Runnning warm on water?

[madocks]

Hello chaps,

Have just converted pc from air to water.

Am running a 1700 at default levels (no clock atm)

Current temps are around 50 celcius, had lower temps when on air

Spec of cooling below.

Pump
One submerged pump chucking 450 litres per hour thru the system

Tubing
All 15mm inner dia plastic sergical stuff, no kinks

Reservoir
Home made res, currently a pop bottle till i can get into work and do a better job

Water block
Home made job, made out of 10mm thick aliminium, with a 2mm coper lid.

Drawing of both can be found, http://homepage.ntlworld.com/bibblem/block.JPG

Radiator
This came from the heater matrix of a mini metro, has 2 fans bolted on and is doing a cracking job, also fits well in case



The problem I am having is that my cpu temp idles around 50 celius on stock voltage, with default FSB and multis. Scared to clock it at all until i can stabalise the temp. On air, this cpu runs about 40 celcius on default settings.

Any suggestions to reduce the temp? usefull info below;

Water in res is cool to start off with, but gradually get warmer in res

No kinks in piping.

Water block is cool (This may be a tip????), water block is calmped to processor pretty tight. So i dont think seating is a problem, or could it be?

Tried to get all air out of system, any tips?


Come on chaps help me out

[Shogun]

404 on the block picture. Make sure the block is seated on the proc level. As tilting it even a little bit would reduce the contact alot. What thermal paste you using?

[megah0]

that pump isn't putting out very much at all.

It could be that there is too much resistance in the system for effective flow.

[madocks]

404 on the block picture. Make sure the block is seated on the proc level. As tilting it even a little bit would reduce the contact alot. What thermal paste you using?

Try now, fixed picture :-)

So the block not sitting correctly would explain the cpu temp problems, but what about my water problems, its heating up , just made a shroud for the fans, am monitoring it for water temp.

[madocks]

that pump isn't putting out very much at all.

It could be that there is too much resistance in the system for effective flow.

Well i looked at the Eheim 1048 pump which chucks out 600 litres an hour, so figured my 450 aint that far off , or am i wrong?

[GAteKeeper]

I have a home made system as well

I first had a heater matrix out of an audi 100 it leaked so I took it to bits and found it full of gunk! This was seriously affecting the performance because all of the water was flowing down 1-2 pipes. You might have the same problem.

This is my custom block:


CAD image1
CAD image2
CAD image2
(Sorry for the CAD images not had a chance to upload the real ones)

I get ~35C under heavy load and 1.85V.

Hydor L30 pump
Custom made rad - same as a ThermoChill 120.3 rad only cheeper
T-Line filling
3 YS-tech 120mm fans @7v sucking through the rad

GateKeeper

[madocks]

I have a home made system as well

I first had a heater matrix out of an audi 100 it leaked so I took it to bits and found it full of gunk! This was seriously affecting the performance because all of the water was flowing down 1-2 pipes. You might have the same problem.


GateKeeper

Cheers gatekeeper, i dont think i have that problem, the water comming out of the rad is nice and clean and seems to be matching the speed of the water comming in.

However, if saying that the water is getting clogged up inside the rad, wouldnt that mean that the water spends more time inside the rad, and gets cooled better?

My water is still increasing

[GAteKeeper]

If you water is only getting hotter then it is gaining heat and not loosing it. The major place it looses heat is the radiator (the whole point of having one!) and so ipso facto that must be the problem.

A radiator works by exposing a large surface area to some cooler atmosphere, if only a portion of the area is getting heated by the water then only a portion can loose that heat into the air - this will affect your performance! The paths though the rads are usually alot bigger in cross sectional area than the tubing that supplying the water so you probably wont notice the loss of paths through the rad on flow (flow in must = flow out for obvious reasons!). All in all the water is speeding through a small portion of you rad and not loosing much heat.

My radiator that was full of crud also passed apparently clean water. I think there are places you can take car radiators and heater maxtricies to be flushed with the right chemicals to clean them out properly. You might at least investigate the cost of doing this as there will be some crud in you rad.

GAteKeeper

[turkster]

What are the ambient temps of the air? It can make a big difference.
Also try reseating the heat sink with new TIM if you can.
When you say the water is heating up what temp is it aprox?

The fact that your pump is only 450l per hour does not neccesarily mean it wont cool as well as a 600+ l per hour system. The actual flow you get depends on the pressure drop (aka resistance) across the system. Your piping is more than good enough, your rad shouldnt be impeding the flow too much either. In any case quite a small flow of water is capable of moving a good bit of heat as long as you can get good enough heat transfer. The problem is that in the case of water it is very difficult to get adequate heat transfer at low flow rates because water is not a good conductor of heat (the reason it is so good in heat exchangers is that it has a high specific heat capacity so can absorb a lot of heat). So to get good heat transfer you need turbulent flow in the block and to a lesser extent in the radiator. This is possibly your problem ie at the relatively low flow rates you have your block and the radiator are too easy flowing to allow adequate heat transfer.

I have a system running on a 600lhr pump through a lot of nasty 8mm and 12mm piping and a passive radiator and I am getting 27c idle and 33c under load at the moment and the block around 25c-30c. This is on a barton 2500 @ 2.3ghz. My block is home made and by no means smooth but it does a very good job. The water in my system is usually very close to the temp of the block.

IT does seem very strange that your system cannot keep a stock clocked xp1700 under 45c

[SilentDeath]

The problem seems to be your rad and fan.
First measure the temp of the water (use your fingers if you have too...) and compare it to the body of the rad.
If the rad is very close, then you are getting good transfer and its the fan thats not doing its job - Make sure it is sucking air from the rad, as this works best and that no air can get round the inlet to aviod the rad.
What is the rad made out of? copper or brass (or a mixture) is best.
Check the rad for buildups and corrrosion, if the rad is alu there could be a lot of corrosion blocking up some of it..

To get the same temps as your prev. air cooling, you will need ~the same air flow or maybe more.

Your block doesnt seem to be restrictive which is good.
How thick is the bottom of the channels to the bottom of the block (base thickness)?
This will need to be 4-8mm at a guess, becuase you are using alu.

The core of a 1700+, iirc is 13 X 7mm, which isnt much. Your block the cooling area is spaced over the whole socket (ideal if you ever want to use a 50*50mm TEC!)

You can check the block effectiveness with the same method above, temp of the water coming out, and the temp of the water going in, this is dT. You then divide this by the wattage you are putting into the block. 1700+'s (IIRC!!!) are rated 50w MAX at stock speeds. You should use a value 80% of this max value, as in real use the CPu will never reach 100% load.

I think pump ratings should be using l/min, or l/s. The l/hour measurement is not good at portraying how much flow is acctually moving. Also you have to consider the pressure curve of your pump. At no load it may be able to pump 450l/hour, but what about with 10 meters of head? (a rather extreme example, some pumps would catch fire with this ) you will get almost no flow.
Use the bucket test to measure the flow rate you are getting. this isnt very accurate becuase the water going out will not go back into the res, and the energy in it will be lost.
A way around this is to use an inline flow meter - this is something I must get for my system.

I also have a home made system, with a diy gpu, nb and hdd block. I have yet to get a CPU block but this is a very early design (lots of changes since) which I might be getting CNC produced very soon. http://server6.uploadit.org/files/kbn2k3-cpuwb2.jpg

GAteKeeper your block has interesting design, does it use a heatpipe in the center, or is that the inlet? I cant understand why the "fins" are above the base, does it reduce thermal transfer? It looks good though! have you got any pics of it finnished, with the top?

madocks for res I use a criticool 6" one, but I would suggest using one of these boxes to base your own from. Its made of polycarb, so shouldnt crack to easly (although the thickness at some points is ~1mm). It should be quite easy to seal - it is designed to be water/splash proof anyway. Its imo the ideal size for a res. I will be using one for another project on my psu, does not involve water though
http://www.rapidelectronics.co.uk/rk...6463&XPAGENO=2

[megah0]

At stock a Tbred 1700 produces 49.35W of energy.

info taken from http://www.benchtest.com/calc.html (A very useful resource for calculating the wattage produced by overclocked CPUs)

My over clocked 2500 produces 92.36W which is dissiapated nicely by my system

www.mgh0.modded.me.uk/WC.htm

By using the °C/W calculation further down the page it performs admirably with a °C/W value of 0.097444781290602 (The lower the better)

[SilentDeath]

I am fairly satisfied with my beta program `cpuburn` (`burnP6`) http://users.ev1.net/~redelm. I measure about 88% of the datasheet power draw through an inductive pickup.

I was curious if the 88% figure was still accurate and checked out his site. I found that he is now recommending the use of a program called runprio to change the priority of running burnp6 to high. He claims an extra 5°C is generated in this mode of operation. I wrote Robert and asked what he thought the percent of power draw was when using runprio in conjunction with burnp6. He answered that it was probably not more than 90%.

Your c/w value will be wrong if you did not take this into account. Also without using the thermal diode to measure temps, they will be atleast 5c out aswell.

[madocks]

Ok, info as requested...........

Water temp on boot = 22 degrees C. After 1 hour of running pc, water temp reaches 26.

Air temp is around 20 degrees, it is 1 am, and the window is open

With regards to the radiator. Its of copper and Brass design, with 4 x 80mm fans attached to it. 2 Of these fans are blowing air onto the rad, and 2 of the fans are on the other side sucking air.

My reservoir at the moment is a bucket, will be improved, bu need to sort ot my main problem first.

Have tried reseating the chip, however cpu temps have shot up to 60 degrees idle. Its gotta be a seating problem. Perhaps i need to redesign my block, ive got a 12 hr slot on a milling machine on Saturday may have a crack at a new one. Have also found a nice chunk of copper .

The blocks base thickness, from water to cpu chip is 1mm.

I am using a thermal probe built into my case for the cpu temp (Thermaltake), with its digi readout.

To measure the water temp i have pinched my water gauge from my fish tank poor fishies.

[SilentDeath]

Your dT between air and water is ok, the problem must be with the mounting, block, or pump. However I suggest using the fans like this as it will make less niose and push more air:
|>>
|>>
Where | = rad, > = fan
Instead of what your probably using,
>|>
>|>
Axial fans are crap when trying to push into restricive objects, like rads

How are you mounting your block? are you using springs or just 4 bolts and nuts?
With springs you get more pressure holding the block to cpu, this squashes out the TIM layer and gives better thermal contact. I cant find anywhere that sells springs that are suitable (at a resonable price), if you know anywhere tell me please

The base thickness for your current block will prevent much heat from reaching outisde the core area. If you make another one form copper, 1.5-5mm is best, depending on how concentrate your cooling is. The dd TDX for example the cooling area is very small, and so has a thin base plate.

Changing the res, so water can freely flow into the pump, will improve flow rates as pump energy is not wasted sucking water in, and vaporising (becuase of lack of water / pressure) the coolant.

[StarkMjolk]

Axial fans are crap when trying to push into restricive objects, like rads

Not entierly true I previously had 4 120mm 230V Fans specified for 24W each to play around with, all built in steel. Quite heavy things.

Tested two of them through my Black Ice radiator which has room for two 120mm fans on each side. The radiator wasn't even noticeable, the air just flew right thorugh it.

Then they did sound as an airplane and sucked more electicity than my VapoChill, but that's a different story

It's true for normal computer fans though, my Enermax fans had a really hard time moving air through that radiator.

[SilentDeath]

Similar to my fan?

Mine is about 13.2w (24v) and is probably about the same as your airplanes in niose I run it at 12v atm to keep it quiet but 9v once I finnish my fan/pump controler.

This one fan has no problem sucking through TWO large heatercores even on min speed (8.5v)
Not tested it at pushing, but i doubt it would have any trouble, I stuck a pencil in the blades for a test, and they didnt slow down :O