WC PC temps seem very poor

[Happy]

Hey,

I have a Lian Li A70B, with three front 120mm intakes and a rear 1200mm exhaust so Airflow shouldn't be an issue.

However I'm currently idling:- (Loop order 1-3)

(3) Q6600-3.6GHz-1.45v - 44C-44C-35C-38C
(2) DFI X38 PWM - 45C
(1) DFI X38 NB-1.5v - 35C

These are in one loop being cooled by a PA120.3 with three 120mm Yates at full whack (1200rpm). The PA120.3 is top mounted with the fans drawing cold air into the case.

My load temps from folding at home and prime95 are :-

F@H
Q6600-3.6GHz-1.45v - 59C-59C-54C-53C (TJ Max 100)
DFI X38 PWM - 60C
DFI X38 NB-1.5v - 42C

Prime95
Q6600-3.6GHz-1.45v - 70C-70C-64C-65C (TJ Max 100)
DFI X38 PWM - 64C
DFI X38 NB-1.5v - 45C

Now I know it's summer, but surely those can't be right

The PC sits on my desk, so it has loads of cool air around it. I have 4 HDs running, along with a OC'd 8800GTX and 8800GT usually running at full load for folding (they max out at around 75C with 100% fan speed).

Any advice, I'm probably taking it to bits next week to see if I can improve on the temps.

Thanks

 

[twist3d0n3]

i've got the same case as you, but i've got an e8400 o/c'ed to 4GHz, with 2 4870's (840MHz core on both) (xfire) in the same loop, with 1 PA120.3 rad. Load temp for the cpu never gets about 55oC, and the GPU's never go about 45oC.

cpu voltage is at 1.42v btw.

That does seem a little high but then again you do have a quad... what waterblock do you have on the CPU? is it seated ok??

 

[Happy]

It's a D-tek Fusion V2 with quad nozzle, I'm pretty sure it's mounted properly, but I will find out next week.

You're putting far more heat into your loop than I am, and you're temps are great. Lets say my CPU block is correctly fitted, why then are my PWM and NB temps so high :/ I don't get it

 

[twist3d0n3]

are they in the same loop as the cpu?

i'd presume that cos you've got a decent cpu clock and upped Vcore/NB volts they're getting toasty, particularly under load, as the fets are having to feed more juice to the cpu etc, and cos of your clock the NB will get warm, but yeah i agree they do seem a bit high...

i've modded my mobo tray so that there's a hole behind the cpu and nb, which i reckon have helped a bit keep those temps down... maybe it's dremel time??

with my setup, the cpu gets a lot hotter than the gpu's for some reason, even tho i'd imagine they both would draw a lot of current under load and both get hot...

what pump do you have btw?

 

[Happy]

They are all in the same loop as the CPU, I have a DDC 18W.

The loop order is:-

Rad>Res (EK 150)>pump>NB>PWM>CPU>Rad

 

[twist3d0n3]

my loop is:

res-pump-rad-gpu1-cpu-gpu2-res

this might be playing a part then in that your loop order is slightly different, from what i've read in the past you should put the rad before the compontents to be cooled, but i can't imagine this'll make a huge difference (maybe a degree or 2).

have you got 3/8th hose or 7/16? i'm using the latter... which if you're on smaller hose maybe that's another factor? but i think the combo of high environmental temps and voltage are going to be the main culprets, so i guess see if you can knock down the volts a tiny bit and see if it's stable. After i played around with GTL's (even tho they're not supposed to do much for duals) i could reduce my voltages by a small amount (better than nothin' i guess)

hope that helps a bit!

 

[Happy]

I using 7/16 as well. I've got 6ft more tubing coming next week, so it's getting destroyed and put back together bit by bit. Cheers for your input

 

[twist3d0n3]

just a thought, what thermal paste are you using? if you haven't got AS5, i'd highly recommend whacking it on all the blocks when you've got the thing in pieces

 

[Duke]

Seems a bit high, however with WC you'll find the idle and load temps are quite similar compared to air cooling which maybe cooler on idle but goes a lot higher on load and takes some time to cool down again.

 

[Happy]

just a thought, what thermal paste are you using? if you haven't got AS5, i'd highly recommend whacking it on all the blocks when you've got the thing in pieces

I've got some MX2, which is meant to be about the same as AS5 (although I'll be reapplying them all since the PWM and NB had blocks fitted already when I bought it second hand.

Seems a bit high, however with WC you'll find the idle and load temps are quite similar compared to air cooling which maybe cooler on idle but goes a lot higher on load and takes some time to cool down again.

Yeah that was one thing I have noticed the temp cooldown is far better on WC.

 

[shadowscotland]

Hi happy

It sound like the airflow is unballanced.
6x 120mm intakes and 1x 120mm + psu out ?

How are the GPU's cooled?

Firstly I'm guessing the psu is at the bottom as the PA120.3 in at the top.
If so thats good - is it also on it's own loop (in bottom out back) or does it take air from the case?

Most people have their rad as extract - as intake heats up the case.
as general rule with 25mm rad fans: slow speed = pull, high speed = push.

With psu at the bottom on own air loop.
Have all four case fans as intake and 3 rad fans as extract.
(but this may mess up the gpu coolers)

Another option is:
Intake at front and extract at back and top is another option
Better if your GPU are extract blower type.

Also black case in direct sunlight (on desk) will not help your temps

 

[Greebo]

Definitely too high. I had a cheapo 25mm dual rad with two xilence fans and my GTX260 and nb in the same loop and with my q6600 at 3.85Ghz with 1.5v it maxed out at 69C on prime. I would expect yours with less voltage and a superior cooling system and less items in the loop to perform much better than mine.

Since you have the rad top mounted, I would have the fans pushing air through the rad and expelling the hot air from the system. That's how I had mine.

 

[Happy]

Hi happy

It sound like the airflow is unballanced.
6x 120mm intakes and 1x 120mm + psu out ?

Yep Although I had the rad fans extracting air before, and the temps were better but still not great. Although when my tubing etc arrives they will be back extracing air so it will be 3x 120mm in and 4x120mm out + the PSU exhausting (PSU fan is facing upwards)

How are the GPU's cooled?

They are both aircooled, but when folding or gaming I run the fans at a 100%, so they don't get too hot.

Firstly I'm guessing the psu is at the bottom as the PA120.3 in at the top.
If so thats good - is it also on it's own loop (in bottom out back) or does it take air from the case?

It takes air from the case and exhausts it out back, since the fan is facing upwards towards the rad.

Most people have their rad as extract - as intake heats up the case.
as general rule with 25mm rad fans: slow speed = pull, high speed = push.

Ah ok, I have noticed that my internal case temps have risen to around 38/40C when CPU + GPUs are at 100%.

With psu at the bottom on own air loop.
Have all four case fans as intake and 3 rad fans as extract.
(but this may mess up the gpu coolers)

Hmm, with the way my PC sits on the desk, there isn't anything above the rads that would trap hot air in. However the back of the PC is only around 5" away from the wall, so if the back in intaking then it would be intaking all the hot air from the GPU exhaust.. might give it a go and experiment.

Another option is:
Intake at front and extract at back and top is another option
Better if your GPU are extract blower type.

I have a single slot cooler on the GT with dumps heat in the case, but the GTX extracts the air through the classic 2 slot cooler. I think that I'll swap the rad fans to be extracting then, as hopefully that should balance it out more.

Also black case in direct sunlight (on desk) will not help your temps

It's actually in the shade all the time luckily

Definitely too high. I had a cheapo 25mm dual rad with two xilence fans and my GTX260 and nb in the same loop and with my q6600 at 3.85Ghz with 1.5v it maxed out at 69C on prime. I would expect yours with less voltage and a superior cooling system and less items in the loop to perform much better than mine.

Dam that's a large difference I think my yates are near identical to the red wings, so thats' a good reference for me to hopefully beat.

Since you have the rad top mounted, I would have the fans pushing air through the rad and expelling the hot air from the system. That's how I had mine.

Thinking about it now, it does seem a bit silly to be bringing all the extra heat into the case, instead of exhausting it :/

This is the state of it at the moment - 1.6Ghz @ 1.16v, internal case temp 29C :/

What are you opinions on the amount of tubing in a loop? I'm thinking it won't really affect the temperature, but I have a large T-Line which I don't need that will be going, which should improve the flow rate.

Cheers guys, I'll report back with how I get on probably towards the end of the week

 

[shadowscotland]

Minimised the tubing lenght for the loop. - have a look at martin liquid lab flow estimator. you'll be amazed how much extra tube or a 90 barb will drop the flow rate.
Or the addition of a res top to the DDC will improve flow (which I guess is what you're doing if you loseing the t-line)

 

[w3bbo]

I think you have a mixture of problems tbh m8.

First off, MOSFET's and NB kick out a hell of a lot of heat, more so than the CPU. Do not be surprised that these are contributing to your high temps.

Flow will also be restricted by said blocks but I wouldn't think it was a massive problem for the DDC18w. Unfortunastely I can't view pics from this pc but you should always go for the shortest amount of tubing possible. That said I personally prefer looping tubing to using 90degree elbows which impact hugely on flow.

Also are you certain the TJMax of the Q6600 G0 is 100c? I thought that was for B3 stepping quads and that the G0 was 95c. Been a long time since I had one so I may well be wrong on that.

Summer temps contribute to your loop temps obviously and some people don't realise just how much. As an example a stable overclocked pc in the winter is not always stable in the summer - mainly due to an increase in temps.

Have you reseated your block? Might be worth a try.

Is the inlet on the cpu block in the centre and outlet at the corner?

You've probably tried all those but worth a mention I guess.

 

[Happy]

Minimised the tubing length for the loop. - have a look at martin liquid lab flow estimator. you'll be amazed how much extra tube or a 90 barb will drop the flow rate.
Or the addition of a res top to the DDC will improve flow (which I guess is what you're doing if you losing the t-line)

Since this was my first WC build, I was unsure on lots of stuff and learnt loads as I went along. I have a dedicated Ek 150 res, however I also installed a massive T-Line (maybe 40cm long ) to help with filling/draining (completely and utterly unnecessary in hindsight and probably a LARGE hindrance )

I think you have a mixture of problems tbh m8.

First off, MOSFET's and NB kick out a hell of a lot of heat, more so than the CPU. Do not be surprised that these are contributing to your high temps.

I know these kick out a lot of heat, but the thing that annoys/I don't understand is that Greebo for instance ran the following of a PA120.2 and was getting better temps than me, with more heat being out into the loop:-

Q6600-3.8Ghz-1.55v>NB 1.5>GTX260

On top of this a guy on CPC is getting a max of around 42C on the NB and around 45C from the default air cooled heatsinks on his DFI X48 (when running a OC'd Q6600 at similar levels to mine) :/

Flow will also be restricted by said blocks but I wouldn't think it was a massive problem for the DDC18w. Unfortunately I can't view pics from this pc but you should always go for the shortest amount of tubing possible. That said I personally prefer looping tubing to using 90degree elbows which impact hugely on flow.

I do seriously think my T-Line is affecting the flow rate, which is why it's being removed. Worth also noting that I use Feser's Distilled water and Petra Tech PTNuke, so I don't think it's clogging and therefore reducing flow etc.

Also are you certain the TJMax of the Q6600 G0 is 100c? I thought that was for B3 stepping quads and that the G0 was 95c. Been a long time since I had one so I may well be wrong on that.

I have spent so much time trying for a definitive answer to this, seems to be a 50/50 split as to whether it's 95/100. I have it set to 100 just to be safe. However 95 would seem much seem far more 'real' given my idle temps and my room ambient.

Summer temps contribute to your loop temps obviously and some people don't realise just how much. As an example a stable overclocked pc in the winter is not always stable in the summer - mainly due to an increase in temps.

I know my temps have/are being affected, but they weren't brilliant to start with. But due to this heat, it's becoming an issue I'd like to be able to understand/fix. My PC is in quiet a small room, although the door and windows are always open.

Have you reseated your block? Might be worth a try.

No, but I'm stripping it down this week and will be reseating all blocks.

Is the inlet on the cpu block in the centre and outlet at the corner?

Yep.

You've probably tried all those but worth a mention I guess.

Cheers for the advice

I just worry if I fall outside the norm for stuff, and this case (temps) I do

Thanks

 

[Bubo]

How much of your T line actually has flowing water in it? If its a normal T line then the | bit of the T will have stagnant water in it. Or are you talking about excessive tube routing to facilitate a T line in the first place?

 

[Greebo]

Hell my rad isn't even as good as PA120.2 and I still did better.

And my nb block was the default one on the Asus maximus formula se and they don't come any more restrictive than that! In fact that killed my cpu temps by 3-4C adding that into the loop.

 

[w3bbo]

You can't really compare temps like for like in different systems with different setups m8, I learnt that hard lesson long ago. For all you know he could be living in an igloo. His board is X48 yours is X38, no doubt there are improvements to the digital pwm which bring about different temps also. Are you using the same temp monitoring software (and the same revision)? There are just too many variations to compare temps to someone elses system m8 - don't worry what anyone else is getting.

Are you sure you have bled your system correctly? There could be trapped air somewhere.

Try the loop without the T-line and see if it makes a difference.

 

[Happy]

How much of your T line actually has flowing water in it? If its a normal T line then the | bit of the T will have stagnant water in it. Or are you talking about excessive tube routing to facilitate a T line in the first place?

Only the -'part' the '|' part is the bit about 40cm long and will be the stagnant part.

Hell my rad isn't even as good as PA120.2 and I still did better.

And my nb block was the default one on the Asus maximus formula se and they don't come any more restrictive than that! In fact that killed my cpu temps by 3-4C adding that into the loop.

I'm hoping that by the T-line I see some changes

You can't really compare temps like for like in different systems with different setups m8, I learnt that hard lesson long ago. For all you know he could be living in an igloo. His board is X48 yours is X38, no doubt there are improvements to the digital pwm which bring about different temps also. Are you using the same temp monitoring software (and the same revision)? There are just too many variations to compare temps to someone elses system m8 - don't worry what anyone else is getting.

Suppose that's true It's just kind of annoying :/

Are you sure you have bled your system correctly? There could be trapped air somewhere.

This is the only thing I'm 1000% sure about, the rad has been help at all angles etc, so now more air to come out

Try the loop without the T-line and see if it makes a difference.

I'm hoping this combined with swapping the rad fans so they extract will cool things down a bit.

Cheers