Mission: Figure out why my temps suck

[Raxxath]

Now that I've just about figured out all the design issues with my rebuild, its time to revisit my poor CPU temperatures. I have a few theories, but I haven't been able to pinpoint the cause yet. I'm hoping for a bit of help from the experts. First, I'll list all of the relevant information.

System specs:
EVGA 8800 Ultra
C2D E6700
Asus Blitz Formula
2x1GB Crucial Ballistix PC8500
10,000 RPM 150GB Raptor
Corsair HX620

Water-cooling:
D-Tek FuZion
Swiftech MCW60
PA120.3
DDC3.2 with Petra's top/ D-Tek dB-1(I have both)
Swiftech Micro-res
7/16" tubing on 1/2" barbs
Only distilled water at the moment

-All fans are Yate Loon Lows at ~800RPM, including the three on the radiator
-Ambient temperature is about 19C
-Everything is at stock clocks for the time being

The CPU cores idle at about 34C, which is simply way too high. The 8800 Ultra idles around 38C. I have Everest, Real Temp, CPU-Z, PC-Probe II, etc all confirming these temperatures are correct, no bad sensors or anything.

I am currently using just the washer in the FuZion, I've tried a few of the nozzles without much luck. I've tried raising the fan speed, there is little to no difference. I've remounted many times with varying degrees of pressure, my TIM application also seems to be correct. I've tried both the DDC3.2 and dB-1, with no difference between the two. I've even removed the video card and replaced it with a passive one, to see if the heat it generated was hurting the CPU. There was no difference in temperature after removing the Ultra.

The radiator does get warm to the touch, but not hot. The same with the air coming out of it.

I am using the Scythe backplate, but I took the cover off, so the sticky part is stuck to the motherboard. Could that have an effect?

I have bled everything as much as I possibly can. I wouldn't say its impossible there's still air inside the radiator, but if there is, I have no idea how to get it out. I have shaken, hit, and run the radiator and blocks in every orientation imaginable to get any air out.

In the rig I'm currently posting on, I run a stock E2180 with a Coolermaster GeminII and a 400RPM fan. It idles at about 23C. Sure, its a lower power processor, but not by that much. My few hundred dollar liquid cooling setup should not perform 10C worse than a 99 cent cooler.

I'm sure I've forgotten a few things(this is a lot to remember!), so I'll add stuff as I think of it. If there's anything else you guys need to know, ask away, and I can post pictures of anything you need. Thanks for any help.

The current testing setup
http://i38.photobucket.com/albums/e119/LordBrona/36e11758.jpg
http://i38.photobucket.com/albums/e119/LordBrona/1ccec376.jpg

TIM contact
http://i38.photobucket.com/albums/e119/LordBrona/Rigtesting015.jpg

[ZOMGVTEK]

Hmmm...

What are you running your processor at, and what vcore as reported by cpu-z?

And those 8800 ultras put off some serious heat... But your temps do sound a little high, not insanely high or anything though.

Any load temps?

[F1ZZY]

Those temps aren't disasterous. If you want better cooling, need faster fans or fans with more pressure.

What's the problem with 34 idle apart from your ego?

What's the load temps like?

[DavidNJ]

You are blowing VERY little air through the radiator...without a shroud?

Do you have a backplate, which reduces the pressure between the bottom plate and IHS?

It looks like the TIM could be thinner...look at some of the other pictures that show a high pressure area for examples...you could also just try a few mountings.

[Conumdrum]

Is there a really restrictive grill on the case part where the fans are? I have a E6600, FC 8800GT block, and a MCW30 on my NB, look at my sig. My ambient idle was a LOT lower than yours.

[Raxxath]

Hmmm...

What are you running your processor at, and what vcore as reported by cpu-z?

And those 8800 ultras put off some serious heat... But your temps do sound a little high, not insanely high or anything though.

Any load temps?

Like I said:

Everything is at stock clocks for the time being

I've even removed the video card and replaced it with a passive one, to see if the heat it generated was hurting the CPU. There was no difference in temperature after removing the Ultra.

And I couldn't give you accurate load temps at the moment, but I believe its in the mid 40s or so with dual Prime95 running. That's only after about five minutes though, not hours like some people test.


Those temps aren't disasterous. If you want better cooling, need faster fans or fans with more pressure.

What's the problem with 34 idle apart from your ego?

What's the load temps like?

Like I said, increasing fan speed does nothing. And what isn't wrong with 34 idle in this setup? This is at stock clocks, and a fairly cool ambient. Summer is coming soon, and I intend to overclock quite a bit. If I don't figure out what's causing this problem, I'll be in trouble. It makes no sense for such an expensive setup to be performing poorly. There's definitely something wrong here.


You are blowing VERY little air through the radiator...without a shroud?

Do you have a backplate, which reduces the pressure between the bottom plate and IHS?

It looks like the TIM could be thinner...look at some of the other pictures that show a high pressure area for examples...you could also just try a few mountings.

Again, like I said:

I am using the Scythe backplate, but I took the cover off, so the sticky part is stuck to the motherboard. Could that have an effect?

Also, there is still quite a bit of air coming through the radiator. Thermochills were built for low speed fans like this.


Is there a really restrictive grill on the case part where the fans are? I have a E6600, FC 8800GT block, and a MCW30 on my NB, look at my sig. My ambient idle was a LOT lower than yours.

Nope, just a normal wire grill, very unrestrictive. And you mean your idle temps are a lot lower than mine, not your ambient temperature, right?

[Ashraf]

You are blowing VERY little air through the radiator...without a shroud?

Really? Do you need a shroud for using low speed fans?

[Conumdrum]

My ambients were a bit higher and my idle temps were lower, and my load temps are a bit lower, lookies at my sig.

[ZOMGVTEK]

I skimmed the OP 3x+ to make sure i didnt miss anything, and it looks like i missed the 'stock clock' part... (I was looking for numbers)

Anyways, i get lower idle temps on my Q6600 at stock clocks with a TRUE, so id say you do have a possible issue... Maybie you just have bad contact with the core below the IHS on your chip?

[Gamekiller]

Anyways, i get lower idle temps on my Q6600 at stock clocks with a TRUE, so id say you do have a possible issue... Maybie you just have bad contact with the core below the IHS on your chip?

I get lower idle temps with my q6600 at idle with 1.504v 3.6ghz with a TRUE also. (But when is it ever idle lol WCG crunch on! :rocker: )

So yeah, big problem. I hope you figure it out because I'm thinking about going to water. Good luck.

[Raxxath]

I skimmed the OP 3x+ to make sure i didnt miss anything, and it looks like i missed the 'stock clock' part... (I was looking for numbers)

Anyways, i get lower idle temps on my Q6600 at stock clocks with a TRUE, so id say you do have a possible issue... Maybie you just have bad contact with the core below the IHS on your chip?

I don't think there's any question that something is amiss. I wouldn't think that something like that would account for such a big temperature discrepancy, but I'm not an expert.

[Raxxath]

Raxxath

switch your hoses on the rad

top= hot/in

bottom cool/out

So... flip the radiator over? I'm curious, why would that have any effect?

edit: The fans are pushing air through the radiator.

[Raxxath]

heat rises

Slowly. Too slowly to make any difference in computer cases that replace the air inside of them in a matter of seconds, or so I'm told.

I really don't think this is a question of the radiator's heat dissipating abilities. If it were, increasing the fan speed and removing the heat production of the GPU would have had some effect.

[SoulsCollective]

I'm inclined to say that there's nothing wrong with your setup, but that the DTS on the chip itself is borked, or that you're using the wrong Tjunction setting for the CPU. Did you ever run this setup under air to give some idea of what temps were like then?

[Ashraf]

So... flip the radiator over? I'm curious, why would that have any effect?

edit: The fans are pushing air through the radiator.

Pull should do the job.

[Raxxath]

I'm inclined to say that there's nothing wrong with your setup, but that the DTS on the chip itself is borked, or that you're using the wrong Tjunction setting for the CPU. Did you ever run this setup under air to give some idea of what temps were like then?

Sensor movement in Real Temp is within spec. TjMax is listed as 85C, which I believe is correct.

I only ran it on air with the crappy stock cooler. It idled at about 45C, with the fan at ~800RPM, stock settings.

[Jedda]

I use a scythe back plate too and have about double the amount of thread showing above your black nuts. I found pressure critical for good temps with this set up.
Need tp know ambient though. My ambient goes up a degree so does my cpu etc Good or bad is all about the delta between them.

[Raxxath]

I use a scythe back plate too and have about double the amount of thread showing above your black nuts. I found pressure critical for good temps with this set up.
Need tp know ambient though. My ambient goes up a degree so does my cpu etc Good or bad is all about the delta between them.

Already said my ambient is 19C. ;)

Can you describe how much force it takes to turn the nuts as tight as you have them? There's a lot of resistance when trying to turn mine, I don't want to damage anything.

Edit: Tightened the nuts down a bunch more, no difference in temps.

http://i38.photobucket.com/albums/e119/LordBrona/a8e94e69.jpg

[desertstalker]

Have you tried OCing the CPU? My E6600 idles at almost the same temps at stock and at 3.3GHz (1.45V).

However when at stock the load temp was ~15C lower. I suspect that the heat dump from the Ultra is causing poor idle temps as the lower the coolant - air delta the less heat transfer.

As an example, my system 2x 8800gts 640M + e6600 @ 1.45V currently room ambient is 16C water temp is 23C and CPU idles at ~28-30C (GPUs @35C). These idle temps are higher than was achieved by a thermalright 120Ultra but the load is much lower.

So... the heat load, even when idling (GPUs are under clocked to 128/300/256) is sufficient for a 7C air-water delta and the CPU water delta is ~6C.

If your load temps are in the 40s on stock then it will be fine, if you double the heat dump by the CPU you will NOT double the temps.

[Raxxath]

I suspect that the heat dump from the Ultra is causing poor idle temps as the lower the coolant - air delta the less heat transfer.


I'm getting the feeling that some of you didn't read my OP too well :p: . I took the Ultra out and tried a passive 2600 Pro instead, so there was no heat coming from anywhere but the CPU and pump. Temperatures did not change.

[F1ZZY]

You appear to focussed on reducing the idle temps at stock settings? Sorry but I fail to see the point in this.

You should be looking at load temps when overclocked as that WILL be a limiting factor.

Stock idling doesn't matter if it's 20, 30 or 40 or even 50 degrees. (as long as that then doesn't lead to high load temps)

Most of the changes you made should alter cooling capacity (diff pumps, fan speeds, no gfx card in loop etc) and you saw nothing. From that maybe you could say that the "bottleneck" is from the cpu die into the waterblock.

Lapping and better mounting might be the only thing to (pointlessly*) reduce the idle temps. (*but this would help load temps too)

[rogard]

try rotating the d-tek 90 degrees ccw

[Soulwind]

1) Have you taken the D-Tek apart and cleaned it? Mine had a VERY nasty amount of rubber extras from the O-Ring. Simply pulled it off of the O-Ring (carefully) and put everything back together again.

2) You could try rotating the block 90 degrees. Don't know if it would help or not though.

3) What's your LOAD temps look like? Honestly, idle temps really don't matter. My CPU idle's pretty high (about 32 [ambient + 10]) but only gains a few degrees at full (Orthos or Prime95 at 4+ hours) load.... (load temps are 38 [ambient + 16] to 40 [ambient + 18]).

[Giannis86]

i agree with F1zzy..in some cases idle temps can be slightly higher than a true for example..

try raising your vcore at stock speeds and stress load the system.. If your load temps are acceptable i wouldnt worry much about it..

a side note as well. if you plan to do some high overclocking you should add the asus fans (not necessary to WC using the fuzion system as it doesnt really improve performance)

[Conumdrum]

Umm and your gonna cool the NB pretty soon too I hope.

[kohlersc]

I would definitely open up and clean the Fuzion if you haven't already. As a previous poster said, when I opened mine up stock out of box, I had to clean off the excess rubber...just peeled it right off and it was fine. Some of it had flaked off already and was in the pins. If I had hooked it up right out of the box I probably would have been quite displeased.

The the other thought, just make doubly sure the thing is bled well...if there is a chance at an air bubble sitting in your Fuzion that may also account for it.

[vitriol_drinker]

I Get 45°C at idle with a Dtek Fusion on a e6750 @ 38000 1.46v,
2MCW60 on 2900XT's
and a Watercool HTF3 Triple

http://www.rsssim.com/images/Big/128017-035_BIG.jpg

With 3 1100Rpm 55 Cfm revoltec fans
:(
(Laing DDC + alphacool top, all 1/2 tubling)

[DavidNJ]

Cooling is really rather simple. You need to transfer heat from the source, the core, to the coolant, and from the coolant to the air.

So, what is your water temp? It should be a few degrees above ambient. If it is 10-15C above ambient, that is the first source of your problem.

Increasing air flow should have a dramatic impact on water temps, 200 to 500%. However, if it is from 4C to 2C above ambient, it won't be easy to see, as other factors can be larger, even variations in ambient temps.

A restriction in the airflow, for example being too close to a wall on inlet or exit, reduces the airflow.

Without a shroud, the fan housing is actually blocking part of the radiator. As a result, it acts like a smaller radiator. Upto 40% smaller when you include the hub and the frame. Some rads have the equivalent of a 1/4" shroud sort of built into the frame. We will have to wait for Martin's tests to judge different shroud depths; however, adding at least 1/2" should help. I add 1 1/4". Note that shrouds can be made from longer screws, spacers, and duct tape for just a few dollars and a few minutes.

The contact between the dies and the IHS varies, and is out of your control. It could be a problem, however that is highly unlikely.

The contact between the IHS and water block is important. you want pretty high contact pressure over the dies. A backplate significantly increases it; without it some of the tension in the mount is released flexing the motherboard. Some tests have shown 2-3C improvment.

Bowing reduces the contact area increasing the pressure over the it; total pressure was determined by the mount. Adding the washer does this on a Fuzion. When bowed, the contact area will look significantly different than non-contact area in the TIM.

TIM has thermal conductivity of around 9 W/mK vs. 400 W/mK for copper that is used in the water block and IHS. It should just fill voids between to two contact areas. There shouldn't be any clumps in the contact spots, especially when bowed. It looks like your layer may be too thick.

As noted by another poster, if there was debris in the system, it could be blocking part of the water flow.

Idle temps in the 30s aren't terrible. Are load temps in the 50s?



http://www.rsssim.com/images/Big/128017-035_BIG.jpg


Does this radiator have round tubes through the fins? If so, it is probably much less effective than radiators with thin flat tubes such as the HW Labs and Thermochills. T

he round tubes are typically used for systems with high temperature differentials in less critical systems where coast is a major issue. Examples would include automotive cooling systems for transmissions, power steering, and rear differentials.

The goal in those applications is just to keep temps within 80C of ambient. In hours we are trying to minimize the differential when the cooled fluid is maybe 2-5C above ambient, and every degree is important more or less.

[vitriol_drinker]

Yes round tubes in it.
Sadly, I have no way to mesure the Water Temp :(
I'm going To try to put the Scynthe ultra kaze 3k rpm to see if the temp will be better.
Edit : I put the 3 Ultra kaze @ 12v > 42-44°C

[C'DaleRider]

Personally, I'd try redoing my loop and see if that helps, if what the pics show is correct.

You're pumping from your DDC to the cpu, then video card, then rad, then res, then back to pump, right?

So you're adding the pump's heat dump, however small, into the water just before it hits the cpu.

I'd try pump to radiator then to cpu, gpu, res, pump.....maybe that'd help....?

[vitriol_drinker]

In fact my loop is
Rad>Pump>Gpu2>Gpu1>Cpu>Rad (I use a T and a Filliport, no res)

[F1ZZY]

I believe he was refering to OP 's post not yours vitriol, you've hijacked someone else's thread LOL

[Raxxath]

try rotating the d-tek 90 degrees ccw

What would that possibly do? I know that I already have it in the proper orientation.


1) Have you taken the D-Tek apart and cleaned it? Mine had a VERY nasty amount of rubber extras from the O-Ring. Simply pulled it off of the O-Ring (carefully) and put everything back together again.


The first thing I did when I got the block was take it apart and trim the o-ring very nicely.


Umm and your gonna cool the NB pretty soon too I hope.

Yes, I have two MCW30s ready, just trying to figure this out until I add them.



Bowing reduces the contact area increasing the pressure over the it; total pressure was determined by the mount. Adding the washer does this on a Fuzion. When bowed, the contact area will look significantly different than non-contact area in the TIM.

I'm not sure if that was all meant for me, or the poster above you, but adding the washer from the nozzle kit does not bow the block. Adding an o-ring does.


Personally, I'd try redoing my loop and see if that helps, if what the pics show is correct.

You're pumping from your DDC to the cpu, then video card, then rad, then res, then back to pump, right?

So you're adding the pump's heat dump, however small, into the water just before it hits the cpu.

I'd try pump to radiator then to cpu, gpu, res, pump.....maybe that'd help....?

First of all, I've redone the loop multiple times. Second, that's not how it works. The differences in heat within a loop tend to vary only a couple degrees.

[F1ZZY]

Seriously bro, stop fretting about 34 idle temps. This could lead to a load OC temp of <50, which isn't an issue.

Like I said, the limitation appears to be die to block (from the testing you have done). Lap and add more pressure to the mount is the only real way to get lower temps I believe.

If I was you i'd be patting myself on the back, overclocking and enjoying a fast and cool pc.

[vitriol_drinker]

I believe he was refering to OP 's post not yours vitriol, you've hijacked someone else's thread LOL

Oops, but in fact the problem is the same :)
(68° Full :()

[DavidNJ]

Oops, but in fact the problem is the same :)
(68° Full :()

I would just replace that radiator. A variety of $60 would be much better.

An o-ring is a type of washer, and bowing doesn't care if the washer is round or flat. However, it does seem that the D-tek washer causes no perceptable bowing. I will have to observe TIM patterns under clamping.

[WoZZeR999]

Have you tested the temp of the water? If you don't have an inline temp sensor, you can use one of those cheap temp gauges and put the probe in with the fins of the rad for a reliable temp.

[Trice]

I know with no AC here in San Diego it gets pretty warm in my apartment. I just redid my loop and idle 28C or so at night with cool air coming in, probably around 65*F and it has gone up to 38C when it gets to be 80+ in the room. I have my GTX in my loop as well and it idles 38-44 depending on ambient.

Your TIM looks kind of funny though, almost as its water based. Also is your rad warm to the touch? That will give you a good judge of if the dump is getting into the water or it its just a bad IHS contact.

[Raxxath]

Have you tested the temp of the water? If you don't have an inline temp sensor, you can use one of those cheap temp gauges and put the probe in with the fins of the rad for a reliable temp.

Nope, I'll see if I can find a thermometer of some sort.



Your TIM looks kind of funny though, almost as its water based. Also is your rad warm to the touch? That will give you a good judge of if the dump is getting into the water or it its just a bad IHS contact.

Its MX-2. I can say that while idle, the radiator stays quite cool. Under load, it does get a little warm to the touch, but not much.

I'm doing some load testing right now, I'll post my results in a bit.

[Trice]

Load test should give a better judgement as to whats going on as well. Try to get a GPU read also, and a load one.

[Raxxath]

Here are some results-

Idle
Core 0: 34
Core 1: 32
GPU: 36

10 minutes of Prime 95
Core 0: 42
Core 1: 40
GPU: 37

20 minutes of Prime 95
Core 0: 42
Core 1: 40
GPU: 37

Stopped, back to idle
Core 0: 36
Core 1: 33
GPU: 37

10 minutes of Prime 95 and ATITool artifact test
Core 0: 42-43
Core 1: 40-41
GPU: 42

Another 10 minutes of Prime 95 and ATITool, with fans on full speed
Core 0: 40-41
Core 1: 37-38
GPU: 40

The ambient temperature increased by about 3.5F throughout the testing, because of all the heat it dumps into my room.

During all this, the radiator got a little warm, but really not much. The bottom copper part of the FuZion that I can touch barely got warm at all. For reference, the GPU ramsinks will burn my finger while touching them for a few seconds.

There's a small difference when changing the fan speeds at load, it seems. I bet if I used a shroud and a pull configuration(which I've been considering for a while), that difference would mostly disappear. Raising fan speed permanently is not an option, and it doesn't have that much effect anyway, so don't tell me that's what I need to do. :p:

[WoZZeR999]

Have you tried anything other than MX-2? Maybe you got a bad batch that was missed during QC. At idle, your core temp and gpu temp should be pretty close, and it looks like the rad isn't the bottleneck (temp would raise more than 1c on the GPU if your radiator was stressed). I would try a re-mount with a difference tube of paste (still use MX-2, just a different production tube). Also try a different method of applying the paste.

[jbartlett323]

dude your temps barely move!!! i wouldn't be worried about that at all!! my q6600 idles about 38-42 on the cores and up too 65ish loaded... i have a small temp problem... you have a system begging to be overclocked!!!

[Raxxath]

Have you tried anything other than MX-2? Maybe you got a bad batch that was missed during QC. At idle, your core temp and gpu temp should be pretty close, and it looks like the rad isn't the bottleneck (temp would raise more than 1c on the GPU if your radiator was stressed). I would try a re-mount with a difference tube of paste (still use MX-2, just a different production tube). Also try a different method of applying the paste.

My thoughts exactly on the radiator.

I do have some Arctic Silver Ceramique from one of my heatsinks I could use, I suppose, though that problem sounds highly unlikely. I was planning on trying the line method instead of the dot method, too.

Can anyone else with a FuZion tell me if the copper baseplate gets warm to the touch?

[Kilyin]

Are you obsessive compulsive? There's nothing wrong with your watercooling. In fact, it seems to be working quite well.

[Raxxath]

Are you obsessive compulsive? There's nothing wrong with your watercooling. In fact, it seems to be working quite well.

A little bit, yes. And there is something wrong with my watercooling. Sure, its working better than air, but its not working close to as well it should be.

[WoZZeR999]

jbartlett323: You have a C2Q and it's oc'd, Raxx has a C2D at stock speeds.


Kilyin: That would be true, except that it's at stock speeds. Plus, people are allowed to want every last bit of performance out of their rigs.

Raxxath: I would try leaving prime on all night, logging the gpu temp, and see if it increases more than 1 degree. I would try overclocking that sucker and see what kind of temps you get out of it at around 3.6.

[Raxxath]

jbartlett323: You have a C2Q and it's oc'd, Raxx has a C2D at stock speeds.


Kilyin: That would be true, except that it's at stock speeds. Plus, people are allowed to want every last bit of performance out of their rigs.

Raxxath: I would try leaving prime on all night, logging the gpu temp, and see if it increases more than 1 degree. I would try overclocking that sucker and see what kind of temps you get out of it at around 3.6.

Thanks for backing me up. I don't feel like I'm crazy for expecting better performance than this.

I think the problem has more or less been pinpointed to the CPU block here, whether its the mount or otherwise. Something interesting to note is this thread (http://www.xtremesystems.org/forums/showthread.php?t=172141) of mine from a while back. Air trapped in the block could be the explanation for both the temperature and noise issues. But even so, I have literally shaken and hit the block in every way imaginable to dislodge any air, and nothing ever came out. I don't know what more I could do.

[WoZZeR999]

Some Proc's just run hot. Find out the water temp and it will be easier to figure out from there. You could also try running your proc at min specs (266*6, lowest volts to prime) and see if your processor is hitting the ambient temps. If you have it running and even with the minimum running speed it's above the water temp, then it's a block/mounting problem.

[Raxxath]

I've been taking everything apart today to get to the bottom of this. I tried a remount using the line method, temps were even worse.

Is there anything wrong with my FuZion here? It appears that the adhesive on the washer is coming out from beneath it, but I really doubt that's the cause of my problem.


http://i38.photobucket.com/albums/e119/LordBrona/3e54f51f.jpg

http://i38.photobucket.com/albums/e119/LordBrona/176424b3.jpg

[Soulwind]

Multiple folks have said that there's nothing wrong with your temps.

You continue to insist that there is.

You're running 32-34 idle and 40-42 load

So I guess the next question would be:

If those temps aren't satisfactory for you, then what temps DO you expect?

Maybe the issue is less one of temps and more one of expectations?

[Raxxath]

Multiple folks have said that there's nothing wrong with your temps.

You continue to insist that there is.

You're running 32-34 idle and 40-42 load

So I guess the next question would be:

If those temps aren't satisfactory for you, then what temps DO you expect?

Maybe the issue is less one of temps and more one of expectations?

You don't see a problem when a $1 heatsink is outperforming a few hundred dollar liquid cooling setup by 10C? There's nothing wrong with a 15C difference between my ambient temp and idle temps on a lone dual core processor with a radiator capable of dissipating hundreds of watts of heat? There is absolutely no question that something is wrong here.

[Raxxath]

yes your missing a seal between top an mid plate

You mean the washer? You can see it in my last picture.

[D_A]

Have you checked the CPU block surface for flatness? If it's concave you'll get really bad heat transfer but with enough TIM you may not see the difference on disassembly.
I'm just guessing, but it seems feasible.

[Soulwind]

Again you didn't actually answer the question, you just keep stating over and over that "it's wrong".

You're comparing air results of a 2160 vs water results on a 6700
-- 2160 = 200mhz FSB, 6700 = 333mhz FSB
(the 6700 runs about 2-3x hotter than the old 2160's to begin with)

Of course you're going to see warmer results on the hotter chip.

Now, if you really want to say that the wc isn't working right then try taking that $1 heat sink off the 2160 and put it on the 6700 (they are both the same socket) and see what kind of results you get.

Then you'll be comparing apples to apples. Of course, I'd be real careful, since I'd expect that $1 heat sink will keep a 6700 at about 75* or so . . .

Anyway, I do hope you resolve this issue to your satisfaction, but I'm done.

[Kilyin]

So you honestly believe that a cheap heatsink is keeping ANY cpu at 4 degrees above ambient? You probably run that chip in a different board too, I'd say you need to add 15 degrees to that temp reading and you might be getting close to accurate.

Why don't you overclock your CPU, put some more volts in it, and really see what the water setup is capable of?

Better yet, install the water loop on the 2180 system and then you'll truly see it's an erroneous reading (Or run the cheap heatsink on your newer CPU like Soulwind said, and compare that).

[NaeKuh]

Is the radiator top mounted?

if it is i think you got a gignormous bubble stuck inside your rad. turn your system off, and rebleed the system and give your case a tilt on each side see if air bubbles shoot out or anything.

[Martinm210]

Try a different TIM compound.

I just did some tests between Artic Ceramique and T-C Grease 0098 from www.tim-consultants.com and noticed about a 3 degree improvement over ceramique.

I also like coolaboratory liquid pro for long term mounts, but the TIM consultants grease is the best I've found for easy application.

I'm also one with a hot processor, not much you can do if you get a hot one, likely a poor soldering job between the cores and the IHS.

I don't think your temps are bad..

[Raxxath]

I took my other rig apart to get the e2180 out, and tried it under the FuZion. It was idling at 13C :p: . That was impossible, of course, so I did some more research, and it turns out the TjMax for the M0 stepping is really 100C, not 85C. Real Temp, Everest, and Speedfan are all incorrect in that respect. So it was really idling at 28C or so, about 6C less than my e6700.

Am I really expecting too much here? I just thought an entire PA120.3 dedicating to only a dual core processor would keep its temperatures closer to my ambient. I am aware that when I overclock, I'll still probably have safe temperatures. I simply expected more for my money.



Is the radiator top mounted?

if it is i think you got a gignormous bubble stuck inside your rad. turn your system off, and rebleed the system and give your case a tilt on each side see if air bubbles shoot out or anything.

While a bubble wouldn't really make sense, because my GPU temperatures were fluctuating so little, I redid the loop anyway with only the CPU block. I bled the radiator very thoroughly, until no air would come out.

Interestingly though, when I try the method explained on the Thermochill website(stand the radiator up straight, with the res higher, then open both to bleed), water simply leaks out of the radiator, while I know there's still air inside. Anyone know why that is?

[smee]

Raxxath, Do you have the FuZion Nozzel kit?

If not, I would order one ASAP, as this could solve your temp "issue". :up:

[Raxxath]

Raxxath, Do you have the FuZion Nozzel kit?

If not, I would order one ASAP, as this could solve your temp "issue". :up:

Yes, I have it, but it makes very little difference.

[NaeKuh]

Interestingly though, when I try the method explained on the Thermochill website(stand the radiator up straight, with the res higher, then open both to bleed), water simply leaks out of the radiator, while I know there's still air inside. Anyone know why that is?

The two chambers at the barb end has a pass through.

Fire went up when someone cut there PA to discover this. However martin noticed it when he flashed a flashlight to notice light bleed from the other side.

They work great despite having this passthrough, hence why we decided to just ignore it and accept it.

[Raxxath]

The two chambers at the barb end has a pass through.

Fire went up when someone cut there PA to discover this. However martin noticed it when he flashed a flashlight to notice light bleed from the other side.

They work great despite having this passthrough, hence why we decided to just ignore it and accept it.

So air gets stuck in there? Does that mean the bleeding method described on the site doesn't work at all, or only with the air in the radiator's tubes?

[NaeKuh]

So air gets stuck in there? Does that mean the bleeding method described on the site doesn't work at all, or only with the air in the radiator's tubes?

no when the radiator is top mounted, meaning you have barbs pointed down, the air accumulates in your radiator because its the highest part in your loop.

As air fills the radiator, water can not replace the air. So it looks like your system is fully bleed but in reality its not.

This is why you need to TILT your computer from side to side while its bleeding if you have a top mounted radiator. Otherwise a solo DDC-3.2 cant provide the required head pressure to push that water in and force the air out.

Get it?

An Analogy, think of sitting in a pool, and getting a cup and fliping it so the opening is touching the water. Bring the cup down, you see how it holds air. Thats whats happening inside a TOP mounted radiator.

solution? Tilt the cup, hence Tilt your radiator.


However it looked like you top mounted your radiator. If you stood it up, this wont be the case.

Top mounted Radiator:
http://i125.photobucket.com/albums/p73/aigomorla/IMG_0838.jpg

[Raxxath]

NaeKuh, my computer is not and has not been inside of a case for some time now :p: . All this testing has been done on my desk, with no case involved. The temperatures are the same as when it was inside my case, too. I am able to tilt and stand my radiator in every orientation imaginable, and no air comes out, so I don't think that's an issue. I was merely wondering why the bleeding method outlined on the Thermochill website doesn't work properly for me.

[Raxxath]

Okay, I just put the e6700 in my other computer, cooling it with my $1 CoolerMaster GeminII heatsink and a 500RPM fan. Its idling at 30C. That's a few degrees less than my water cooling setup. The TjMax is listed correctly as 85C on both computers. WHAT IS GOING ON HERE? :shrug:

[NaeKuh]

Sorry i dont know whats wrong then.

>.<

to be honest i think your temps are where they should be at. You cant judge a system by its idle, you need to compare it on load.

However pull your gpu off the loop and try that.

[leoftw]

those temps that you're getting sound good to me . I think you're over reacting .

[Raxxath]

However pull your gpu off the loop and try that.

Already did that long ago. :p:


those temps that you're getting sound good to me . I think you're over reacting .

Perhaps, but I can't help but think something is wrong when I get slightly better idle temperatures on a CM GeminII than a FuZion, PA120.3, etc. I still have hope that you guys are correct, and that the overclocked loaded temperatures will be good.

Then again, the people with overclocked quad cores, more blocks in their loop, and higher ambients getting better temps than me kinda make me feel like something is wrong. :(

[NaeKuh]

Already did that long ago. :p:



Perhaps, but I can't help but think something is wrong when I get slightly better idle temperatures on a CM GeminII than a FuZion, PA120.3, etc. I still have hope that you guys are correct, and that the overclocked loaded temperatures will be good.

Then again, the people with overclocked quad cores, more blocks in their loop, and higher ambients getting better temps than me kinda make me feel like something is wrong. :(

....

Lesson 1. Everyone Lies.

:rofl:

Anyhow, no two systems will ever perform the same. Even when i just retube, the system acts differently.

You cant ever say oh his temps are X so my temps must also be X. You need to load her up and see how she does.

[Snyxxx]

The 8-10° delta between idle and load means your system is working like it should.

If you want to be anal, then the problem lies with either the IHS to die bad contact or CPU waterblock contact.

Everyone here is trying to help yet you will not let go. Try a different chip.

[Conumdrum]

Hate to be his car warrenty care manager................

[Polizei]

You don't see a problem when a $1 heatsink is outperforming a few hundred dollar liquid cooling setup by 10C? There's nothing wrong with a 15C difference between my ambient temp and idle temps on a lone dual core processor with a radiator capable of dissipating hundreds of watts of heat? There is absolutely no question that something is wrong here.

It is capable of dissipating hundres of watts of heat when there is a 10c different in ambient/water temps. If your cores are 15c above ambient, then you sure as hell don't have water temps 10c above ambient.

[Raxxath]

It is capable of dissipating hundres of watts of heat when there is a 10c different in ambient/water temps. If your cores are 15c above ambient, then you sure as hell don't have water temps 10c above ambient.

I'm not really sure what you're trying to say there, but I hope you're not suggesting my lone dual core is dumping that much heat into the water.

Anyway, can anyone else with a FuZion tell me if their baseplate gets warm under load? I did a quick overclock to 3600 mhz, 1.45v, and loaded it up for a few minutes(that's all its stable for, I have a lot more work to do :p:). With the overclock, the idle temps were about 41/37, and load temps were in the low 50s. Even so, I couldn't feel any heat from the FuZion baseplate, when I can easily feel it from the NB heatsink or GPU ramsinks.

Another strange thing to note is that core 0 is always a few degrees warmer than core 1, even after all these remounts.

[D_A]

It's not unusual to have variation between core temps. Differences in the internal diodes can account for some of it. Variations in the thermal junction below the IHS can account for more. Variations in the quality of the cores could account for the rest. Or all of that could be bollocks. If it's not a huge difference, don't worry about it.

[Raxxath]

Raxxath you didn't say where your located? equator, north or south hemishpheres. go to locale rent shop an see if have infared TEMPERATURE METER for rent.

this will give a better view of what your temps are doing on the over all with cpu performance

I've already ordered one. Hopefully it will make my life easier.

[Polizei]

I'm not really sure what you're trying to say there, but I hope you're not suggesting my lone dual core is dumping that much heat into the water.


Look at published graphs of radiators... they all specify their heat dissipation with a 10c diffrence between ambient air and water temperature. Your dual core is not putting out enough heat to get your water 10c above ambient, no matter how good the Fuzion is at moving heat from the processor to the water. The higher water/ambient air delta there is, the more heat you are going to be dissipating to the air. Without a huge heat dump to the water, there can't be a huge heat dump to the air.

Aside from that, how are you applying your TIM? Pictures?

[Raxxath]

Aside from that, how are you applying your TIM? Pictures?

I apply MX-2 with a dot in the center, then I press the waterblock down on top of it. I wiggle it as much as possible from side to side to get air out, but waterblocks don't have a very large range of movement as you know.

I tried the line method once, but the temps were worse. I may have done the line in the wrong orientation, though. Does anyone know which way the line is supposed to go in relation to the writing on the chip?

The last picture in my OP will show you the contact I'm getting. I'll see if I can take some more pictures of the TIM with my newer, better camera.

[Polizei]

Try less TIM. You don't need to cover the whole IHS, just right above the cores.

[Raxxath]

Try less TIM. You don't need to cover the whole IHS, just right above the cores.

I really don't use much. I think I used slightly more in that picture than I have been lately, because there usually isn't that much coverage.

I just went through 30 minutes of Prime95 at 8x450, 1.45v, with the maximum temps being 55C for both cores. The FuZion baseplate still isn't noticably warm to the touch, but the radiator is a bit warmer. Doesn't the fact that both cores stay within 1C of each other under load and around 4C during idle suggest that the sensors really are off?

[Polizei]

I really don't use much. I think I used slightly more in that picture than I have been lately, because there usually isn't that much coverage.

I just went through 30 minutes of Prime95 at 8x450, 1.45v, with the maximum temps being 55C for both cores. The FuZion baseplate still isn't noticably warm to the touch, but the radiator is a bit warmer. Doesn't the fact that both cores stay within 1C of each other under load and around 4C during idle suggest that the sensors really are off?

Read up on the 45nm dual core's temp sensors being "really off." They are saying processors at stock voltages and speeds are at 45c idle on water and 75c load. Some report one core is 10c hotter than the other.

Intel has had awful IHS/die contact ever since they released Core2 in August 2006. Yours is actually very good in terms of accuracy.

[smee]

Raxxath. I still don't see why your making such a big deal about these temps! :confused:
They just fine to me, thats exactly what I'd expect with your chip and setup.

[Raxxath]

Raxxath. I still don't see why your making such a big deal about these temps! :confused:
They just fine to me, thats exactly what I'd expect with your chip and setup.

Because they are a lot higher than I expected. Take a look at this (http://www.xtremesystems.org/forums/showthread.php?t=176531&highlight=q6600) thread for example, or the earlier posts in this thread. Barring people lying or reporting incorrect temperatures, people with similar setups, but often higher ambients, more blocks in their loop, and quad cores are getting better temperatures than me. Because of that, I can't help but think something isn't right.

And yes, these temperatures may be fine for now, but when summer comes around, I'll be forced to reduce my overclock at this rate. My room temperature can reach 35C+.

[Polizei]

My E6600 is a hot chip, just like your E6700.

[Serpentarius]

have tried injecting more thermal compound? or change to other thermal compound?

from ur image attached on ur 1st post

[Raxxath]

A little update: I've tried just about everything suggested to me, including rotating the FuZion, and trying all different amounts and applications of TIM. So far temperatures have only been worse. One of the last options remaining seems to be lapping my CPU, so I'm looking into that now. Could my processor really be so screwed up that I drop ~10C on load temps after lapping? Is that unheard of?

What really pisses me off is this thread. (http://www.xtremesystems.org/forums/showthread.php?t=187558) He has a quad core, a 5C+ higher ambient, almost the same loop as me, and actually had lower temps, both load and idle, and yet still recognized it as a problem. And he found a simple solution that reduced his temperatures drastically, to the point that his load is only slightly higher than my idle. How the :banana::banana::banana::banana: can someone tell me that my temperatures are normal?

[Kilyin]

A little update: I've tried just about everything suggested to me, including rotating the FuZion, and trying all different amounts and applications of TIM. So far temperatures have only been worse. One of the last options remaining seems to be lapping my CPU, so I'm looking into that now. Could my processor really be so screwed up that I drop ~10C on load temps after lapping? Is that unheard of?

What really pisses me off is this thread. (http://www.xtremesystems.org/forums/showthread.php?t=187558) He has a quad core, a 5C+ higher ambient, almost the same loop as me, and actually had lower temps, both load and idle, and yet still recognized it as a problem. And he found a simple solution that reduced his temperatures drastically, to the point that his load is only slightly higher than my idle. How the :banana::banana::banana::banana: can someone tell me that my temperatures are normal?

He's lying (lots of people exaggerate)... or he's not putting an Orthos/prime load on the CPU. I can rip DVDs and call that my load temps... doesn't make it true.

[Kilyin]

And as another example of someone not being completely honest...

http://www.xtremesystems.org/forums/showthread.php?t=179142

this dude claims he has a CPU load temp of 41, with a quad (QX9650) and 1.46 vcore. Sorry, but that's a load of horseshat.

[Raxxath]

He's lying (lots of people exaggerate)... or he's not putting an Orthos/prime load on the CPU. I can rip DVDs and call that my load temps... doesn't make it true.


And as another example of someone not being completely honest...

http://www.xtremesystems.org/forums/showthread.php?t=179142

this dude claims he has a CPU load temp of 41, with a quad (QX9650) and 1.46 vcore. Sorry, but that's a load of horseshat.

I really doubt they're purposely lying- what do they have to gain by that? Are you honestly suggesting that thousands of people on these forums are lying about their CPU temperatures? I can understand some people reporting incorrect temperatures due to software measuring them wrong. Even so, there are plenty of reputable people who know the proper software to use(like RealTemp) that report temperatures that back up my thinking.

[Snyxxx]

Maybe they are using "REAL TEMP" and getting cooler temps. ;)

[Raxxath]

Maybe they are using "REAL TEMP" and getting cooler temps. ;)

Very funny, but I'm using Real Temp too. It happens to report the same TjMax as Everest and CoreTemp for my CPU.

Are there any more suggestions for me? It seems like lapping is my last hope. If that doesn't work, I guess I'm just really unlucky, huh?

[Polizei]

I bet that your processor has horrible contact between the die and then IHS, internally. Or you have air in your radiator. Or your pump is on the fritz.


I really doubt they're purposely lying- what do they have to gain by that? Are you honestly suggesting that thousands of people on these forums are lying about their CPU temperatures?

You've heard the term e-Penis before, yes?

[F1ZZY]

Put your $1 heatsink on the same chip in the same mobo and compare that to your w/c temps.

How many people need to tell you that your temps are fine??

[smee]

Ok, heres a suggestion Raxxath.
Take your CPU chip out, put on the back plastic cover, get some 1500 grit (or higher) sand paper, and just rub it back and forth on the top of the chip for a couple seconds.
This will tell you if your chip is concave or not, my had some droops on it, so I lapped it.

I believe this could be your greatest problem. :up:

[Raxxath]

I bet that your processor has horrible contact between the die and then IHS, internally. Or you have air in your radiator. Or your pump is on the fritz.

I think bad contact is likely the cause, but I already ruled out the radiator and pump a long time ago.



I've pretty much read thru this whole thread, and from what I can tell, your temps are fine and well within specs for the CPU and setup. You've torn down the setup and re done everything several times from what I can tell/read. 42c under Prime 95 load, that's great temps from some of the testing I've done.

Yes, those temps would be great, but those tests were all done on stock settings. After overclocking, it idles higher than it used to load.



Be that as it may, I have one suggestion. Since your checking this with a bench type setup and the rad is out of the case, gently pick up the rad and shake back and forth. Do you hear fluid sloshing around. If so, there's a large air pocket in the rad. If you have enough slack in the tubing, and can do it without distrubing the set up to much, try tipping the rad while the pump is running so the inlet/outlet is on top. Then tip back and for slowly while shaking. Do you see any air bubbles flushing out? Only reason I suggest this is I've have several bench setups and seen problems with bleeding out the air from the rad.

It cannot be the radiator. First of all, if it were, my GPU temps would have suffered much more when it was in the loop as well. Second, I know exactly what that sloshing sounds like from redoing my loop so many times, and I've shaken and tilted the radiator in every way imaginable. There isn't any air in it.



Just thought of something else. Hook the pump RMP sensor to a MB fan port that the MB can read RPM's off of. That pump should spin around 4500 RPM +/- 300 RPM. If it's well below that speed, might be a problem there and the flow rate thru the loop.


I've already confirmed its not the pump. I've always had the RPM sensor connected, and it always reads a little lower than 4500. I had the exact same temps with a different pump as well.



Ok, heres a suggestion Raxxath.
Take your CPU chip out, put on the back plastic cover, get some 1500 grit (or higher) sand paper, and just rub it back and forth on the top of the chip for a couple seconds.
This will tell you if your chip is concave or not, my had some droops on it, so I lapped it.

I believe this could be your greatest problem. :up:

I wasn't sure exactly how the CPU flatness test with a razor works, but I tried what I thought I remembered reading. I held the blade to the IHS and shined a flashlight from the other side. Lots of light came through on each side of the very center, which would indicate its definitely not flat. That is, if I did that test correctly.

[smee]

I wasn't sure exactly how the CPU flatness test with a razor works, but I tried what I thought I remembered reading. I held the blade to the IHS and shined a flashlight from the other side. Lots of light came through on each side of the very center, which would indicate its definitely not flat. That is, if I did that test correctly.

To make sure, get some 1500 grit sand paper and just rub it on top of the cpu for a couple of seconds.
You will see what I mean, you'll immediately see if your cpu is flat or not and where it dips.

[brianm602]

I am not speaking from personal experience but I have heard that those DDC's dump a fair amount of heat into your loop, could this be a contributing factor?

[Raxxath]

I am not speaking from personal experience but I have heard that those DDC's dump a fair amount of heat into your loop, could this be a contributing factor?

In a word, no.

[eXa]

Why are u putting so much trust in the internal temp censor? what u should be looking at is the idle\load delta wich is fairly good in ur case.
ur censor probably reports higher temp than it is while the guy with 5c above ambient probably has a censor wich reports lower temps than it really is...

[HaCKs]

One of my systems has:

E6700
8800 gt sli g92

with dd mc-tdx and full coverage g92 blocks on the cards
black ice gtx 360 with 3 yate loons

and I get almost the same exact temps.

The following picture is idle, but I am primming right now, so give me a few hours and I will let you know. P.S this is in my small bedroom that had the AC on all night.

http://images.dr3vil.com//files3/190/ownedd.JPG

[Raxxath]

One of my systems has:

E6700
8800 gt sli g92

with dd mc-tdx and full coverage g92 blocks on the cards
black ice gtx 360 with 3 yate loons

and I get almost the same exact temps.

The following picture is idle, but I am primming right now, so give me a few hours and I will let you know. P.S this is in my small bedroom that had the AC on all night.

Those look like much better temps to me.

I'm pretty convinced that its my IHS. I read through the lapping thread here last night, and saw plenty of people who reported dropping ~10C. I also tried spreading a very thin layer of TIM on my processor a few times, and my idle temps were always extremely high, indicating that there was no contact.

I do have a question though, how much pressure should I apply while lapping?

[Soulwind]

I do have a question though, how much pressure should I apply while lapping?

You don't.

While lapping, the weight of the cpu (and the inevitable weight of your fingers) is all you need. Just go slow and steady. The last thing you want to do is to apply pressure (since it will almost certainly be un-even pressure) and mess your chip up even worse.

[Snyxxx]

You may have read this already, but here goes anyway:

http://www.thetechrepository.com/showthread.php?t=71

[Polizei]

Those look like much better temps to me.

I'm pretty convinced that its my IHS. I read through the lapping thread here last night, and saw plenty of people who reported dropping ~10C. I also tried spreading a very thin layer of TIM on my processor a few times, and my idle temps were always extremely high, indicating that there was no contact.

I do have a question though, how much pressure should I apply while lapping?

Lapping won't fix anything if you have bad contact between the die and the IHS.

[Raxxath]

After hours and hours of sanding, I put my lapped CPU back in and booted up. Aaaaand... zero difference.

Actually, it was even worse than that. When I first booted up, I saw that my temperatures had dropped 10C, and was very excited. Then I looked at my clock speed. Apparently it dropped to below stock when I took the processor out. When I redid my overclock, things were back to normal. :(

So the verdict appears to be that I have a crappy chip, whether its bad internal contact, or a messed up sensor. I have really bad luck with these things.

[warmage]

hours and hours of sanding seems like u lapped it way to much :P

Rax what Vcore your running at now ? , do you set your own volts or is it auto volts from the board ?
if you dont mind can you give your new temps idle/load/ambient

do your block/rad feel warm now ?

[FrogBite]

If you dont get any temp. drop from lapping you probably did something wrong.

[disruptfam]

I am not speaking from personal experience but I have heard that those DDC's dump a fair amount of heat into your loop, could this be a contributing factor?

are you sure about that mate

[smee]

It's not suppose to take you hours and hours of sanding to lap a CPU.
You may have lapped it so much, that the only (or main) thing the waterblock is hitting is the CPU bracket.....

It only took me 20min to lap my quad, better temps too.

[Raxxath]

hours and hours of sanding seems like u lapped it way to much :P

Rax what Vcore your running at now ? , do you set your own volts or is it auto volts from the board ?
if you dont mind can you give your new temps idle/load/ambient

do your block/rad feel warm now ?


It's not suppose to take you hours and hours of sanding to lap a CPU.
You may have lapped it so much, that the only (or main) thing the waterblock is hitting is the CPU bracket.....

It only took me 20min to lap my quad, better temps too.

The reason it took so long was that I used only the processor's pressure, and I sanded dry. That's what I was told to do. I barely sanded down after the nickel was gone, so don't worry, there's plenty more copper. Should I do it again with a little water?

My overclock is 450x8 1.425v. Idle temps are Core 0- 40C, Core 1- 37C. Load temps after about 20 minutes are 55C for both cores.

[smee]

The reason it took so long was that I used only the processor's pressure, and I sanded dry. That's what I was told to do. I barely sanded down after the nickel was gone, so don't worry, there's plenty more copper. Should I do it again with a little water?

Ok, yea water is much better.

What grits did you use? You are suppose to go 400-800 and if you want to (but not necessary at all) go to 1200 grit.

Anything after 800 grit does nothing temp wise, just makes it look shiny, and shiny is not always the best.

And when you lap, go one direction for 10 seconds, turn it 90 degrees, sand another 10 seconds, turn 90 degrees... etc.
Make sure to go from the top of the sand paper to the bottom, keeping it straight as possible.

At the end of 800 grit, it should look like this:
http://i254.photobucket.com/albums/hh106/parkman14/578d21a9.jpg?t=1211034869

[gxavier]

Your temps seem fine to me! Maybe try an EK Supreme, just for kicks?

Or if all else fails, invest in a water chiller!!! :)

[Raxxath]

Ok, yea water is much better.

What grits did you use? You are suppose to go 400-800 and if you want to (but not necessary at all) go to 1200 grit.

Anything after 800 grit does nothing temp wise, just makes it look shiny, and shiny is not always the best.

And when you lap, go one direction for 10 seconds, turn it 90 degrees, sand another 10 seconds, turn 90 degrees... etc.
Make sure to go from the top of the sand paper to the bottom, keeping it straight as possible.


I used 320 until all but a spec of nickel was left, then used 600 for a while. I did exactly as you say, except replace 10 seconds with 30.

[Jimmer411]

40C load temps is pretty good actually. Are you expecting to run at ambient?

[Raxxath]

40C load temps is pretty good actually. Are you expecting to run at ambient?

40C load temps? What are you talking about? I swear, half the people that post here haven't read anything in the thread.

[gxavier]

40C load temps? What are you talking about? I swear, half the people that post here haven't read anything in the thread.


10 minutes of Prime 95
Core 0: 42
Core 1: 40
GPU: 37
20 minutes of Prime 95
Core 0: 42
Core 1: 40
GPU: 37
Stopped, back to idle
Core 0: 36
Core 1: 33
GPU: 37

10 minutes of Prime 95 and ATITool artifact test
Core 0: 42-43
Core 1: 40-41
GPU: 42
Another 10 minutes of Prime 95 and ATITool, with fans on full speed
Core 0: 40-41
Core 1: 37-38
GPU: 40



Hmm

[Chickenfeed]

The temps you are reporting are within 2-3C (both load and idle) that I had with my current loop when I still had my E6600 (at stock). My Fuzion is a stock V1 and the cpu was unlapped at the time.

Again, I'd suggest you try clocking that thing higher, keep the GPU in the loop and tweak things from there. I know your not impressed with your current temps but remember your at stock. Like was said earlier in the thread, a dual core and pump isn't enough to make a PA 120.3 sweat. I'm sure you didn't purchase and setup this loop to merely run a dual core at stock so who cares where your idle is at 2.66? Your temps at load with a full loop are what you need to concern yourself with. Once you overclock your dual core, GTX and if you add chipset (S) to the loop then you can concern yourself with temps. If the addition of the above items does little to the cpu temps then your rad is doing its job. If they adversely change, better fans are needed for the rad to manage the extra heat.

Do remember many people who report their temps don't explicity write out their procedure and methods (let alone corresponding proof). There can be large discrepancies as people use different load tests ( even within Prime 95; small FTTs create the largest heat output and many seem to use large FTTs or mixed ram/cpu tests)

Without the full story on a machine and the testing procedures of the user, you can't fairly or even close to accurately compare two systems or loops. Like was mentioned before many things attribute to the differences in a given cooling system so direct comparisons aren't easy or entirely realistic to make.

Just a few of my own thoughts :)

[Raxxath]

Hmm

How many times do I have to say it, those tests were done at STOCK settings.


The temps you are reporting are within 2-3C (both load and idle) that I had with my current loop when I still had my E6600 (at stock). My Fuzion is a stock V1 and the cpu was unlapped at the time.

Again, I'd suggest you try clocking that thing higher, keep the GPU in the loop and tweak things from there. I know your not impressed with your current temps but remember your at stock. Like was said earlier in the thread, a dual core and pump isn't enough to make a PA 120.3 sweat. I'm sure you didn't purchase and setup this loop to merely run a dual core at stock so who cares where your idle is at 2.66?

Yet again someone who apparently isn't paying much attention. I ran it at stock before, I am now and have been overclocked to 3600mhz for some time.

[gxavier]

I think you should just tear down your water loop, throw it in the trash, and save us your abrasive attitude. Then you can go back to your "1 dollar heatsink!"

People are trying to help, but you aggressively dismiss them as stupid, or incompetent. Watercooling isn't for you.

[Chickenfeed]

Yes I realize you ran it stock before but I am interested in your temps with the system overclocked with both the GPU and CPU in the loop. You never mentioned if your 450x8 load of 55C was with the full loop (GPU in and overclocked as well). Regardless 55C load is an acceptable safe load temp (using small FFT stress ) if your GTX is in the loop. If your IHS is indeed bad theres nothing you can do short of a new CPU. I find the base of my Fusion does get rather hot to touch under load so if yours isn't then your thoughts on the IHS may very well be right. I had lapped my E6600 (helped perhaps 1C) and my Q6600 isn't lapped but the base still gets really warm.

[Raxxath]

I lapped the CPU again, with water this time(much, MUCH easier and faster). I think my temps dropped about 2-3C idle and load, but its hard to tell with the colder weather today.


Yes I realize you ran it stock before but I am interested in your temps with the system overclocked with both the GPU and CPU in the loop. You never mentioned if your 450x8 load of 55C was with the full loop (GPU in and overclocked as well). Regardless 55C load is an acceptable safe load temp (using small FFT stress ) if your GTX is in the loop. If your IHS is indeed bad theres nothing you can do short of a new CPU. I find the base of my Fusion does get rather hot to touch under load so if yours isn't then your thoughts on the IHS may very well be right. I had lapped my E6600 (helped perhaps 1C) and my Q6600 isn't lapped but the base still gets really warm.

Sorry, but I'm not going to be able to give you the temps of the overclocked CPU and GPU in the same loop, because I don't intend to run it that way again. The GPU will be getting its own MCR220, while the CPU, NB, and likely SB will be on the PA120.3. And yes, as I've said many times before, 55C is a fine load, but my ambient will be increasing by about 10C as we head into summer, and so will my temperatures. Those temps might still be fine for some people, I was simply expecting more for my money. Though its likely not my cooling's fault, but rather the processor's. Interesting on the hot baseplate. I know touch isn't a very accurate method of measurement, but I really feel next to no warmth from the FuZion, but plenty everywhere else.

Oh well, I'm tired of spending time on this- thanks to those who tried to help. I bet my temps will drop quite a bit with the next processor I get. :rolleyes: