing a lot of extrapolation using a trendline that doesn't even really match the few datapoints you have, especially not at the extremes of the data :-/

[NaeKuh]

blah andy i just want to know what performs better and at what flow cost.

thanks please!

[AndreaBZ]

[AndreaBZ]

As I said earlier,

That information means absolutley nothing without the asscociated core temps and ambeint rad air in temp. At least not to me.

Do you have those plots also?

andyc

http://www.xtremesystems.org/forums/...d.php?t=206542

[Waterlogged]

[autobot]

I hate to say this, but what is the purpose of this chart?
It just shows that 2 setups with identical computer and cooling components equal the same results.

[LockBreaker]

If D-Tek not soon releases their 1366 mounts for the Fuzion V1 I think I will be going with this one for my 1366 Core i7 setup.

The block looks great, but hows the performance? Saw Malik had some numbers, where the Luna performed better comparing to the GTZ.

[Sarg]

my next build is on the I7 chip and WCing it, so You have yet another set of eyes on ya, tis okay no pressure I like the way it looks and yes that goes for the funky little binky light on it. So yes seeing some Stats on this puppy will nice.

[LockBreaker]

Anyone who knows if the BP 7/16" ID 5/8" OD will fit the Enzotech Luna without those 45 degree fittings from BP?

[Leeghoofd]

I just received my 2 samples ( with one brand new block, can't tell the name for NDA) from Enzotech will test them next week on the I7 platform (when my probes get here) against the EK supreme, Swiftech GTZ and co...

Again tests will be kept simple and reproduced in coretemp charts at stock, 3.6-3.8 and 4ghz clocks (like the swiftech review) but also with added in and out watertemps...

[OdinEidolon]

I just received my 2 samples ( with one brand new block, can't tell the name for NDA) from Enzotech will test them next week on the I7 platform (when my probes get here) against the EK supreme, Swiftech GTZ and co...

Again tests will be kept simple and reproduced in coretemp charts at stock, 3.6-3.8 and 4ghz clocks (like the swiftech review) but also with added in and out watertemps...

when will this new block be avaiable you know?

[Leeghoofd]

I will post some piccies later on...





Exterior looks alike, but inside it's totally different... can't wait to test these blocks...