PDA

View Full Version : Crazy Temps!


Spinout113
07-28-04, 10:22 AM
Hey everyone!
Thanks for all your help! Now I have my swiftech setup up and running pretty smooth. I had 2 rads, a cpu block and a gpu block, a swiftech fill and bleed and a pump. My question is that my temps are higher w/ water then I had when I was using a sp-94 w/ a enermax fan. Using an abit mobo (temps are off by 10c), I idle at around 50c and load at bout 62c (so 40c and 52c real temps). I have the setup pump-> fill and bleed - rad> cpu -> rad -> gpu -> pump. I am using arctic silver 5 on the cpu. My specs are below..any ideas?? Thanks!


-Jason

Abit IC7-Max 3 Mobo (bios 15)
P4 3.0C running a 3.4Ghz (fsb 227, vcore 1.65v)
1gb ocz pc4000 ram (2.8v)

Kaso
07-28-04, 10:47 AM
Hows the block sitting, are any of the pipes exerting pressure on it, they can sometimes tilt the block of the cpu

Spinout113
07-28-04, 11:28 AM
The cpu block is seated fine. It is really hot inside the case. The air inside the case and the rads are very warm. I have 1 intake 80 on my hds, a blowhole 80m w/ a rad pushing air out of the case and a rear 80 pulling are in the case. Is this enough?? What can I do about it?

Guderian
07-28-04, 11:41 AM
Replacing the fill and bleed system with a t-line might help flow a little.

sucking cool air through the rads will lower the loop temp, but raise the case temp. and vice versa.

you might want the rear fan as an exhaust, so your not sucking hot psu air back in the case. get air flow going front to back; bottom to top.

ZachM
07-28-04, 11:45 AM
Did you try remounting the block? You may think it's sitting fine, but it's really not. Make sure you use an even, paper thin layer of AS5

There could be bubbles trapped in the block. Take the block off the CPU, and move it to a low point in the system. Shake it, and that will work out any trapped air.

mumrah
07-28-04, 12:17 PM
AS5 takes a while to fully take effect. that might be a part of it.

Graystar
07-28-04, 12:35 PM
BIOS 16 claims to provide a calibrated CPU temperature.

nartac
07-28-04, 12:43 PM
BIOS 16 claims to provide a calibrated CPU temperature.

it claims to but doesn't. Bios 1.6 is not as stable as 1.5. Additionally, IC7 series temp reports about 10-13 degrees celcius above actual. Read (http://www.bleedinedge.com/reviews/abit_asus_temps/abit_asus_temps_01.html)

jenkx
07-28-04, 01:18 PM
The fill and bleed system is not a problem with flow. Make sure your pump is working and make sure you blow the hot air out of the case. A normal silent fan will do. Just keep the hot air out. Your system setup sounds a lot like my prior system. Hot air inside of the case was an issue for me until I got a good fan to push it out. The water gets hot and the heat dissipates through the tubs and into the case.

Those temps you have sound like your pump is not moving water. Make sure there are no hidden kinks or other restrictions. In my current rig I had it running an shut it down to do further work and in the process unknowingly disconnected my pumps. When I turned it on and did a temp reading idol was at 56c. I was like what the hell!!! I was checking the seating of the block. I turned the fans on high and since there is no air in the tubs I can’t tell when the water is moving and when its not. I then felt the pumps and they were not vibrating and it was then that I figured out what the problem was.

With that fill and bleed system make sure you opened the middle valve when you close the other two!!!!!

Graystar
07-28-04, 02:03 PM
it claims to but doesn't. Bios 1.6 is not as stable as 1.5. Additionally, IC7 series temp reports about 10-13 degrees celcius above actual. Read (http://www.bleedinedge.com/reviews/abit_asus_temps/abit_asus_temps_01.html)

Hmmmm. Interesting. After looking at that article I think the Abit readings are probably close to accurate. Only real way to tell is to read the CPU diode but I never seen or heard of anyone making a diode reader for Intel.

nikhsub1
07-28-04, 02:29 PM
it claims to but doesn't. Bios 1.6 is not as stable as 1.5. Additionally, IC7 series temp reports about 10-13 degrees celcius above actual. Read (http://www.bleedinedge.com/reviews/abit_asus_temps/abit_asus_temps_01.html)

Oh man, don't get me started... it is garbage like that 'test' that perpetuates the misinformation. He is measuring temps OF THE HSF, not the CORE OF THE CPU! His temp readings have as much in common with actual CORE TEMPS as my arm pit. The IC7 series of mobos read the ON DIE THERMISTOR in the CPU, Asus does not. Asus uses an insocket thermistor which is exactly what this bozo is reading in the HSF. All this stupid test poves is that the 2 secondary heat paths that are being measured (his silly probe) are close. Please see THIS (http://www.bleedinedge.com/forum/showpost.php?p=28448&postcount=6) post by someone else who knows. In conclusion, this test shows in no way shape or form that Abit reads high or that Asus reads low.

nikhsub1
07-28-04, 02:31 PM
Hmmmm. Interesting. After looking at that article I think the Abit readings are probably close to accurate. Only real way to tell is to read the CPU diode but I never seen or heard of anyone making a diode reader for Intel.
The Abit boards DO read the CPU diode which is why they report higher, and IMO much more toward reality.

Spinout113
07-28-04, 03:49 PM
I feel the tubes and see a few small bubbles moving , so the pump is working....for the blowhole...should i have it pull in or push out...thanks!

DumpALump
07-28-04, 03:49 PM
nikhsub1 is right. I hate when everyone says abit temps are 10C too high. Where the hell do they get this info from? Unless they can measure the temps exactly at the core then they have 0 basis for their statement.

Graystar
07-28-04, 04:00 PM
The Abit boards DO read the CPU diode which is why they report higher, and IMO much more toward reality.Yeah, that's what I figured. I didn't make this clear but what I meant was that an external reader would be best to determine the temperature. The board could still be reading the diode incorrectly, and it would be nice to take it out of the equation.

jenkx
07-28-04, 05:46 PM
Funny how we always manage to change the subject of a post. Have the blowhole blow air out.

Graystar
07-28-04, 07:07 PM
Funny how we always manage to change the subject of a post. Have the blowhole blow air out.
Quite right. I second Guderian's and jenkx's suggestions. You must have outside air go through the rad first. You don't want rear fans blowing in because there's PSU exhaust in the rear (hot air). So air should go in from the bottom/front and out through the top/rear. That's why you see a lot of rads in the bottom-front of the case.

Spinout113
07-28-04, 07:37 PM
Unfortuantly, I cannot have the rads on the front due to hard drives being there and having no other place to put them. I reseated the cpu w/ little effect. I idled at 48c
:( I am using Fluid XP+ and a blue uv dye if that makes any difference. I changed the rear fan to blow air outside the case. Something else is causing a boot up temp of 48c because w/ the case open and in the middle of the room w/ AC goin freezin i still start at 48c even after an hour cool off time w/ the system off. Any other suggestions. BTW, I have a 9700 Pro and my cpu in the loop, no NB. Thanks for your help!!

Spinout113
07-28-04, 08:08 PM
Here is a pic

Graystar
07-28-04, 08:09 PM
External CPU diode readers have demonstrated that CPUs heat up nearly instantaneously. It makes sense that you'll never see a lower temp...even from a cold boot. By the time you get into the BIOS to see the temperature, the CPU is well heated. So I wouldn’t worry about that.

Oh, I would definitely change the routing. Go from the pump output to the first rad, then the second rad, then the CPU, then the GPU, then back to the pump. Have the rad fans blowing in...it's not the best but better than blowing out.

nikhsub1
07-28-04, 08:13 PM
I cant tell by the pic, but have both fans blowing INTO the case, do NOT suck hot case air through the rads.

Spinout113
07-28-04, 08:27 PM
Well...what can I change w/o moding my case
Would changing the routing really help that much?? Thanks!

z0n3
07-28-04, 08:29 PM
Why did you use fluid xp and a dye? the dye makes it able to conduct electricity, unless you know where to get a non-conductive dye. plz tell me if you do.

Spinout113
07-28-04, 08:35 PM
I was unaware that the dye was conductive. ANyways, Should I be using something else? Could the fluid XP cause this high temps?

z0n3
07-28-04, 08:37 PM
No its not from the fluid xp. Try some of the solutions stated above and then tell us what happens. Also if you want the water to change color with out additives use coolsleves.

jenkx
07-28-04, 08:44 PM
Turn the back rad on its side if you have space and put a fan under it. If the case has a fan on the front reverse it to blow air out as the rad fans blow air in.

z0n3
07-28-04, 08:46 PM
In your pic, it may just be my eyes ,but it looks like you have a kink where the tube connects to the cpu block.

Guderian
07-28-04, 09:44 PM
Unfortuantly, I cannot have the rads on the front due to hard drives being there and having no other place to put them

are they behind the pump? cause it looks like you could get one of em below the 3.5 bay if you push the pump out. If this is the case I would move the one near the PSU.

Spinout113
07-28-04, 09:45 PM
It was a kink but I used cool sleeves to keep it unkinked lol. So...will chaging the air flow from back to front really help out my temps. What should I aim for as temps? If i decide to rebuild the system, should I run the rads in parallel? ANy other advice will help!

Spinout113
07-28-04, 09:46 PM
If I should do anything, what is the order that I should try things in? Thanks!
This high of temp seems to be much worse then rad\fan placement. Is there anything else big that I can look for? The one thing that sticks out to me is the parallel\series rads.

Graystar
07-28-04, 09:55 PM
If I should do anything, what is the order that I should try things in? TYhanks!
Try the easiest thing first.

Spinout113
07-29-04, 11:13 AM
I have already tried changing the fan directions and remounting the cpu block twice. Should I try unhooking the vga block and considering taking it out of the loop? Would the next step be rebuilding the system? Thanks!

-Jason

jenkx
07-29-04, 12:11 PM
Swiftech claims a 5c-7c drop in temps with two 80mm rads in parallel. If your CPU is idling at 50c a 7c drop would put you 43c which is more respectable.

The fact that Abit boards report higher temps than Asus boards, however each arives at its temp, should not be overlooked. Chances are you’re within acceptable limits. I had an Abit IS7 on my first watercooled rig and it reported super high temps after I replaced the Heat sink fan with a waterblock. I think removing the CPU cooling fan may cause the board to panic in a since and report higher temps. Try pluging a fan into the CPU spot on the board and see if it makes a difference.

Spinout113
07-29-04, 09:01 PM
Within the next few weeks, I will def try moving to parallel. Anything else big that I can try? I have a fan plugged into cpu fan power, and it does nothing (theoretically, it shouldn't do anything but you never know ;) )