My WW upgrade results
 

I upgraded from this homebrew block: (the one on the right)
http://prod.bsis.bellsouth.net/coDat...76P1010007.JPG


to the Whitewater (polytop). My pump is an eheim 1048 and I use a Geo Tracker heater core and rheovolted 120 mm Panaflo L1a.

At the speeds/Vcore listed in my sig, my load (cpuburn) temps dropped from 63C to 59C. Ambient is 27C. Running the fan at full 12V drops it 2C.

Temps are a bit lower Folding: 52C.


BTW, those temps seem high across the board. I don't think my diode reads correctly.

I know that idle temps aren't reliable but why is mine 47C??

Wow, 182 views and not one comment?

It is just your home brew block we know nothing about so the 'difference' means very little. If it were a known BillA tested block lets say, then we could oooh and ahhhh. I'm glad you like the block though :D

What's your water temp?

I don't have a waterproof thermocouple, but I can at least guesstimate by taping it to the copper outlet tubing of the heater core (and put insulation over it).

That'll work.

More data:
With the fan running at quiet speed:
cpu diode temp under cpu burn: 57 C
when folding: 52C
idle: 47C
true idle with via regedit hack: 43C

Water temp (measured by placing a diode on copper tubing in the system) is 34C under cpuburn.

Again, ambient air temp is 27. Case temp is 32C.

Turning the fan to 12V drops the temps above by 2C across the board.


diode temp at cpu default speed and voltage: 42C under cpu burn, 36C @ true idle


These temps still seem high for a white water/Eheim 1048/heater core, considering the water temps. That's a 6C gradient from air to water, and a 23C gradient from cpu to water. My springs are cranked down (7 turns actually) and there is a cured coat of AS1.

whats your ambient?

27C

From your water temp, your cooling solution is working ok (7 deg C gradient).

The problem is your CPU temp measurement and/or the AS application: you're only supposed to have a razor thin layer of the stuff on both the block and CPU core. If you've done it right, then you ought to look into a better temp measurement, or alternatively, just leave it alone: it's probably just fine.

I followed the AS application instructions. This tube is years old, but it didn't look like it had separated when I applied it.

This is the rig btw:
http://prod.bsis.bellsouth.net/coDat...tercooling.jpg

Any suggestions/comments appreciated.

Does anyone know where to get those really thin thermistors (that fit between the block and cpu)?

Yeah... the tubing could do with an overhaul ;P.

~ Boli

i hope your coolant of choice wasnt feces :eek:

Man whats wrong with your water !!! Never seen that brown dye before.....

That's what happens to tubing after using water wetter after a while. The stuff does the same thing to the coolant overflow tank in cars (the application that the product was designed for).

I had used some on a previous setup.

I'll stick to my neon green water.

I hope hyperlube doesn't make my coolant look like dog crap over time.

Update: I coated the exposed wires of one of my thermistors in jbweld to make it waterproof.

My direct water temps are actually 2C higher than I had measured before. That makes the cpu-water gradient 21C (still too high) and water-air 9C (not great either, but my single fan is rheovolted to near silence).

I'm going to check the WW for flatness, and probably lap it.

Yea........running my fingernail over my Dtek WW I could feel some rather obvious valleys, so as soon as my lapping kit arrives I'm gonna take it to 1500.

I pulled the block and the base is definitely not flat. Mild surface grooving aside, the block is convex along it's long axis. You can rock it back and forth almost a full millimeter. I've started lapping it with 230 grit, but it may not be coarse enough to take this out. No wonder my temps were in the crapper.

An update:

I loosened the allen screws on the poly top, 1/2 turn each and that took some tension off the base and flattened it out a bit. No, it doesn't leak.

I lapped the base with wet-dry paper, 230 to 600 grit. I mounted it with a fresh coat of artic alumina (a new tube). The whole job resulted in load temps 2C cooler.

I also replaced the tubing...no more brown. :p

I noticed that my water temps with the pump running but the PC/fans off are quite a bit over ambient. It may account for most of my water/air gradient. More data later.

Watch out with the tightening.....my first dtekww came with only one o-ring and leaked a lot the first day, but then completely stopped leaking for 3 days....might wanna tighten em up a little more.

More data of interest:
With the PC/fans off but the pump on my water temp is 5C above case temp, or a total of 8C above ambient air.

Turning the PC on and running folding@home only adds 1C to water temp. CPUburn adds one more degree C.


So, my rad is working, but can the 1048 be heating the water that much?

I'll check the water temp with the pump AND radiator fan running, but the PC off to know for sure.