RAD position change make a difference???

[chazz469]

I changed my setup in two ways because I wasn't happy with the temps I was getting.

The old way:

rad-->WB-->pump/Res-->

* RAD was sitting on top of the power supply inside the case. I was also using a coffee can as the res, and decided that I wanted something a little less ghetto.
*1x120mm Enermax fan exhausting the rad (with ghetto cardboard shroud) and 1x80mm fan blowing into the rad (again with ghetto cardboard shroud.)
*tap water with Water Wetter

The new way:

rad-->WB-->pump/res-->

* rad is now outside the case, and I am using an outdoor PVC electrical conjunction box as the res.
*Changed fans!!! same 1x120mm Enermax with same shroud, and a new 120mm Enermax (no shroud yet) - I am planning to get some galvanized aluminim to make the shroud(s) soon.
* Distilled water with Water Wetter.

Notes:

The new way is running hotter by about 6-8°C.

Why is this happening? I had thought that the PS would be heating up my radiator, thus causing the water to run warmer overall. Therefore, moving to the outside of the case would help it to cool more efficiently. I also thought that the resevoir wouldn't make that much difference because of the material. Bongs are made out of PVC right?? I knew that it was better to pull air thru the rad, as opposed to push it, but I thought that effect could be amplified by adding another fan, at the same direction of flow (pushing into --> sucking out of, i.e. the first 80mm fan). By adding the 120mm fan I thought that effect would only increase. The old way ran anywhere from 36 (at boot) to 41 (after about 2 hours or so of UT). Now I'm running @ about 44°C aftre about 2 hours of SETI. And it's not wanting to seem to settle below 42/43°C.

Now I am reading this from MBM, on an MSI KT3. I can accept that there are variation in temp readings, but I have a basis with the same board,and I would think that the variation would be the same - air cooling --> the oldway (1st H2O attempt) --> new way (2nd H2O attempt). Meaning that even if the temps aren't accurate, I should still see a decrease right?

What's making me hot??????

(Ambient is about 24-26°C - CompuNurse)
(PS is about 28-29°C - CompuNurse)

One final note:

I had bought an XP 1600+ AFTER installing the water-system. I'm half tempted to go back to air just to see what I get with the XP&HSF combo.

[Random Nonsense]

It could be lower air flow through the case.... The in socket thermister is highly sensitive to airflow.... now that you have reduced the amount of air being pulled away from the general area (no longer exhausting air throught the rad) you may be causing a sort of stagnation point around the socket, try "hanging" a 80mm fan inside the case to agitate the air around the socket area, see what this does to the reported temps.

[chazz469]

OK....I have an 80mm intake in the top-side of my case that I used to provide fresh air to the rad while it was in the case. So I tried putting the other side of my case back on, and the temp dropped from 41 to 39°C.

So I get what you're saying, and definitely appreciate the input, Random. But I think that there might be something else that is preventing this system (admittedly not the best) from reaching it's full potential.

[Random Nonsense]

What rad are you using exactly? you may just be not pushing enough air through it, the enermax 120mm isnt the most potent fan in the world from what i know.

[chazz469]

I'm using the Danger Den "Cooler Cube". From what I've read, here and on other forums, the enermax fans may not be the best, but they are relatively quiet and seem to have about 70-90 cfm (depending on rpms - currently running @ ~2050 rpms). For me, they are also the most available. It's hella easier to run to Fry's and pick up a couple of these than try to order something better online.

Maybe I should look into it a little more....

BTW...ambient temp is ~24-26°C, chipset is at 34°C.

[Random Nonsense]

Something is wrong there, using a home made aluminium block, and NO radiator i had an 11c delta at full load after about 8 hours. you should be running cooler!

What are the full specs of the watercooling system? block? pump? hose diameter etc

Also when you moved the rad, is it possible that the new hose routing is levering the block off the CPU slightly? Are there any kinks in the tubing? is the pump having to lift the water higher?

[chazz469]

OK...the system:

Maze3, Cooler Cube Rad, 3/8" ID silicon tubing, Eheim 1048 (128gph), dual Enermax 120mm on the rad @ ~2100 rpms according to MBM (max spec. is supposed to be 2300 rpm).

Here's a pic of the current setup:

http://members.cox.net/hard10/images/NewSetup-top.jpg

(pardon the aesthetics - its a work is progress :-D )

I did check the screws on the block, and there seemed to be a little slack in a couple of them, so maybe the hose is levering it a little, but I tightened them down with only a degree of difference (41 -> 40 w/ case off). I don't want to crank it down too much, as I don't have a shim yet. Since my case is vertical in alignment, in there anyway to measure that the block is plumb? What else could be wrong?

[Random Nonsense]

Hmmm just took a look using radiate, your temps actually seem to be about right for a C/W of .15

try putting the Rad and reservoir on a shelf at the same height as the CPU...... that would mean the pump isnt having to work at all, apart from to push water throught the pipes, no lifting of any sort.

[chazz469]

Originally posted by Random Nonsense
...using radiate, your temps actually seem to be about right for a C/W of .15

Radiate? What's that? And C/W of .15? Can you translate for the W/C noob?

Originally posted by Random Nonsense
try putting the Rad and reservoir on a shelf at the same height as the CPU...... that would mean the pump isnt having to work at all, apart from to push water throught the pipes, no lifting of any sort.

This is basically the way it was before. Old res/pump was sitting on the pc next to the one it was cool (about the same level as the CPU), and the RAD a little lower than now on the PS. Temps were still in the 38-42°C range (36 at boot - 37-38 idle - 40-42 at load.) I'll give that a shot and see if it helps.

Thanks for all of your input.

P.S. Will posts result on this tread instead of starting a new one....

[Random Nonsense]

Radiate is a program that tells you the heat output of a cpu, i set it to your specs (1680 mhz athlon XP, 1.9 volts, assuming you are set at 1.85, its probably nearer 1.9) and the temp with an ambient of 25 c and a C/W of 0.15 came to 39 C

C/W is kinda hard to explain but it is used to tell you how high above ambient your temp will be

your CPU is radiating 90.3 watts, multiply this by the C/W of the sink (or waterblock) and this gives you the number of degrees above ambient your CPU will run at! A dangerden Maze 3 is probably about the 0.15C/W mark in your set up

Comes out at 39 Celsius

scary, i guessed the 0.15 to start with and it came out spot on!!!

[The Spyder]

Hummmm......... try Rad- Waterblock-Pump resivor-
Your CW is HIGH- something is WRONG-
SPyder

[Random Nonsense]

Originally posted by The Spyder
Hummmm......... try Rad- Waterblock-Pump resivor-

He is, i think the temp is a little high but it is 90 watts of heat after all!

[robbie]

Try PUMP-WB-RAD-RES.
Rob

[Holst]

You need to buy a temperature sensor and cheak your water temperarure.
If its more than about 5*c over ambient then you need more fans and/or a better radiator.

If your water temp is ok then there are two other possibilitys.

1 - you have moved your rad to a more elevated position and thats making your pump work harder and reducing flow rate. If at all posible try to run your system in the same horizontal plane, that way you keep your flowrate high.
(I doubt you would loose 6*c from this alone)

You may also have just got a bad contact on your CPU, check you have no hozes moving the block and re-seat it if necicary.

I should realy take some of my own advice as my system is running at a rather high 47*c after 3 days at full load. It should be quite allot lower than that, although my water temperature is rather high cause I have a crappy homemade rad and very little fannage.

[I Am A Sponge]

I have no experience with water cooling at all. But from your picture it looks like your pump would be working a little too hard to pull the water up from your CPU. Top that off the turn from the res to the radiator looks a little tight, but that just could be a little trick on my eyes.

Anyway, as I said, no experience with it. Just.. agreeing with everyone else I suppose.

[chazz469]

The turn from the res to the rad IS a little tight, but not restricted. I thought about throwing an elbow in there, but thought that it might restrict the flow. And if the problem is with flow, because of relative elevation or whatever, then obviously I don't want to make the problem worse.

OK....to address the idea of °C/W....I wasn't really sure what that was, but I've done a little research, and I understand it a little better. But I don't understand how you calculate the heat wattage output of the CPU. Watts=amp*voltage, right? How do I know what amperage is going to my CPU? Right now the BIOS reports my Voltage at 1.8, which Random assumed that it was 1.9. I'm not afraid of the math, so if anyone can explain it......drop science!

[Random Nonsense]

A program called Radiate does all that for you, i think its on AOAfiles take a look

[chazz469]

I found and downloaded Radiate, but I didn't see any watercooling setups in the coolers.dat, and I didn't see the Athlon XP. So I went with the Athlon MP, and set the °C/w to .15 , like you said, and came out with what you got. Except that my bios is reporting the CPU voltage as 1.8 which should give me a temp of 37°C.

Also, I would think that the C/W would be calculated by the total surface area of the system, which with a VERY ROUGH calculation comes out to be, like .86, which doesn't make a whole lot of sense. I must be calculating wrong.

So, can you stick a compunurse in water? I would imagine that my water temp is pretty close to the CPU temp, since, isn't that the point? But I would like to test it.

I am thinking about getting a shim which will 1) allow me to crank down the nuts on the block and 2) set the stage for a pelt. Especially if this is all that I can expect from my MAZE3, as is.

[Holst]

I wouldnt recomend using a CPU shim.
I think they are totally unnecicary and more trouble than they are worth.

My system is running at 48*c with a 108watt CPU and an ambient air temperature of 21*c.

Delta T = 48-21 = 27

then CW = Delta T / wattage = 27/108 = 0.25

Which isnt bad at all.

Im reading from the XP internal diode so my temperatures read slightly higher than you might expect and my radiator is far from brilliant.

Ill upload the program to calculate wattage here.

[chazz469]

OK, so by your calculation==My ambient = 23/Current CPU = 37/Delta T = 37-23 = 14
CPU Wattage (according to your handy-dandy tool - n1 BTW)

Wattage = 65.2W :. CW = 14/65.2 = .21

Random had said that my setup should be a CW of .15, so something isn't working as efficiently as it should. If we solve for Delta T or Wattage using the CW of .15, I would have the following:

Delta T of 9.78 OR Wattage of 2.1W (Which I would think would be impossible. )

So that means that I should be able to shave at least 4 more °C off my current CPU temp, giving me a temp of 33°C@23°C ambient.

Radiate calculates the wattage at 124.5W (for my custom XP specs) which is almost 2x what the OC-tool says. For the MP it says 80.8W which should give me a temp of 35°C. Basically this is a lot of theory, and theory doesn't always reflect the practical application. But TO ME it does indicate that I could be doing something better.