No need for "devising"...I have a full coverage DangerDen block sitting in my desk drawer.
Not gonna use it though, for two interconnected reasons.
The card is an old 6800GT and is scheduled for replacement at some point (as is the entire system actually) and watercooling the vid card adds complexity to the system out of proportion to the benefits.
Right now, I could undo two screws and have the Apogee Drive off the CPU and the motherboard is free.
Add the vid card into the loop and everything gets a whole lot harder.
If I had a decent card to work with, my attitude might be different but as it sits the maintenence issues outweigh the value of the card.
Also, the configuration of the Apogee Drive's barbs really limits hose routing options.
It was clearly designed as a CPU-only application (although they claim that you can add other blocks to the loop) and adding the vid card would probably triple the amount of tubing required.
Minimizing the tube runs is a good thing- this is not an especially high output pump anyway and the aesthetics of the loop are better with less hose.
I have a backup 7200 card that is passive and that's probably the easiest solution.
Now for basic performance info.
Bear in mind that this info does not constitute a rigorous evaluation of the waterloop but given the bounds of my limited equipment, I feel it does show some relative results.
Prior to converting, my aircooled idle temps ran to the mid to high twenties.
I never bothered stress testing while under air.
To get an idea of room ambient I used a Craftsman InfraRed thermometer...
That's 16.8°c, BTW.
Just a foot to the right of the case is a window (closed) and outdoor temp is @ -12°c so yeah, it's a bit nippy this morning.
I ran a stress test called Orthos (works both cores) for twenty minutes.
Temps had pretty much stabilized after five minutes and varied little after that, so twenty minutes seemed adequate.
Basically, with watercooling my fully loaded temps are about the same as air-cooled idle temps.
The ▲T (the difference between ambient and CPU temps) is right at 15°c which I would consider pretty bloody good.
Next I opened the window and gave the PC a blast of arctic air.
I have long considered building some sort of duct from the window to the PC case, this test was just to see what I might gain from doing so.
Naturally, room ambient plummeted as well...
Brrrr.
I then stopped the stress test and idle temps were...
...ridiculous.
All in all, a fairly revealing- if not terribly comfortable- experiment.
Finally- for the curious and skeptical- a shot of a few of the air cooling methods I tried...
None of them are absolute state-of-the-art but they aren't complete dogs, either.
If I were to go back, I'd probably use the Zalman 7700 (essentially, the oldest of the designs shown) as it had the best performance/noise ratio of the group.
The Zalman 9500 outperformed it's older sibling but the fan had a very annoying click, a failing I considered remedying by removing the proprietary fan and sandwiching the sink with two 120mm Scythes.
That probably would have been stellar and I may do it at some point just to see but for now it'll just go back on the shelf.
I may go to Home Depot and check out my ducting options today.
If anything happens, I'll let you know.
Bookmarks