Asetek R&D for WCing

[BillA]

http://www.asetek.com/DownloadArea/M...eport_v1.0.pdf

interesting process description

[wijdeveld]

Thanks for the link,
You're welcome to discuss the article in here or on the Asetek forum

[Brians256]

Originally Posted by wijdeveld

Thanks for the link,
You're welcome to discuss the article in here or on the Asetek forum

Thank you for being so open about how you designed your water cooling system! I call it system, because it is obviously designed to work together.

I am surprised at some of your results (in the graphs), but I believe I know why.

First, I would like to comment on your design goals. All decisions flow from attempting to meet stated (or in a pathalogical case, unstated) goals. It was great to see how your decisions flowed from your clearly stated goals.

1. Easy bendable non-kinking hoses.
Good goal but lacks specifics. Perhaps you should shoot for a minimum bend radius that you wish your hose to achieve without kinking? As you state your goal, it cannot be tested. In fact, because there is no way to test your stated criteria, you don't show any verification of this goal other than saying you meet it.

2. Optimal hose diameter (not bigger than necessary)
Excellent. Short goal and to the point. Easily testable.

3. Hose resistant to hydrolysis and aging
This is another fuzzy goal. What level of resistance do you want? Perhaps you should state a lifetime that you wish your tubing to last? How about maintaining clarity over the stated lifetime (this would help specify your coolant additives too)?

4. Radiator capable of removing 400W of heat
I don't believe an engineer wrote this requirement. You can remove 400W of heat with a chunk of styrofoam if you accept a high difference in temperature between coolant and ambient air. A better goal would be a maximum delta-T between ambient air temperature and coolant temperature for a given wattage of heat input. An even better goal would be a maximum delta-T between ambient air temperature and the CPU/GPU/Northbridge on-die temperatures (simulatable of course). This goal should help define the radiator you choose, the fans you put on it, the pump you specify, etc... It is a true system goal.

5. Optimal Price/Performance ratio
Unless you define what units you will use for "performance", this is too fuzzy to test. It is probably much easier to look for the best performance you can achieve for a given price.

6. No aluminum in contact with water (due to corrosion issues)
Great goal. Succinctly stated, and it includes your reason for the goal. Easy to test.

You said:

The hose test confirm that increasing the flow, both with a bigger pump and larger hoses, do not have any positive impact on performance.

Surprised me at first. However, you give the impression that your decision is valid for all block styles. This would, of course, be false. You didn't evaluate different block styles at different flow rates, did you?

There is much more of this type of testing in your document that can be misleading to the casual reader. I think you might have gotten much better performance (5C or better) if you used a larger pump, larger tubing and a different style block. Perhaps others will disagree. However, it is likely that your decisions were based upon making a much more cost-effective system than what I describe. A company can go broke if it does not follow the "good enough" principle and tries to go for performance at all cost.

In any case, thank you for exposing your design process. It was very interesting to see how you achieved your decisions!

[bigben2k]

I'm not sure which Forum at Asetek would be the right one.

I believe that we've discussed this paper before, no?

[Les]

Oh dear.
Chapter One
Temperatures are a joke - 0.1c not suffice to give an accurate wattage.
Notwithstanding accuracy:-
Water temperatures indicate a flow-rate of ~ 5X in the Half-moon as the Spiral/maze giving similar performances( Table1and Fig1.9).
Yet Fig10 shows, at the same flow-rate, the Half-moon design giving better performance than the Spiral/maze design.
Unless I have on my fuzzy head there is something amiss.

[BillA]

Ben
the paper is a year old, no doubt others have seen it, I had not

note that there is another link to a paper of Arjan's in the testing section,
http://forums.procooling.com/vbb/showthread.php?t=9536
same vintage but I had not seen it either (or do I forget ?)

I will comment on both, but after others have had a crack at it

I intend to leave the Asetek forums be; they are of, by, and for Asetek
(and I don't wish to re-visit the European/US WCing differences)

[wijdeveld]

Let me clarify some things first. The Aseteks R&D report is a joint venture between Asetek and me. I'm responsible for modeling part. I'm NOT an employee of Asetek and I'm mainly interested in the physics behind cooling and translating that into an understandable numeric model to evaluate performance and possible improvements of the cooling solution (also for phase change setups).

So, my test bank is a computer and some source code I own which when compiled is called a flow model.
Yes, the paper is old, we wrote just after finishing the Waterchill development. Since then Asetek has introduces an alternative block with better performance but with a higher back-pressure. There are also some improvements on the ‘old’ block, especially with regard to the radiator (Yes I know, the 400 watt claim is a bit of marketing nonsense since it is not defined at which water temperature/flow rate that performance is reached, the water could be cooking)

If you’re interested in a more general comparison on (non air) cooling techniques follow the following link:

http://www.home.zonnet.nl/wijdeveld/cooling/part1.pdf

Part 2 has been published in parts on the Asetek forum since their design was to much intertwined in the numeric tests to publish this on personal title. You might want to look into:

http://forum.vapochill.com/showthrea...&threadid=2819