New swiftech storm review. Enjoy

[ricecrispi]

http://www.systemcooling.com/swiftech_storm-01.html a recent review.

Don't you love looking at this block. Thing is so cool.

[cotdt]

i'm going to use this block to cool my peltier

[AngryAlpaca]

That's an inherently bad idea. The Storm doesn't have a very large cooling area on the base. It's designed for, at most, a 14*14mm CPU, not a 40*40mm TEC...

[bigben2k]

Way to go Robotech!

[Cathar]

Originally Posted by AngryAlpaca

It's designed for, at most, a 14*14mm CPU

Restate that as a 20x20mm IHS covered CPU die, and I'll agree.

[cotdt]

oh no! is there a 50mm x 50mm mega version of the storm / stormG4?

[gazorp]

Originally Posted by cotdt

oh no! is there a 50mm x 50mm mega version of the storm / stormG4?

NO! That is not the focus of these blocks.

[Mathelo]

Cathar,

The reviewer suggest using a filter with this block. Is that really necessary?

Louis

[jaydee]

That was way more than a review. That was a lesson + a review.

Probably should use a filter. Any little thing will plug those jets.

[Cathar]

Originally Posted by Mathelo

Cathar,

The reviewer suggest using a filter with this block. Is that really necessary?

A filter is not necessary so long as the loop is clean, and filled with clean coolant, and is closed to the open air (as in dust/lint can't get into the water).

[csimon]

a filter? what kind of chunks would he expect to catch?

perhaps icecycles ...LOL

[Dave]

Good article guys, and it helps my support work.

I often get email asking if they should use 1 or 2 MAG's with a Storm block, and I have been suggesting 2.

From your charts, it suggest only a 0.5-0.8 DegC difference in true CPU temps from one to two pumps, if I read correct, Stew?

[Cathar]

Originally Posted by Dave

From your charts, it suggest only a 0.5-0.8 DegC difference in true CPU temps from one to two pumps, if I read correct, Stew?

Yeah, around about that. A ~0.5C improvement would be a reasonable bet. Depends on the CPU heat load really, as well as the size of the CPU die itself. Smaller dies will yield larger improvements. I'd say a 0.2-1.0C sort of improvement range.

Robotech was nice enough to generate a performance vs pumping power graph:



Click here for source of graph.

The MAG is a ~1W applied-to-the-waterblock sort of powered pump for purposes of that graph. Putting 2 in series would probably get you about 1.75-2.0W applied pumping power (after factoring in power loss due to other components).

As you can see from that graph, the Storm doesn't need a particularly powerful pump to do its thing.

[Risky]

This block is using a central inlet and single outlet. In curios as to why this chosen over dual (or more). Is this because the dual outlet setup is inefficeint or of marginal benefit or becasue recombing them in a Y-piece (as opposed to running dual lines to the reservoir) caused resistance?

[jaydee]

Originally Posted by Risky

This block is using a central inlet and single outlet. In curios as to why this chosen over dual (or more). Is this because the dual outlet setup is inefficeint or of marginal benefit or becasue recombing them in a Y-piece (as opposed to running dual lines to the reservoir) caused resistance?

If I remember correctly Cathar tested both dual and single outs and found there was hardly a measurable difference between the two and Sigle in and Single out was MUCH easier for the end user to route hose. Cathar can correct if wrong.

[Dave]

Thanks Stew,
Just wanted to double check before passing info to our customers,

[Les]

Originally Posted by Cathar

......Robotech was nice enough to generate a performance vs pumping power graph:

http://www.leesspace.com/images/Web_...dPwr-Graph.gif ...........

Tend to have more faith in Swiftech's "red-cross" version as being more representative of cpu performance.


Traditionally Swiftech(Billa) uses a 10x10mm die which I suggest is more comparable to cpu cooling. I do not accept Robotech;s assertion " the relative relationship or ranking of waterblocks tested on different systems should be relatively consistent ." Some of my thoughts on Die-size are illustrated here (from one of Incoherent's yhreads


Suspect that the "thermal resistance", shown by Swiftech, is the virgin resistance as determined by Incoherent . This does not include the contribution of TIM and "sensor offset" to the Die/Coolant temperature differential.

[eander315]

At first glance, you might assume they are just plain, drilled holes. Actually they appear to be bored and have a tiny spike sticking up in the center of the bottom – amazing! This obviously creates more surface area and serves as a great conduit for heat transfer into the thin base.

Aren't those little "spikes" in the center of the cups designed to create turbulence, rather than increase surface area? And how would they serve as a conduit for heat into the thin base? Don't they mean out of the base? And doesn't the heat travel more efficiently through the thin base than the thicker spike? Maybe I just misread previous discussions on jet impingement cooling.

[jaydee]

Originally Posted by eander315

Aren't those little "spikes" in the center of the cups designed to create turbulence, rather than increase surface area? And how would they serve as a conduit for heat into the thin base? Don't they mean out of the base? And doesn't the heat travel more efficiently through the thin base than the thicker spike? Maybe I just misread previous discussions on jet impingement cooling.

Increase surface area is sure a added benifit. As for the conduit for heat it really is subjective to perception. Is the water cooling the base or removing the heat? Either way it is heat transfer.

[Long Haired Git]

If you look at the pictures showing fluid velocity during impingement, the area where the spike is located has lower velocity flow....