Acceptable ambient/cpu temp diffs?

[Skeptic]

What would be considered acceptable differences in temps from ambient amd cpu while watercooling and overclocking?

I am currently 10C difference and barely overclocking my AMD 2600+ to 2100mhz. This is with the Maze3, Hydor L30, Black ice Xtreme, and using a mixture of Distilled Water with Hyperlube.

One thing to note is Hyperlube foams and it takes a LONG time to get the air bubbles out if you aren't careful the 1st time you fill up your system.

I just wanted to know what is considered a good ratio is all. My current temps are 22C ambient, 29C idle, 32C load.

[GenGoku]

If it helps any, these are my stats, though i just started watercooling 2 days ago.

Rig:
AMD XP1700b+ @ 2.2GHz (1.71v actual MBM5)
GeForce 4 Ti4200 (not o/c'd)

eheim 1250 -> dtek pro core -> white water -> maze4 gpu -> eheim 1250 (in series)

Idle:
room temp - 81.6F (27.5C) digital thermometer
CPU temp - 116F (46.6C) MBM5 / 107.4F (41.8C) probe behind mobo
GPU - 102F (38.8C) probe behind card

room to cpu delta ~ 34.4F (19C) / 25.8F (14.3F)

i'm not too sure how much the gpu is affecting the water temps as my system is usually used for work and web-browsing purposes. i'd say 10C difference is pretty good, a lot better than me

[winewood]

Skeptic,

be skeptical on those temps. Knowing the CW of those as .26 at perfect conditions, thats more of a stretch. Do you have a fan blowing on the sensor? I would not be quick to think those valid. HOWEVER with the cooler ambient and a good rad.. maybe..? I think 10C from system temperature is possible. But not from ambient.

[Skeptic]

Well, the thermocoupler is placed behind the cpu in the same way jaydee did it,, and that was a real pain in the neck for me to do. There are fans inside the case, but now blowing directly near the thermocoupler.

Can fan air get inside on the back of the cpu well enough to change the readings drastically?

At the hottest point of the day, I sit at 27C ambient, 37C load, so the 10C appears to remain constant with a varying + or - 1C either way on both ambient, and load temps. Seems to be fine, does anyone have better?

I am redesigning the current system actually to provide better cooling from the rad by removing it from the case, and putting it on top of the case. I am designing a bracket system so it will sit 2 inches from the case with 2 small fans blowing air underneath the rad, and 2 120mm fans sucking that air through and out above the rad. I figure this will get some decent airflow through the rad, maybe cooling the water 1-2C more? I don't know, but I am giving it a shot.

In addition, I will be placing my res on the back of the case and putting it 4 inches above the rad and adding an air trap/fill tube to the res. Since it will be the highest point by default, and also at the utmost end of the line, it should be an effective airtrap.

Only thing I am missing is a How To to remove the heatsink fan combo on my GeForce FX 5900 ultra vid card. I am very skeptical of removing it and wasting a $500 card.

Anyone know of a how to anywhere?

[GenGoku]

removing an epoxy'd hs on the gpu can be tricky. i used a blowdryer over the heatsink for a while and that seemed to soften up the compound (i've also read some folks putting the card in the freezer overnight). place a credit card on the pcb and a flathead screwdriver or butter knife as leverage to pry off the hs. be very careful not to force it off too much as the epoxy can take a piece of the gpu with the hs--as it did to my geforce4 ti4200 but was lucky no adverse affects happened to it. *knock on wood*

if you're lucky, your hs is only held down by push pins and are a sinch to take off. if you already know this, ignore this part as it'll be for others reading this post. on the backside of the vid card, you will see a small plastic cone-like thing sticking out. all that needs to be done is to squeeze the cone enough so it will fit though the hole.