Phase 1: Continued...

The tower style air trap/reservoir/eyesore is just as described earlier.   Here is a picture of it during construction.  The cool part is the 1" hose running from the "Y" to the manifold ... it is a good indicator of any air left  in the system.  Oddly enough, once the system was filled, this became a  place for small amounts of air to get trapped.  Since the flow is slower here (especially with the 350 GPH pump), little air bubbles got trapped along the top of the tube.  I suspect that when the more powerful pump gets put on in the future  than this will stop.  One great benefit of this type of air trap is that  because of the width of the vertical shaft slowing the downward flow to a crawl,  the length of it giving air plenty of time to free itself from the current, and the initial upward flow giving the air bubbles an even better shot at escaping the coolant, it only takes one pass for all the bubbles to be removed from the flow.  Truly, you cannot design a better functioning air trap than this.  I was amazed when I first fired up the system at how there were never any "little bubbles".  There were either big bubbles or just straight water.  It took only five minutes of fiddling with the bypass valve (to force the air out of the radiator) to fully prime the system and remove every last bit of air.   The only remaining pockets of air are in the uppermost portions of the manifolds, and this cannot be helped, though it will never be a problem in the performance of the loop.

See the bubbles?  Kind of weird looking, but other than in the high points of the unused areas of the manifolds, there is no air  anywhere in the cooling loop.  I don't know, it gives it character.

Initially, since there was added strain to the pump because of height, the  miraculous temperatures I got in the proof-of-concept run were hurt somewhat.  All temps across the board went up 1.5 degrees.  I would hypothesize that the coolant temp actually remained the same, and that the lower flow (I estimate  less than 1 GPH through observation of flow

through the 1" line ... please don't ask for equations since that is a very unscientific statistic) is what caused the swing.  Due to the  fact that the flow rate is so low, I will keep the bypass fully closed at this  point.  If I open the bypass, I don't think enough water will pass through the radiator and the temperature of the coolant will suffer for it.

There were only two problems with this whole evolution.  Problem one was that there is room for two hard drives or floppy drives.  If you want a  floppy, you're limited to one drive, or you just have to cram the second drive  into spare space.  I didn't like that idea, and being that I have two hard  drives, a floppy drive, and a ZIP drive, I got rid of the floppy and ZIP drives.   I'll just access them over the network in the future (on machines that don't have multiple hard drives).  I don't like having components hanging free in  my cases.  The second problem involved the LEDs at the front of the case.  When I fired the computer up for the first time, they didn't fire up with it.  I checked the wiring, and they were wired in right, but they just didn't work.   LEDs are cheap, though, and I'll probably be modding them later to indicate HDD/NET usage instead of HDD/Power anyway.  I know when my computer is on ... I want to know when it is accessing the network or being accessed (maybe by our friends in Redmond?).

Once everything was running right, I closed it up again, set my Linksys router next to the case, and my Lexmark laser printer on top.  All that is left to add is the fans for the box, though even without the fans (I still only have the desk fan blowing across the radiator) I have a temperature spread of 38-43C.  One thing that I already notice is that  this setup is already putting off significant heat.  Within minutes of  closing the door to the room (70 sq. ft.) it starts to heat up significantly.  It'll be fun to see what happens when three machines are on the loop!

One other thing to add is that the internal case fans are  quiet.  The PSU is an Antec Truepower 480 with automatic fan controls,  and the front fans are two Chieftec 80mm fan running at 5V, with one pushing  through the other in a wind tunnel arrangement.  Between them  they make very little noise (the case must be open in order to hear them).  When the fan in the front is turned off,  you can hear a slight bass vibration from the pump, but your ear must be  within one foot of the cooling box in order to hear it.  You can  slightly feel the vibration from the pump when you hold the 1" return tube  on the reservoir.  With the Comair Rotron Caravel fans running on a  ceiling fan controller, I should be able to tune them down to near silence, creating one bad-ass cooling system with nearly no noise.  Currently, the only noise coming from the system is from one of the hard drives ... it  is a constant high pitch whine that never ceases.  It almost sounds like someone using a circular saw at a distance of 400 yards.  Quite  annoying, and I may install some sound deadening material.

Since my system temps are still slightly better than what I was previously getting with the BIX and football reservoir in the more compact form, I'm going  to leave well enough alone and call an end to this phase of the project.   The Caravel fans should be in at the beginning of December, and my second PC (an  Athlon 700 slot A) will be added by mid-December.  At this time I will also be adding coolant additive (Hyperlube) in an experiment to see how much it  affects the efficiency of the coolant loop.  Check back soon and see how  the project is going!

Appendix:

This computer is being used as a basic HTTP/FTP server for intranet use as well  as a gaming rig.  The rise in temps from when it was initially hooked up to the cooling loop to when it was put into its new case is also probably due to  the equalization of coolant temperatures over time, since it takes a while to appreciably raise temps in the coolant mass, which currently sits at 4.25  gallons.  Three days after initial installation of the block into the new case, my CPU temperatures have dropped to 38C/44C, probably due to the thermal compound setting.  All in all, Phase 1 is a smashing success.

Phase 2 of the series is up! Click here to go there. (12/21/02)

Coming Soon.... Phase 3.

If you have any comments or Questions please PM me on the Pro/Forums