• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Wacky temps - just board problem?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

craneage2001

Member
Joined
Mar 14, 2002
Location
Wisconsin
With my current setup (see sig for details), my temps seem abnormally high. I've tried resetting my block three times so far and each time I remove it the thermal paste is beautifully spread out in an extremely thin layer as it should be. The water in my res is usually just a bit warm after folding all day. Before I reseated my block, I was getting load temps of 42°C. I reseated and I'm getting a load of 43°C with the increase in temps probably due to the reapplication of AS5...

Any thoughts on the subject? These high temps are aggravating, especially if they're incorrectly reported...
 
It says version 6.00PG...it's the newest one from the website (12/24/03). I'm actually hitting 44°C load right now...I'm guessing the AS5 is the reason for the slight increase. However, the temps should be about 10°C lower according to established CW ratios for the DBX.

EDIT: I flashed to the Preacher BIOS and I'm still hitting loads of 42°C...


Trogdor the BURNINATOR!
 
Last edited:
First off what is your overclock? If you are overclocking those temps sound about right. Temps are very hard to diagnoise or compare to others. You should be looking to see if you are actually able to over clock higher versus your air setup. Also where does your motherboard read temps, is it from the cpu diode or the cpu socket.

Also for example my setup is a black ice extreme, mcp600, mcw5000 water block and a bay res. I have it overclcoked at 220x10 with a vcore of 1.725. My temps according to the bios, which is an in socket temp, is 39 idle and 44 max load with prime95.

Good luck ;)
 
My board reads from the socket...my processor at current speeds/voltages puts out a bit over 90w of heat, whereas yours generates about 80w of heat. From most tests, the RBX gets a C/W ratio of around .10, which theoretically puts me at roughly 30°C. As my setup is relatively simple, I think my actual CW should be close to the tested ones...am I missing something? I've seen users running the A64s hitting loads of 35°C or so with the same setup...
 
craneage2001 said:
My board reads from the socket...my processor at current speeds/voltages puts out a bit over 90w of heat, whereas yours generates about 80w of heat. From most tests, the RBX gets a C/W ratio of around .10, which theoretically puts me at roughly 30�C. As my setup is relatively simple, I think my actual CW should be close to the tested ones...am I missing something? I've seen users running the A64s hitting loads of 35�C or so with the same setup...

hmm - lowest i've seen quoted for the rbx is 0.13.....

AS5 is supposed to get better over time for several days iirc - maybe you're not giving it enough time?

are you sure it's completely bled out...?
 
1) Even at .13CW that puts me at 33°C...

2) I ran my last seating over a month and the temps did drop just a hair over the first week, down to around 41 or so...

3) My system bleeds quite fast and yes it's clear of air...
 
craneage2001 said:
1) Even at .13CW that puts me at 33�C...

2) I ran my last seating over a month and the temps did drop just a hair over the first week, down to around 41 or so...

3) My system bleeds quite fast and yes it's clear of air...

hmm sounds like may be the NF7-S temps again (i have one too...) - possibly the block is better at cooling the chip than the thermal probe (since it doesn't read off the die)

you don't mention fans in your sig - i take it the ones on the heatercore are up to it... and that its not high air temp>high water temp>high cpu temp...

not familiar with the preacher BIOS myself...
 
I've got a cold-rolled steel shroud I folded and a 120mm fan on the rad...the water in the res (post cpu, pre rad) is lukewarm...I'm going pick up a suitable thermometer to get the exact temps of that water. However, being pre-rad, its going to be a bit warm. BTW, my rad fan is in a pull config, drawing fresh air from the outside of the case.
 
What is your room temp? Can you post a picture of your setup? Do you have good case airflow (hot air may be building up around the socket, giving false readings)?

The tested C/W of the block generally cannot be used in calculations the way you are trying. They are only useful for comparison to other blocks tested on the same test setup.
 
I don't have my digital cam on me at the moment, but here's the setup:

1) All of the narrow lines are 1/2"
2) The larger lines (to and from Mag3) are 3/4"
3) There are no sharp bends in the tubes (almost exactly like pic)
4) There is also a 80mm intake on my side panel that was there when I got the case...
5) The 120mm on the rad is pulling a good amount of air, easily able to hold paper up against the front of my case.
 

Attachments

  • waterflow.jpg
    waterflow.jpg
    37.3 KB · Views: 148
I don't think anything is wrong. You are running at 2530MHz @ 1.68v on a chip that is supposed to run at 1.83GHz @ 1.45v. You are talking about a large voltage increase along with a large overclock. If you were saying your temps were over 50C then you should be concerned. Remember that chip is rated up to 100c.

I think you are just being thrown by the published C/W ratings. They are there for comparison purposes only. Overclockers' waterblock tests all say "NOTE: These results are NOT comparable to tests done by others." at the bottom of the results. This is there for a reason.
 
One thing to note: The mobile Bartons are no different than the standard Bartons, aside from being "the cream of the crop" or "hand-picked" for lack of a better term. These chips are simply the Bartons that can run at 2500+ speeds at low voltages, making them good for mobile platforms. Due to this stability at low voltages, they overclock extremely well at higher, standard Barton vcores. A mobile Barton at 1.7v and 2500mHz will put out the same heat as a standard Barton at 1.7v and 2500mHz. It's just that only a few Bartons will actually run at 2500/1.7.

Also, I don't see any reason why tested CW ratios are irrelevant...given the number of watts of heat produced, a block can only remove so much at a time, whether it's a cpu core or other heat producer for testing. This amount of heat removed can be calculated by the temperature of the 'core'...With the tests (see below), the estimated GPM of the RBX/Mag3 combo was ~2GPM. Testing the RBX at 2GPM, it resulted in an approximate CW of .10 or so. Allowing for a large range from their tests to my results in the 'field', a CW of .13 gives at most 35°C. I realize that one cannot take the testing results as the gospel, but it would be reasonable to assume that the results would be at least close.


RBX / Mag3 Data
 
craneage2001 said:
One thing to note: The mobile Bartons are no different than the standard Bartons, aside from being "the cream of the crop" or "hand-picked" for lack of a better term. These chips are simply the Bartons that can run at 2500+ speeds at low voltages, making them good for mobile platforms. Due to this stability at low voltages, they overclock extremely well at higher, standard Barton vcores. A mobile Barton at 1.7v and 2500mHz will put out the same heat as a standard Barton at 1.7v and 2500mHz. It's just that only a few Bartons will actually run at 2500/1.7.

using the calculator here, i get just over 99Watts for a barton 2500+ at 2.5 & 1.7volts
http://www.benchtest.com/calc.html
 
I just noticed something...you don't know the water temperature going into the block, so how are you calculating the CPU temperature? Looking back at your calculations it looks as if you were using the air temperature. You can't use the air temperature, you would need to know the water temperature going into the block for the calculations to be even close to accurate.

Also, you said the water in the Res is LUKEWARM; lukewarm is a comparative measure to your body temperature (~98.6F/37C) so that water is pretty warm if you are saying it is lukewarm.

Assuming the 99W, 42c load temp, and a .13 C/W the water temperature is probably around 29C, which could be close to reality. The C/W is dependent on flow rate, so it is probably alittle off.
 
I'm away from the PC at the moment, so I'll measure water temps on Sunday. However, I can only measure them just after exiting the cpu block. How much of a temp difference do you think there is between the water entering the block and the water leaving the block?
 
craneage2001 said:
I'm away from the PC at the moment, so I'll measure water temps on Sunday. However, I can only measure them just after exiting the cpu block. How much of a temp difference do you think there is between the water entering the block and the water leaving the block?

not much at all - especially at 2 gpm - under a degree rings a bell...
 
Back