Overclock.net banner

Can you achieve ambient temps with enough rads?

546 views 27 replies 10 participants last post by  christoph 
#1 ·
As the title states.

5930k 1.33v gets over 70c on real bench. 360mm rade and 240mm rad on cpu with external pumping too gpys in sli.

Temps coming inside can get down too 8c. However the temps are so high on the cpu.

Starting too annoy me with the cash ive spent.

Cpu block reseated. Second pump added too the loop no change in temps.

Seriously considering buying 4x480 rads bolting them together too see if i can lower the temps down too something sane.

Reckon it will work?
 
#2 ·
You have plenty of rad. You need to maybe look at flow problem or maybe just not a great block on the cpu. What cpu block?
 
#3 ·
I haven't done my Watercooling yet, but you should try adding another 360mm rad, is this all one waterloop, or is it dual loops? If adding a second rad doesn't help, you have the second pump, try to make a separate loop for the GPUs, so they won't interfere with CPU temps, like what temps do you get when playing games? Isn't real bench just a synthetic test that is supposed to stress the CPU to the max?
 
#5 ·
Its the ek evo supremacy. The temps quoted are on the hottest core. The rest sit around 64-65.

Setup is d5-cpu-360x60-external 240x30rad-condenser-gpus-second d5-280x60rad-res.
Quote:
Originally Posted by Moparman View Post

You have plenty of rad. You need to maybe look at flow problem or maybe just not a great block on the cpu. What cpu block?
 
#6 ·
What size tubing are you using? What order is your loop in? What model fans in what configuration?

Don't want to overwhelm you with questions but the more we know the more we can tell you
 
#7 ·
Are all the rads moving ambient air over them , Is your flow really strong, and are you using a tim like Coolermaster liquidmetal ultra? Adding more rads will probably not make much difference. At 1.33v 70c on a synthetic benchmark is VERY good. Not really sure why you're upset.
 
#9 ·
Quote:
Originally Posted by Dimebagg View Post

The temps are the close using intel burn test and aida64. When gaming like gta5 i get around 41-45c and 21-28c on the 980tis.

Do the ek blocks come preassembled to fitted right away? Because i pulled it out of the box and plonked it straight on.
Yes, they come good to go. You should be very happy with those temps.
 
#10 ·
Quote:
Originally Posted by kl6mk6 View Post

Are all the rads moving ambient air over them , Is your flow really strong, and are you using a tim like Coolermaster liquidmetal ultra? Adding more rads will probably not make much difference. At 1.33v 70c on a synthetic benchmark is VERY good. Not really sure why you're upset.
The outside rads (under house) have outside ambient on fans. Temps coming after cpu load ov 70c sit at 19-24 (no load on gpu) the 2 pumps in question are a pwm d5 at max speed and a d5 vario on setting 5. The tim used is the gelid extreme ( can this go bad after time?)
 
#11 ·
Quote:
Originally Posted by LmG View Post

What size tubing are you using? What order is your loop in? What model fans in what configuration?

Don't want to overwhelm you with questions but the more we know the more we can tell you
The pipework changes to get the stuff under the house. So internloop is 15mm. External pipework changes too 10mm air line with push/lock fittings with elbows and 4 valves inline.

So with temps going into the cpu of around 20c or ambient i shouldnt expect with even more rads too lower 70c i see on benches anymore? When it was resl frosty a few days ago my ambient was just over 1c and the temps coming ontothe gpus from the external was 8c. With an input of 60c from the cpu.

How does one make the temps all the same? A car radiator?
 
#12 ·
Quote:
Originally Posted by Dimebagg View Post

The outside rads (under house) have outside ambient on fans. Temps coming after cpu load ov 70c sit at 19-24 (no load on gpu) the 2 pumps in question are a pwm d5 at max speed and a d5 vario on setting 5. The tim used is the gelid extreme ( can this go bad after time?)
I've heard of tim drying out, but its not likely. You could try and use CLU and get a couple degrees cooler, but you're pumping alot of wattage out of that chip and the water can only wick it away so fast. I found that adding more rads to my setup added restriction and raised my temps a very small ammount. If you want ambient temps look into LN2 or dice, cause it wont happen on water.
 
#13 ·
Quote:
Originally Posted by Dimebagg View Post

The pipework changes to get the stuff under the house. So internloop is 15mm. External pipework changes too 10mm air line with push/lock fittings with elbows and 4 valves inline.

So with temps going into the cpu of around 20c or ambient i shouldnt expect with even more rads too lower 70c i see on benches anymore? When it was resl frosty a few days ago my ambient was just over 1c and the temps coming ontothe gpus from the external was 8c. With an input of 60c from the cpu.

How does one make the temps all the same? A car radiator?
Well damn. I didn't realize this was an externally radiated setup, that is out of my league.

Just curious, did you already post this setup somewhere else? Should I have seen it somehow? I don't see any info on your setup, but I'm just getting back into being active on OCN
 
#14 ·
Alrite so the crux of the concersation is achieving ambient isnt possible on water. Adding a monster external setup with 4x480mm may not lower my temps either. Is it limited by how fast i remove the heat from the cpu?

Ie flow?
Quote:
Originally Posted by kl6mk6 View Post

I've heard of tim drying out, but its not likely. You could try and use CLU and get a couple degrees cooler, but you're pumping alot of wattage out of that chip and the water can only wick it away so fast. I found that adding more rads to my setup added restriction and raised my temps a very small ammount. If you want ambient temps look into LN2 or dice, cause it wont happen on water.
 
#16 ·
You cannot achieve sub-ambient temps without somehow exposing your heat wasters (rads) to sub-ambient temps. This may be with LN2, dry ice, or radiating externally to an environment with a lower temp. No matter what you have to prepare for condensation if you're going to try for sub-ambient cooling
 
#17 ·
Quote:
Originally Posted by Dimebagg View Post

Alrite so the crux of the concersation is achieving ambient isnt possible on water. Adding a monster external setup with 4x480mm may not lower my temps either. Is it limited by how fast i remove the heat from the cpu?

Ie flow?
Yeah. The cpu at a given voltage will be a certain constant temperature above the water cooling it. Like you said, if the water is 10c cooler, you bench 10c cooler. The thermal properties of the tim and metal transfer the heat at a given rate, as well as the water moving the heat away, and there is no way to tranfer and move 100% of that energy instantly. I hope that makes sense.
 
#18 ·
Maybe I missed it, did you measure your water temperature?

Basically you have delta T between CPU die and water (dT1), and between water and air (dT2).

To reduce dT2, you get more radiators, more fans, faster fans. E.g. if your dT2 = 20 C, getting twice as much rad space will bring it down to 10 C.

dT1 is tougher to change and is determined by the waterblock, thermal paste and most notably by whatever Intel puts between the die and the IHS.

Point is, you need to attack the right area, the one where you have problems.
 
#19 ·
Quote:
Originally Posted by LmG View Post

You cannot achieve sub-ambient temps without somehow exposing your heat wasters (rads) to sub-ambient temps. This may be with LN2, dry ice, or radiating externally to an environment with a lower temp. No matter what you have to prepare for condensation if you're going to try for sub-ambient cooling
I dont want sub ambient. All j want is for the cpu to better reflect the ambient temps of outside. Thats why i askef if mounting a huge external setup bigger than what have nowoutside give me the temps im looking for?
 
#20 ·
I have not measured the water temps. Only the difference between leaving the cpu and arriving at the gpus (gpus not loaded)

I guess the advice i need then is what area shouldvi attack. I am prepared to build a monster external if thats the issue. I added a second pump thinkung it was a flow and head issue. Temps didnt change. A swingbof 10c ambient was about 7-10c change in max temps on cpu.
 
#21 ·
Quote:
Originally Posted by cmpxchg8b View Post

Maybe I missed it, did you measure your water temperature?

Basically you have delta T between CPU die and water (dT1), and between water and air (dT2).

To reduce dT2, you get more radiators, more fans, faster fans. E.g. if your dT2 = 20 C, getting twice as much rad space will bring it down to 10 C.

dT1 is tougher to change and is determined by the waterblock, thermal paste and most notably by whatever Intel puts between the die and the IHS.

Point is, you need to attack the right area, the one where you have problems.
I tried too quote you its being funny. I have not measured temps of water only temp diffences on die between cpu and gpu. I have no idea if adding even more rads will lower my temps or not.
 
#22 ·
I'm pretty sure the basic answer to your original question is no, and that's not the point of basic water cooling anyways. basic water cooling is about lowering the temps of an overclocked component to the point of being effective. if your looking for ambient or sub-ambient temps then that's a whole other category.
 
#26 ·
Quote:
Originally Posted by Norz View Post

1. dig a 10feet hole in the backyard
2. buy a BIG truck radiator and throw down in that hole
3. connect the tubes from radiator to a Iwaki MD70R pump
4. fill the system with water + 10% glycol (car antifreeze)
5.plug into youre custom waterloop ....and smile... you got temps below ambient in youre loop
smile.gif
and this is why I asked about pressure;

isn't that pump way too much for the loop??

I'm buying a 400 gph water pump...
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top