Temps go up but won't come down. Help Please.

I'm having an odd problem. My rig is watercooled, specs in sig. As of recently, I added a new 3.0C and my temps started doing something wierd. They start out around 80/26C, as I start to stress my comp, they go up a little maxxing ou around 86/30C. Well they don't come down even when I'm not stressing it anymore. They just go up from there. So lets say I have it under load and it goes up to 30C, and stays at 30C. The next time I have it under load it goes up to about 35C. And then the next it will top out around 41C. It only goes back down after I have let the system turned off for a good while. What could be causing this? I've got good air flow through my Rad. At least I can feel plenty of air coming through it. And I've even had my side panels off for quite sometime now. Any ideas?

software. i delete a creative folder thinkn it would be aight since i uninstalled the drivers for the sound card and installed the sound drivers for the onboard sound. ever since my comp would go to load on its own and temps would shoot up. I had to reinstall windows to get rid of the problem. Maybe all that overclockn and crashing have somehow corrupted some files causing ur comp to be on load all the time and temps going up.

Maybe your block needs to be remounted. Maybe you used too much thermal compound, maybe not enough? The block needs good contact.

Yes, I could be software, I guess. But then again I can watch my processor load and it doesn't seem like it's under load. Also, I mounted the block really well so I know it's making good contact. My 2.8C never got even close to this hot even under full load. Something is going on and it's pretty weird. I might try reinstalling windows and remounting the block whenever my new P4C comes in tomorrow. Oh and what do you guys think about using mosfet heatsinks. The ones that sidewindercomputers has? These Mosfet Heatsinks

IMO, they wont do much. I made my own MOSFET heatsinks and they didnt help anything, mine even have a 40mm fan on them. They look neat though and its a somewhat fun project.

i have mosfet sinks from sidewinder too. they didnt do much. Now im more worried about them falling off and shorting something out. if your cpu all of a sudden spikes to a 100% load even for a sec ur temps will shoot up. so dont be surprised, and yeah reseat ur waterblock before reinstalling windows.

well here's my take on it.
well you say you get 26c at startup... then you stress. i'm wondering if you wait a good while before you stress your computer. because it takes a lot longer for water to heat up and while you think it's at 26c idle when you start, after 30 minutes or so, it may be a good deal higher maybe 30 or so, like you said...

if you stress before the water has a chance to catch up, it'll read lower, particularily if you don't stress for too long. Then when you stress again, when your water has reached equilibrium, it'll go from like 35-41.

Depending on how much water you have in your system, it'll vary the lag between the temps. More water = bigger lag cause there's more of a medium to absorb the heat, and vice versa. Also it takes a good while to cool the water. Shutting off the computer won't help much cause stagnant water in tubing and reservoirs aren't being dissipated too well.

Long story short, there's a huge difference in the response time between water and air...

I don't water cool..but it kinda makes sense to me...

Very good point. I had actually thought about that before as a reason for the seemingly huge leap in temps. Although I've tested it before and even if I start my comp up and let it sit here for about 30 minutes and come back to it the temps are hanging around 84-86F. Then I'll stress it, and it jumps up to about 96F and only comes back down to about 93. Then I'll stress it again and it will jump up to about 105 and only come back down to about 98. And the only other thing that I have to go on is the fact that my 2.8C's temps never got above 95F at full stress, and always dropped significantly at idle, usually back to around 83-86F.\

Edit: My comp has been on for about 30 minutes now, and I've only been doing light work on it so far and it's running right at 27C, my mobo is running at about 25C. No games, movies, editing. Just internet use mainly so far. At this moment I'm going to stress the CPU.

3 minutes later: Okay so I just stressed it and it went up to 32C.

10 minutes later: Sitting at 30C. Stressing it again right now. Went up to 37C.

5 minutes later: Came down to 35C. Stressing again. Went up to 40C. Only came down to 37C

Perhaps you have a heat dissipation problem somewhere in the loop that just recently popped up ? Have you double checked your pump to make sure it still works ? lol..

I remember one time when I played Half-life without my pump on for something close to an hour.. Imagine a Duron at 70C and your computer screaming at you when you start mbm

is your radiator clogged? it sounds like your water system is storing heat.... take your rad apart from your fans. i know once on a HSF for me, i was seeing +5C because over time the HSF was clogged with dust... maybe same with your rad.

I don't think it is. Only been using it for a few weeks, and temps were good with my old proc. I think this proc is getting way to hot for some reason.

Myhre...that is weird bro, not even sure what to tell ya. Have you moved the case from before, is it recirculating warmer air through the heater core, or are ambient temps rising in the room or around the case? It wouldn't hurt to remount the heatsink, just to rule that out.

That's just bizarre...seems like the heat isn't being dissipated for some reason. My take on mosfet sinks...they don't hurt, but if you've got good case airflow you probably won't see much benefit.

I think I'm going to remount the heatsink first thing tomorrow when my new mobo comes in. I have to admit, AS5 is a lot thicker than the thermal compound I've had to use recently due to no AS5 being sold locally. And I did put some AS5 on this CPU. I'm thinking that I may have applied to much, because I'm used to the other compounds having a very low viscocity. Anyway I'm going to check that out tomorrow. Hopefully I'll be past 250FSB tomorrow afternoon when I get home from work. I'm going to droop mod the board before I mount it to the case, but after I make sure the board is not DOA of course. Thankyou for all your thoughts and comments on the subject guys. I'll keep you all posted on whether or not I find the cause/problem and am able to find a solution. Thanks!