Temperatures IDLE and Full difference: 10º and 20º (this one with OC) too much no?

[HitMaker]

I have this water cooling system:

-Danger Den D4
-Swiftech Storm G4
-DangerDen Maze4
-BlackIce Pro 2
-Tube 10mm (i broke one of the 1/2'' barbs :rolleyes: :p: , until i receive it back -new one-, then i will use 12mm tube ;)) --> i don´t know if i have speak clearly, my vocab is not good enought to speak of water cooling stuff, i am sorry :p:
-Fillport

So.. i picture here:
http://www.telefonica.net/web2/hitmaker20/rl_apa%F1o3.jpg

My weird temperatures:

Stock cpu Without OC:

IDLE : 32º
FULL LOAD: 42-43º

With OC 275x10 1.7V (i was just testing the proper operation of my RL system):
IDLE: 33º
Fullload: 52º
:confused:

And my aerogate shows (with the temperature reading thing, whats its name in english :( ?) me:. That "thing" is mounted touching the heatspreader (lateral part of the IHS)

Without OC: IDLE 33º FULL 36º
OC: IDLE 34º FULL 39º

The thing is if i touch the water block at full load its completely cold (not at all, but cold), here in Spain we have today 33-35º and my room is around 30º.

So, the IHS causes this Temperature "problem"? i think the swiftech is mounted properly (i used the manual)... maybe the IHS (heatspreader) isn´t on touch with the core at all no?

Any recommendation or advise???

[One_Hertz]

The big diiference between stock and oced temps means that your rad can not take that much heat load ;)

[Ancient_1]

I kind of agree with what One_Hertz said but to have that much difference between load and idle it really looks like you could have a bad mount.

As far as your readings from a probe it would be hard to get one placed that would actually show anything near the temps that the CPU actually operates at. If you could mount on the backside of the cpu you could be close but other than that a probes reading will be way off.

[CrimeDog]

i blame your ihs

if you wanna confirm it try storm, d4, rad and see what temps are like with good mount

[Special_K]

onboard temperature sensors are not very accurate so do not believe the temperature it says

your aerogate "probe" is better but is not at the hottest part of your block

120.2 radiator is enough for CPU and chipset - from picture, it looks like more airflow needed for radiator

[aMp]

If the temperature rise happens quickly at all, the problem isn't the radiator. Even without a rad it would take at least a few minutes to significantly heat up the volume of water in a typical 1/2" ID loop.

So, that leaves a poorly mounted block, poor IHS contact, or a terrible internal temp sensor. The latter seems least likely to me, since even when a sensor is way off, it's usually consistently off, i.e. it's not going to measure a 20C difference when the temp only jumped 3-4C.

I've got sort of the same problem. Idle, my H2O'd Venice reads ~27C; a sudden loading bumps the temp to ~36-38C almost instantly.

-aMp-

[nikhsub1]

A few things, first, it looks as if the tubing is really pulling on the CPU block, not good. 2, your rad is in a horrible place AND you are blowing INTO it with hot case air, reverse the fans at least. 3, I would bet you dont have enough mounting pressure on the CPU, you need to crank it down.

[Ancient_1]

your aerogate "probe" is better but is not at the hottest part of your block

A probe on the edge of the IHS which is in direct contact with the water block and is away from the heat source is better? The only thing I see that measuring is the temp of the waterblock at the edge of the IHS. The block itself should of absorbed most all the heat of the CPU by the time it would get to the edge of the IHS. The only way the IHS would be significantly warmer than the block would be if they were not in contact.

I agree that on die cpu readings are not accurate either but should give a closer reading than a probe on the edgr of the IHS.

I still think poor cantact with the block and IHS could be the problem, a bad mount is the only time I have seen greater than a 3~5° difference in idle to load temps.

[HitMaker]

I think there is no problem with the system, is running perfect. Maybe is the airflow how Special_K says, the radiadot es good enough for these system.

But maybe is the IHS, i think i will take it off next time i mount the 1/2'' barb on the storm G4, i will need disassemble the entire circuit..

For the IHS any advise?? use a small cutter and put it into the IHS just a little and border the IHS no?? (because of the bridges inside/under de IHS no?)..

Maybe the diode sensor is broken, but if it shows ~41º full load without OC i don´t understand how it can go so higher with the OC.. i don´t know.

To improve the airflaw what about placing the fans under the radiator and suck the aire outside the case?? they will work better??? i should also move the Virge 2mb card to an upper PCI port (my 6800GT is in the first aid room :stick: )..

I know the blocks don´t work as a heatstink, but how can the cpu block be so "cold" at full load with the OC? do you the separator of the socket (on the tope of the socket, not the one under) is thickness than swiftech expect?? (so maybe the pressure of the block is not enough powerfull no?)

I dont know, but for sure the circuit is mount good enoght :p: , i have made a few modifications before...

Thanks a LOT!

[HitMaker]

If the temperature rise happens quickly at all, the problem isn't the radiator. Even without a rad it would take at least a few minutes to significantly heat up the volume of water in a typical 1/2" ID loop.

So, that leaves a poorly mounted block, poor IHS contact, or a terrible internal temp sensor. The latter seems least likely to me, since even when a sensor is way off, it's usually consistently off, i.e. it's not going to measure a 20C difference when the temp only jumped 3-4C.

I've got sort of the same problem. Idle, my H2O'd Venice reads ~27C; a sudden loading bumps the temp to ~36-38C almost instantly.

-aMp-

Thats what i was thinking.. the IHS

A few things, first, it looks as if the tubing is really pulling on the CPU block, not good. 2, your rad is in a horrible place AND you are blowing INTO it with hot case air, reverse the fans at least. 3, I would bet you dont have enough mounting pressure on the CPU, you need to crank it down.

I know the fans have not the better conditions to work.. but i don´t think they will make the temperature decrease more than a few ºC. By the way i stayed with the case opened and i got just 1ºc less ;)

About number1, what do you mean with pulling on??? i don´t really understand it. If you meant what i think yo do, i am using two "elbow" so the tube is has no problem there ;) and a D4 with 3.2 of pressure should be able to handle that.

At point 3 is my second though, but i follow the instructions carefully, swiftech should know that some mobos could have a different separator (thickness)

I will check all of that, do you recommend me to mount the block without the top socket's separator?

[One_Hertz]

I know the blocks don´t work as a heatstink, but how can the cpu block be so "cold" at full load with the OC?

The waterblcok will always be cold because it has relatively cool water running through it and the heat from the processor never gets to the top. It goes to the water which is then cooled by the rad.

[HitMaker]

The waterblcok will always be cold because it has relatively cool water running through it and the heat from the processor never gets to the top. It goes to the water which is then cooled by the rad.

Yeah i know, but i had a zalman WB2 and at least it got warm at full load...

Ok... maybe tomorrow i take off the top seperator of the socket and mount again the cpu block and see... but i think i should try to take off the IHS and maintain the separator and see how it goes, if it doesn´t get better then i will take definitively the separator off..

My thoughts are on the iHS.. but also is just the bloque/ihs contact.. in fact i don´t know.

[Ancient_1]

Also if you do take it apart it might be a good idea to check and make sure your coolant was clean enough to not of plugged up the jets and cups with debris from your coolant (that is the only drawback I see for a storm in general use).

[HitMaker]

Also if you do take it apart it might be a good idea to check and make sure your coolant was clean enough to not of plugged up the jets and cups with debris from your coolant (that is the only drawback I see for a storm in general use).

mmm i have filled the circuit 3 times already (i had to dissamble the system twice) so i guess its clean enough ;). I saw black sediments from the radiator the first time i drained the circuit ;).

more opinions?

[Ancient_1]

Did you check the jets and cups to be sure they are not clogged each time?

Not trying to be critical, just trying to help with a potential problem (it could be the inaccuracy of the on-die diode but I dont think so). The thing that makes me feel there is a problem is the large idle to load temp difference. What the actual numbers are to me dont mean much since two of my systems (motherboards) with the same cpu and cooling read about 17° diffent from each other and I believe the one that reads highest reads the closest to actual.

[One_Hertz]

well actually when you think about it at your OC settings you are pumping out 1.84 times the heat in watts then the stock processor by itself... Its only natural for your delta T to rise a similar amount and a little extra because of the higher water temp...

[heater918]

About number1, what do you mean with pulling on??? i don´t really understand it. If you meant what i think yo do, i am using two "elbow" so the tube is has no problem there and a D4 with 3.2 of pressure should be able to handle that.

From your pic. it appears inlet tube to cpu block is pulling on the block, could be causing uneven pressure. From pic. looks like elbows go from cpu outlet to maze.