Switching to WCing only lowered my load temp by 3C?

gingo · Nov 13, 2004

This is ridiculous.
I had an SP-97 with a 92mm Panaflo on my XP-M 2500+. Its quite a nice heatsink but I figured I could do better with water. So I got my parts, set up the loop, leak tested for 24 hrs and it was all going good. I turned on the system today and my load temps under Prime95 or after a game of CS:S are only 3C under what I had before, if that. This is very disapointing. Idle temps seem to be higher now. The parts used are:
Swiftech 6002-A w/arctic silver 5
Black Ice Extreme
AquaExtreme 50Z pump
I have two aluminum 120's in a push/pull setup on the radiator with shrouds.

Another odd thing is my case temp seems to be higher now, sitting around 31C compared to 26C it used to be. And it is about 2C lower than normal room temp!

Strange happenings I tell you.

Btw, I know it may take awhile for the AS5 to 'break in', but in the past I have never noticed much, if any, temp decrease. And if it did drop a good 2C that still doesn't put me to where I think I should be on water.

At least it looks cool...

thorilan · Nov 13, 2004

well the stickies do have a section that talks specificaly about this

greenman100 · Nov 13, 2004

grrrrr

your case temps will appear higher, as will your CPU temps

there are not fans now to cool down your motherboard, so it and the sensors on it are heatng up

another reason tnot ot trust motherboard temps

I bet you can up your overclock, man

gingo · Nov 13, 2004

greenman100 said:
grrrrr

your case temps will appear higher, as will your CPU temps

there are not fans now to cool down your motherboard, so it and the sensors on it are heatng up

another reason tnot ot trust motherboard temps

I bet you can up your overclock, man

Well, I've heard all about temps not being an accurate measurement of cooling. But since I am measuring the temps from the same system they should be of relevance. I always hear people saying how their temps have dropped quite a bit after switching to water, shouldn't this apply to them to then?

Thorilan, I'll go look,,,

greenman100 · Nov 13, 2004

gingo said:
Well, I've heard all about temps not being an accurate measurement of cooling. But since I am measuring the temps from the same system they should be of relevance. I always hear people saying how their temps have dropped quite a bit after switching to water, shouldn't this apply to them to then?

Thorilan, I'll go look,,,

did you bother to read what I said?

gingo · Nov 13, 2004

greenman100 said:
did you bother to read what I said?

Of course I did. But maybe you could explain to me why some people NOTICE dramatic decreases in temperatures while others (me in this case) dont see much of any.

pwnt by pat · Nov 13, 2004

'Cause they still have good airflow around the socket and mobo sensors. The same is why some people notice a dramatic drop in temps after installing a fan to the back of the socket. It really doesn't do that much, just enough to trick the sensors.

gingo · Nov 13, 2004

pwnt by pat said:
'Cause they still have good airflow around the socket and mobo sensors. The same is why some people notice a dramatic drop in temps after installing a fan to the back of the socket. It really doesn't do that much, just enough to trick the sensors.

Good enough airflow to drop temps by 10C +?
Besides the only fan I dont have any more is the 92mm Panaflo I had on my SP-97. Although now I have 2 120mm fans in the case.
It just doesn't make sense to me. I want to beleive you, and I guess I have no choice, but it just doesn't make sence.

I'll try to get pics up soon.

eobard · Nov 13, 2004

Take some twist ties and mount your 92MM Panaflo roughly above the waterblock (It doesn't have to be exactly centered, this is just a very rough test anyway) and see how that affects your temps. They're right that sensors that come with computers aren't the most accurate but you are right from a layman's view that if it's the same system then it's not unreasonable to expect the results to be equally inaccurate, you may not know your exact temps but you will have a "better rough idea" of the change from air to water. I'm not saying this is a good way to compare temps, but without getting some more capable sensing equipment it's probably the best "layman's guess" that you can make.

The reason other people may get a noticeable change in their reading could be something as simple as them having a strong fan nearby to cool their northbridge and that airflow makes up for the lack of flow around the cpu socket that the CPU fan made.

Etacovda · Nov 13, 2004

Trusting NF7 sensors is just plain stupid.

They vary by up to 10 degrees, just by rebooting...

Keiron · Nov 13, 2004

Since you flatly refuse to read the sticky, I'll sum it up. NF7-S's use a socket diode to measure CPU temps. The socket normally would be cooled by air and would give you a quazi-accurate tempreature reading. Now, with WC the socket below the CPU containing that temp diode (you know the blue thing that triggers the 2-tone noise of doom) heats up, and there is no air to cool it. Try testing how far you can overclock and compare it to a previous air-based setup.
The 10ºC or greater temp drop you are looking for would require having CPU+Motherboard that supports on-die temp monitoring, or a decent mod (you can build a CPU temp monitor, it's on the mainsite somewhere or another).

sunrunner20 · Nov 13, 2004

Etacovda said:
Trusting NF7 sensors is just plain stupid.

They vary by up to 10 degrees, just by rebooting...

You are limiting that to JUSt NF7 sensors?! Try ALL temp sensors on every motherboard I know of will vary.

gingo · Nov 13, 2004

Etacovda said:
Trusting NF7 sensors is just plain stupid.

They vary by up to 10 degrees, just by rebooting...

Actually I have noticed this, not by 10 degree's, but a change nonetheless.

Glad to know about the socket diode and all now, makes a little more sense.

Aphex_Tom_9 · Nov 13, 2004

Etacovda said:
Trusting NF7 sensors is just plain stupid.

They vary by up to 10 degrees, just by rebooting...

also, even when using a vapor cooler with good cpu contact, when cpu temps should be well in the negative range, it still reads positively because it's just the socket reader. when i first installed my prometeia, i thought i had done it wrong because i was getting above-zero temps and i was getting pretty angry, then i realized it was just the stupid socket reader...

schmide · Nov 13, 2004

My temps just dropped 7c, 3 days after first system start. I was getting scared for a while, now I'm happy.

oswald · Nov 14, 2004

reassuming - what is the good method to get the REAL temps? Digidoc5? Hardcano13?

greenman100 · Nov 14, 2004

oswald said:
reassuming - what is the good method to get the REAL temps? Digidoc5? Hardcano13?

fluke makes some good thermometers

9mmCensor · Nov 14, 2004

greenman100 said:
fluke makes some good thermometers

Whats the best way to use one to get good results, without hampering thermal xfer though.

greenman100 · Nov 14, 2004

9mmCensor said:
Whats the best way to use one to get good results, without hampering thermal xfer though.

...stick a thermocouple next to the core?

9mmCensor · Nov 14, 2004

greenman100 said:
...stick a thermocouple next to the core?

that doesn't imped heat xfer, but how accurate is it?

Switching to WCing only lowered my load temp by 3C?

Member

Member

Disabled

Member

Disabled

Member

Member

Member

Give me a break Senior

Member

Member

Masta Tweeka!

Member

Member

Member

Registered

Disabled

Disabled

Disabled

Disabled

Similar threads