Watercooling - temps to high?

[Cappu]

I just hooked up my watercooling gear. It's been lying around for somtime, and I finally got time to install it. It's a DD Maze 2 with 3/8" fittings mounted with Artic Silver 3 a DD Cooling Cube also 3/8" fittings with 1 low noise papst 120mm fan and an old 120mm Delta fan on each site and a Eheim 1048 pump.

I have some "trouble" I would like some answers to.
First, my max load temps is way to high i think.
I mounted a Senfu temp probe on the site of my t-bird core.
It says 34c in idle and 41c in full load with Toast, 38c with CPU stability cpu warming.
MBM says above 55c. Ambient is just around 20c
Isn't that crap temps considered, I'm running watercooling?
Before I intalled my watercooling i was running the good old Taisol 74xxxxx alu-cooler. With this cooler i could oc my t-bird to 1333mhz at defalut vcore. Now with watercooling, I get the EXCACTLY same oc, and it doesn't help to raise the vcore to 1,85v.
Really hope some of you can help me




My system is as follows:
ABIT KT7A - RAID
T-Bird 1000@1333 default vcore
384mb Apacer PC133@CL2
Hercules GF 2 Pro
SB Live! 5.1 Player
Adaptec 2904 SCSI card
3com 10/100 nic
IBM 40 GB 60 GXP (two of those suckers died on me! So no raid for me until rma)
Plextor 40x speed SCSI cd-rom drive

Enermax 431 watt psu
All stuffed in to a Addtronics 7896A with 2*120mm papst in the front, 1*80mm papst above the cpu, and one 92mm papst in the back.

[eX&TriC]

same problem here ...
found no sollution yet

[Brad]

Your rad isn't very good, it doesn't cool well and reduces flow a lot, and your fans mightn't be moving enough air. The block is fine though.

The other thing to remember is that you have no fan blowing onto the cpu area now, and your mbm temps will be quite a bit higher with no air there

[Cappu]

Ye, I know the rad isn't the best, but it shouldn't be bad either?
I bought it when it was the only thing I could get here in Denmark.
I read about the onboard temp probe probs, and that's why I mounted the Senfu directly on the core.
But.... I would still like to know if those temps, aren't to high?

[Brad]

nahh, they aren't really high, at the same time I don't think they are too bad, you might want to remount the block, and things like that

[Cappu]

Ok....
Btw, I'm using demineralized water (70% )and antifreeze(30%) as fluid.

[kemical]

I had (still have) the same prob... But you should find out what your water temp is under load, and what your ambient temps are(in and out of case) as well. This will tell you if its a mounting problem or a radiator/flow/block problem... For instance, my ambient room/case temp is around 22C, my water temp under load and idle goes between 24-27C before and after my radiator(very good) but my CPU load temp goes upwards of 40C (measured by my DD5) or more or less depending on what sort of over-clock im going for, my idle is around 32C. 40C isnt great, but what you should notice is that my water temp hasnt moved more than 3 degrees while my cpu temp is going from 8-12C. This tells me that the heat isnt being moved effeciently from the cpu to my water. So, by remounting my block correctly I should see a rise in water temps and a drop in my avg CPU temp under load. Ive remounted a couple of times and have gotten quite a few different temps, none that im happy with yet, but I think you get the idea. So, I would try and remount your block a few more times to rule out that issue, and lap it (if you havnt already) before you look at your water system (which looks like a good one to me in my noobish opinion). Once your water temps start to rise you can measure how effective your radiator/fan setup is.

After comparing my on board temp sensors (MBM) with my DD5, I know that the MBM doesnt give accurate readings. So I would get a few more probes.

But im no pro and dont have much experience (see: "Hot Stank"), which is why my temps suck - so you dont have to take my advice.

-kem

xp1500 @ 1.59ghz | Iwill xp333-r rev2.1 | 2X256 Kingston pc2700 | Radeon 8500 retail | 431W Enermax Whisper | 2x 40gig Maxtor 133's in RAID 0 |

**Cooling** | DD Maze2-1 | DD Chipset Block | 1/2" ID hose and fittings | ehiem 1250 | BIX shrouded to a 172mm 200+cfm Rotron Monster | DD5 | Custom inline Res |

[Fixittt]

If it were me, I would do 2 things, get a shim. A good quality shim to garantee that your block is flat. U may have to modify it, or the block for the temp probe. I would rather U madoify the block but cutting a small channel into the bottom for the probe to sit into. I have also found that some of them slim probes, if you put them in to far, will cause the block to lift on that end. So check that. Also a push/pull fan config on the cude might help, more air thruethe rad is always a good thing. Also a good lap, and very very very very very (U get the idea) thin layer of thermo compound will also help. Remember the coumpound is not there to cover the whole chip, but just to fill in mirco gaps between the 2 surfaces. many many people get that wrong. Also, with the shim installed ( I highly recomend this) U can crank down a bit more preassure on the block, there for cranking the block to the core. What type of pump U using, for that cube style U need a beefy pump. remember, the Cube style rad and the maze 2 both reduce flow rates. So U need to compensate. If you are running the pump in the res, and upgrade to a bigger pump, then the bigger pump will produce more heat, so keep it out of the res. Also something else to remember about the maze blocks, there are alot of 90 degree bends, and alot of places for air to get trapped. So beofre mounting the block to the board. Run the cooling system and rotate the block several times, smacking it on its sides, top and bottom. I literally bang them around to insure all air is out. I hope this is of some help. Another note, I would mount the cude if U can so that the water flow goes into the top of the rad, and feeds downward thrue it, this may help improve flow. As it is not trying to lift the water thrue all the bends of the tubing.

[Cappu]

Well... first of all thanks for the great answers!

I measure my temps with a senfu probe directly on the site of the core, and also with the onboard sensor via mbm. The difference is up to 12 degrees!.

My radiator is mounted above my PSU in my Addtronics 7896A, with a duct going from the top. So the case temp shold not influence my rad.

I have mounted the Maze 2 with Artic Silver 3 following the instructions on Artic Silvers site. Wich means a very thin layer on the core and rubbing the block with Artic Silver before mounting to fill all the small gabs. I havn't lapped though.

I did run the cooling setup for aprox. 24 hours before mounting it. Banging the rad. and the block to realease as much air as I could.

But I think I'll try to meassure my water temps and ambient temps to see if it's a mounting prob or anything else.
But how should i meassure the water temps? Is it ok to just stick a probe on the copper tubing on the rad?

[kemical]

I used a couple of "T's" just before and after the radiator, I placed a probe in each one so that just the head of the probe would be exposed to the water flow. I put a thin coat of silicon on the probe itself to protect it from corrosion or anything else I put in the water... Not entirely sure if it was necessary as the end of the probes dont appear to be conductive and also appear to be ceramic... A thin coat should not affect my temps by anything worth worrying about, as long as both of them are the same. Once the probes were in place I filled the rest of the hole with lots and lots of marine GOOP. Made sure there were no air bubbles and that it wasnt oozing out into the water path. This seems to work ok. Im guessing youve already filled your system, so you probably dont feel like draining it to put in any T's. You could do the same thing by putting a probe into a res or fill tube if you have one. Im not an expert, but I dont think that measuring the temp by placing a probe on the radiator itself will give you an accurate water temperature because the air from the fans, the temperature of the inside of the case and the thermal properties of the radiator itself will give you too many variables to give you a good reading... Anyone try it?

The shim is a good idea... I think im going to try it myself, or look into a different mounting method, the nylon screws worry me.

200 dollars for the next generation main board,

300 dollars on the finest radiators, pumps, and waterblocks...

Watching your CPU smoke because the 2cent nylon fastener broke.... priceless...