Lets see.
Quote:
Originally Posted by Tachyon
Why 3 independent pumps.
Well two reasons .. first I wanted to have the coolest coolant available to each of the critical devices. Second - since water, like electricity, follows the path of least resistance, I did not want to give it a chance to favor one device over another for any reason. In a design where you have a header or "Y" adaptor there is nothing stopping some or all the coolant from flowing down a single path.
|
I am not sure you understood my suggestion in my last post, I suppose I was not being clear on it.
In a loop set up in series, the temperature difference on either side of any given component is almost negligible, assuming your loop has good flow rates.
If you ran the loop in series as follows, Cool rez>CPU>GPU1>GPU2>NB>SB>Warm rez, the difference in water temperature between the beginning of the loop and the end of the loop would be maybe,
maybe 5 C if you had very poor flow.
Do some searches on these forums, they are a wealth of information about the details of WCing.
Now if you still want to run independent parallel loops then I will drop it, but I think you would be better off running your loop in series. Run the numbers, you might be surprised at how much energy water can take before it shifts 1C.
Quote:
Originally Posted by Tachyon
Speaking of water blocks....
I'm really liking the design and results I'm seeing on the new D-Tek Fusion for the CPU. On the GPU's probably going to be danger den unit (when available) NB & SB ..hmm hadn't really decided .. they're not really that critical.
|
I will restate that the NB and SB probably don't need that kind of cooling unless you are seriously abusing them.
The important thing to know about the D-Tek Fusion is it is jet-impingement. That means reduced flow rates, although the Fusion claims to have fair pressure drop for a jet-impingement block.
Quote:
Originally Posted by Tachyon
....sizing down the tubing....
Here is my thinking here - you have 2 or in my case 3, 3/8 coolant lines each with a fixed flow rate.
It's pulling out of the cool supply rez - thur the loops to the warm rez tank.
Now - what size tube/s would you use for the lines going from the warm tank to the cool tank via what ever I use to remove the thermal energy with. Answer: in my case 3 loops x 3/8" (9/8ths or 1 1/8th) or I'm ok to round down to 1" with the understanding that there is going to be a slight flow increase in those lines.
|
I think by starting at the run between your warm and cool rez you may be placing priority at the wrong end of you loop, from a design justification stand point.
Better to start where the action is, that is the water blocks and the pumps.
Quote:
Originally Posted by Tachyon
In a single loop system where you have 1/2” or 3/8” tubing throughout it’s a none issue since you’re minding the conservation of additive flow rates. (there are none to add!) But in a multi-looped system or one where you’ve divided the flow some how … I would think that to be efficient you’d need to mind those additive flow rates.
|
What I do not understand is when people take a 1/2" supply line and divide it into two 1/2" lines via a "Y" (or worse yet a "T").
From a total flow volume and rate stand point this would seem to be problematic or at the very least inefficient. A 1/2” input to two 1/4” outputs “Y” adaptor is what I’m talking about here for example. Am I way off here or is this just not an issue? Has this been tested before?[/quote]
I would argue that is even less efficient to handicap a loop(s) by undersizing
any portion of the tubing than it is to use a Y or T.
Again, plenty of good sources if you do a couple of searches.
Quote:
Originally Posted by Tachyon
...you are considering a chiller rather than a radiator...
Well I've got some ideas there that I'm not done messing with in my brain yet. But I've just read about the Switftech chiller with it's performance levels and it looks interesting.
|
I understand how that is, I have been mulling a case design around in my head for more than a year now and only have it a third finished. No rush.
Quote:
Originally Posted by Tachyon
Ever notice how your basement floor is always cold .... mine is .... and it's < 8' below me... Why not place a radiator/fan down on the basement floor to utilize the lack of thermal energy to remove some from my system.
|
Don't have a basement, but I walked down you train of thought. You may be surprised where you end up. If you had a big enough basement it would be possible, but keep in mind that earth is a wonderful insulator.
You will almost certainly end up with a large volume of air to use as a heat dump but if you leave your computer on you will eventually be no better off than if you had stuck with more traditional methods.
Quote:
Originally Posted by Tachyon
In a sealed system pump heads are irrelevant (theoretically) The pump is just acting as a recirculation pump. In this scenario anything could be used as it’s “in the basement”. Like a full sized automotive radiator with a real 12vdc fan on it .. ok way over kill …. But you get my drift. You could use say a automatic trans-cooler with integral cooling fan quite nicely.
|
I don't think your first statement is entirely correct. In reality, you pump still has quite a task just keeping the water flowing within a loop inside one's case. Adding a ten foot vertical probably won't help temperatures any. The less tubing the better.
-----------------
It seems to me that you still haven't mentioned the entire story, yet.
I get the impression you already have plans (and parts) in place or almost so.
Care to enlighten my sagging mind.