My current project is shaping up
 

Greetings all. I just finished assembling my current water cooling setup tonight; right now I am doing some leak-testing before install it on my chip.

My loop is set up as follows: resevior --> pump --> waterblock --> radiator --> resevior. I built the majority of my system in a separate housing (made from an old AT case I had laying around) which is going to reside approximately 9' from my computer. Inside the "pump house" is my Little Giant 3-MD-SC pump, which is connected by some 3/4" ABS riser and a 3/4" garden hose connection to my 3" ABS resevior that holds 1.39L. My block is a D-TEK TC-4 with 1/2" fittings, and my radiator is a Chevette heatercore (modified for a 5/8" hose barb on the intake) with a low speed 120mm Panaflo connected to it via a custom tin shroud. I am using Clearflex60 Nalgene 1/2"x3/4" tubing for all of my runs, and 5/8" hose barbs for all my connections, except those on my block. I haven't done a flow test as of yet, but I am adding a pressure gauge tomorrow once the plumber's goop dries; this should allow me to determine the flow rate directly from my pump's H-Q curve.

I don't have any pictures as of the moment, but when my girlfriend comes over this weekend she's going to bring her digital camera, and I'll post some after that. In the mean time, I'm interested in any comments or questions you have.

It'd be interesting to see some temps, if you can read the CPU diode.

I am in the process of acquiring a diode reader from MeltMan, and when I get that I'll be certain to post results. As for now, when I get my block installed (probablly tomorrow) I'll post my temps vs. those I got with air cooling as a reference.

I installed my pressure gauge this afternoon, and during regular operation, it reads a value of approximately 8-8.5 psi (18.5-19.5' H2O) pressure. From my pump's H-Q curve, this translates to between 150-220 gph of flow through the system. I am fairly happy with this flow, but am not sure I trust this gauge 100% (I bought it at Crappy Trash for $6.99CAD, and its resolution isn't great... I am looking for a better gauge). Remember, stay tuned for pictures of the project in progress, coming this weekend.

Nice pressure, but you need to remember that the pump inlet pressure is negative (maybe -1.0 psi), so your flow is actually a little higher...

There shouldn't be negative pressure, as I am drawing water from an open resevior... since the resevior is 12" tall there is less than 0.5 psi pressure at the inlet (0.43 something-or-other). However, this small amount is not noticable with my current crappy pressure gauge that has a 2 psi resolution. However, I bought a shiny new pool pump pressure gauge that has a range of 0-30 psi, and is marked in 0.5 psi divisions. I'll post the accurate pressure tomorrow.

Well, it's negative, relative to atmospheric. The pump induces a pressure differential, regardless of the atmospheric pressure.

What you measured was the pump outlet versus atmospheric.

I think it was pHaestus that took some readings like that, or maybe it was someone from OC.... but I remember the negative reading at the pump intake.

What you (also) need is pump inlet versus atmospheric: that way you'll know the pressure increase from the pump in the closed loop.

Alternatively, you can do what BillA (aka unregistered) suggested: run a couple of Ts at the inlet and outlet, hook up some vertical tubing, and calculate the height difference in the columns of water, then translate it in psi. The tubes need to be of some length though...

It doesn't quite work that way in a system with a resevior. That's because the resevior is open to the atmosphere, and hence, has no pressure relative to atmospheric. It's a simple matter to write Bernoulli's equation between the surface of the resevior feeding the pump and the pump inlet itself. For ease of calculation, suppose that I place my datum at the centreline of my pump intake (i.e. along the centreline of the impellor shaft). If one assumes that the flowrate is actually 150 GPH = 0.15773 L/s, and substituting the relevant values into the equation, one finds that the pressure term at the pump inlet is equal to:

p/gamma = 0.276m - losses (measured in m H2O)

This was simply calculated as ~0.29m water (the elevation head of the resevior) minus the velocity head of the fluid at the pump inlet. Now, given, I don't know the losses that occur through the short connection between the resevior and the pump, but I doubt that they equal more than ~0.276m H2O, given my flowrate and pipe diameter... I could be wrong, but I doubt it.

One other thing. If the pressure at the intake were negative, then the flowrate would be less than I predicted since the pump is using some of its head to lift the water to the intake. The flowrate from the pump would only be greater than I determined from the H-Q curve if the pressure at the intake head was positive (i.e. if the intake pressure is 1 psi, and the pressure at the pump discharge is 8.5 psi, the pump only imparted 7.5 psi to the water, not 8.5).

Yeah, I remember that thread. If you recall, I was the one that pointed out that it wouldn't work too well for someone with a powerful pump (i.e. since my ceilings are a lot lower than the 21.5' cutoff for my pump...). I even joked about using mercury as a manometer fluid, regardless of the fact that it would be dangerous and less accurate than water. :D

I got a better pressure gauge installed, and this one reads approximately 8.5-8.75 psi pressure. This means that my flow is somewhere around 100-120 gph... not quite as good as I thought, but I am still quite pleased. (Like they say, ignorance is bliss. :D )

Actually, I meant that the pump is taking a negative pressure (say -1.0 psi) and outputting 8.5psi, which translates into a 9.5 psi pressure increase through the pump, so yes you're right, it means less flow.

I've got to find who measured that...

Like I said, negative pressure on the intake is going to be there if the system is inline, but it's not in mine because the potential head of the water in my resevior prevents it. Please don't make me install another pressure gauge to prove it... :cry:

Out of curiousity I just conducted a flow test (with my beautiful girlfriend as an assistant) for my new rig. I get a flowrate of 12L/min (720L/h or ~190 gph). Needless to say, I am very pleased with these results, as they exceed my target of 2 gpm. Also, I think that these results show that my pump intake has a positive pressure (my H-Q curve shows a head of ~8.25 psi, or 19' H2O for this flowrate, indicating that the remainder of the pressure measured on my pressure gauge is intake pressure from the height of my resevior). I'll have some pictures posted by the end of the weekend. Thanks for your interest.

Here are the pictures I promised:

http://www3.sympatico.ca/merobbins/Watercool/05.JPG
A picture of my complete setup, including the 9' run of hose to and from my block, and my Radio Shack thermomiter which I am using temporarily for the water temp. readings.

http://www3.sympatico.ca/merobbins/Watercool/03.JPG
Here's a closeup (relatively that is) of my tin fan shroud and duct; I made them with tin snips, a file, a drill and a punch riveter.

http://www3.sympatico.ca/merobbins/Watercool/01.JPG
Here's one that shows my pressure gauge, located at my pump's discharge.

http://www3.sympatico.ca/merobbins/Watercool/07.JPG
Here's a shot of the top of my reservoir, showing its clear acrylic top (which is cracked... I need to get some thicker plastic I think, as this is only 1/8"). I am thinking of lighting it with LEDs; we'll see how I decide to finish the project.

The next thing I am going to work on is finishing up the case. I still need to build a small power supply for the fans (ok, right now there's only one, but I will need another to help cool the pump when the cover goes on... it's a hot little beast), and which includes a pass through plug for my pump so there's only one power cord going into the back of this thing. I am playing with the idea of an acrylic side panel too, though my girlfriend says that I'm in trouble if I start puting windows in things to show them off :cry:... in any case, I'll post a few shots once I get this whole thing completed.