new guy new wc system

[greenman100]

ok, to improve upon your setup

install a shroud for that fan, $1 for a tupperware shroud is fine

sell that waterblock, get a MCW6000/TDX/WW.

your temps are way off, W/C system is pretty good, but PLEASE put a shroud on that fan....

 

[Cyrix_2k]

ok, to improve upon your setup

install a shroud for that fan, $1 for a tupperware shroud is fine

sell that waterblock, get a MCW6000/TDX/WW.

your temps are way off, W/C system is pretty good, but PLEASE put a shroud on that fan....

agreed

You're KILLING your temps without a shroud...

 

[way2ez4u]

I just rebooted and went into the bios and it did read 18c.
I guess its right...
wow kind of cool. no pun there.

Does anyone know how to mount or were to mount a temp sensor for the cpu ?

Thanks
Bmo

I would get second opinion from a separate sensor. I wouldn't trust it until.

 

[Bmo]

OK I had some sheet AL so I made up a shrowd for the rad. After I was done cutting and bending the metal, I sealed the inside and dynamatted the out side. Mounted the fan and reinstalled...

By doing this my temps dropped 3-4c, of course I am not sure what there really are but reading 13c.

How much of a improvement will I see if I get a different CPU block?
I am have a friend that is a CNC guy make my GPU block for me so thats taking care of... But do you think that I should order a different one or maybe have one made like the "MCW6000/TDX/WW"

Thanks for the input
Bmo

 

[Westside05]

I don't see any point of replacing your block while you are having 13c..

I wish I could have those temps...=](only if sensor probe is working perfectly^^)

 

[greenman100]

OK I had some sheet AL so I made up a shrowd for the rad. After I was done cutting and bending the metal, I sealed the inside and dynamatted the out side. Mounted the fan and reinstalled...

By doing this my temps dropped 3-4c, of course I am not sure what there really are but reading 13c.

might get another 5 or 6 C from going to a better block

go read about the R-Type at procooling if you want a CNC block

 

[lclark2074]

Hi all my name is brian, and now I am a waterholic...

After lurking around reading alot and checking out all the cool pics here I wc my system.
I have a gigabyte 8IPE-1000 pro-G MB with a P4 2.8 800fsb running at 3.08. Right now I am only cooling my cpu. Until I get my gpu block in.

The main question is that at idle my temps on the cpu are 17-18c, after a game of UT2004 the temps raise up to 23-24c but cool down fast to idle temps in around 2-3 minutes. Is this normal temps? I know that everyone rig will be different... I am using MBM5 for the readings. I do have temp sensors from my hard drive coolers, But I don't know were I would but them to get a real reading from the cpu...

Well just looking for some input.

thanks
Bmo

65F –32 * .555 = 18.315C
70F -32 * .555 = 21.109C
18C * 1.8 + 32 = 64.4F
as you can see i think your bord neads to be cal

 

[greenman100]

65F –32 * .555 = 18.315C
70F -32 * .555 = 21.109C
18C * 1.8 + 32 = 64.4F
as you can see i think your bord neads to be cal

and I think you skipped the whole thread

 

[Cyrix_2k]

OK I had some sheet AL so I made up a shrowd for the rad. After I was done cutting and bending the metal, I sealed the inside and dynamatted the out side. Mounted the fan and reinstalled...

By doing this my temps dropped 3-4c, of course I am not sure what there really are but reading 13c.

How much of a improvement will I see if I get a different CPU block?
I am have a friend that is a CNC guy make my GPU block for me so thats taking care of... But do you think that I should order a different one or maybe have one made like the "MCW6000/TDX/WW"

Thanks for the input
Bmo

Nice job on the shroud

I'm in complete agreement of greenman100... I would recommend buying, not making, the Swiftech (www.swiftnets.com) MCW6002.

 

[[AK]Zip]

best watercooled system has a C/W of about .15

and that's a cascade SS, huge pump, huge rad, huge fans

so the CPU cannot be any cooler than 36C load, given his overclock

and I doubt he has a cascade SS, huge pump, huge rad, huge fans

so he's off by at least 11C, I'd guess closer to 20 or 25C

I have one better then that.

0.11290322580645161

room temp of 24*C full load of 38*C 1.78V 3.92GHz OC 124Watt CPU

-Alex-

 

[greenman100]

I have one better then that.

0.11290322580645161

room temp of 24*C full load of 38*C 1.78V 3.92GHz OC 124Watt CPU

-Alex-

measured by what?!

 

[Cyrix_2k]

I have one better then that.

0.11290322580645161

room temp of 24*C full load of 38*C 1.78V 3.92GHz OC 124Watt CPU

-Alex-

NOT POSSIBLE!!!!!!!!!!!!!!

 

[[AK]Zip]

My dad did all the measurements. He is an electronic engineer. He measured all the temps and did the math. He was soo surprized that he did it again and same results.

-Alex-

 

[nikhsub1]

I have one better then that.

0.11290322580645161

room temp of 24*C full load of 38*C 1.78V 3.92GHz OC 124Watt CPU

-Alex-

And that ASUS board you are getting your CPU temps from is likely to be reading 15-20C too LOW. Rerun your calc with a load of 55C and see what C/W you come up with. We don't doubt that your calculated C/W is 0.11290322580645161, the point is, your temp measurement is not accurate.

 

[rogerdugans]

Congrats on joining the ranks of water coolers, Bmo.

But I post to remind everyone of the errors inherent in computer temperature measurement and comparisons....

Motherboard sensor readings are notoriously undependable.

Variations in the sensor hardware itself throws things off.
Variations in the code used to interpret the reading and show a temperature exist between BIOS revisions and manufacturers.
Software that shows a temp in an operating system may also "adjust" the reading.

It is impossible to get an actual cpu temperature below ambient using standard water cooling alone, but it is NOT impossible to get a reading showing a temp like that.

Just don't get too caught up in the temperatures you are shown in bios or an operating system, folks.

Monitoring temps is a great way to check for improvement in a cooling system by comparing the old with the new and it is helpful for spotting problems in a system, but indicated temps are NOT the "holy grail".

The above doesn't apply to anyone who has spent thousands of dollars on specialized temperature monitoring hardware, of course, but it does apply to anyone using motherboard sensors or the simplistic devices that we often add as well.

 

[[AK]Zip]

And that ASUS board you are getting your CPU temps from is likely to be reading 15-20C too LOW. Rerun your calc with a load of 55C and see what C/W you come up with. We don't doubt that your calculated C/W is 0.11290322580645161, the point is, your temp measurement is not accurate.

I am not using my mombo to give me temps. My dad placed a flat thermal probe on the CPU. Between the CPU and WB with the sensor side facing the CPU. My motherboard is actually reading temps much lower then this sensor is. Trust me both my father and I know what we are doing when getting temps.

 

[nikhsub1]

I am not using my mombo to give me temps. My dad placed a flat thermal probe on the CPU. Between the CPU and WB with the sensor side facing the CPU. My motherboard is actually reading temps much lower then this sensor is. Trust me both my father and I know what we are doing when getting temps.

I beg to differ... first, you should NEVER put anything between the WB/HSF and the CPU, you will have terrible contact. Second, would you measure the temp of the inside of your oven with a temp probe mounted on the outside of the oven door? Would you measure the temp inside your refrigerator with a temp probe mounted on the outside of the fridge door? I hope you answered no. What you are doing with that external temp sensor is exactly that, you are measuring an external CPU temp that first has to get through the IHS... Your temp readings are useless.

 

[rogerdugans]

[AK]Zip- note that using an external thermal probe measures the surface temp of a cpu and is therefore going to give a reading considerably lower than a core temp reading: a cpu's core is much hotter than its surface.

Which brings up yet another example of how comparing temps between systems is pretty pointless: variation in WHERE temp is measured also adds big discrepancies.

I have two systems with external temp monitors- cheap ones: a Digidoc 5 and a Vantec Nexus

One system reads the socket thermistor, the other reads the on-die temp.

On-die temp system:
BIOS - 52c
Vantec - 36c

Socket therm system:
BIOS - 42.5c
Digidoc 5 - 31c

With an ambient temp of 24.6.

Which readings could I use to ACCURATELY compare temps?

The ambient one is the only one I would consider reliable.