Temps too hot?

[ArchAngel777]

Originally posted by: Dark Cupcake

Originally posted by: ArchAngel777

Originally posted by: Dark Cupcake

Originally posted by: ArchAngel777

Originally posted by: SerpentRoyal
Idle should be around 40 to 42C with 75F room. You'll need room temp north of 90F to idle in the mid fifties.

It all depends on the air blowing over the heatsink. If you have a standard case, you are looking at a around 5-7c increase over room temperature. So if the air blowing over the heatsink is 35c (BTX Specification for case temperature) or so, your idle is going to be ~45c at Tjunction. It wouldn't matter at that point if you installed a huge ass heatsink, as you can't lower the temperature of the IHS (Tcase) lower than the air blowing on it, and the 4 cores (Tjunction) sitting under the Tcase are ~10c hotter. The simple formula is to simply add computer case temperature + 10c to find your idle temperature. This doesn't always apply as some people have ducts bringing in room temperature air for the CPU, which then, does not play by the case temperature rule. At any rate, just log the temperature of the air being sucked in by the CPU fan and add 10c, that will be your idle. If you have poor contact with the IHS or a concave chip, you may need to add a few degrees.

Mine q6600 B3 sits on 80C full load at 20C ambient and 84C full load at around 32C ambient. So whats the most likely cause of such small diff in load temps from huge temp diff in ambient temps? concave chip?

The heatsink is 120 ultra extreme, 2 fans and as5

The quads throttle at 82c, so that is why your temps are not higher than 84c under full load. Load CPU Rightmark and verify this, it will tell you if a core is throttling.

FYI - At 82c it will lower clock speed, if temps still rise, it will cut voltage, if that doesn't work, it will climb until shut off time. (100c Tjunction or Motherboard Thermal threshold). It looks like the thottling was able to stop your temps dead in their tracks at the cost of performance.

if TM2 Throtling did occur i should be able to see it in cpuz and everest where the frequency and voltage should reduce? That does not happen, and it stays at constant 3.15ghz at proper voltage. I have not noticed performance drop either.

Do q6600 have TM1 throttling like the p4s used to have? where instead of reducing clock speed a duty cycle is introduced.

Plus don't the G0 stepping start throttling at around 80C (their TJunction is only 85C), I read that throttling point for B3 was 95C.

You probably wouldn't see it because it only affects the cores that are over that temperature. If you really want to be certain, download CPU Rightmark and run that during your load test. Check all four cores... I am guessing you will see some throttling. But, who knows...

 

[MonKENy]

im sorry I didnt mean to thread jack. ill make my own thread if its a problem. but im just looking for a yes or no. yes its running to hot or no it seems fine.

 

[moforles]

Thanks for your responses everyone.

Here's an update on my status. I suspected that the Intel stock fan thermal pad was ruined, because I had to remove the fan once to replace the motherboard, so I used Arcticlean to clean the heat sink and cpu, applied some Arctic Silver 5, and reseated the Intel Stock fan.

Unfortunately, the Core temperatures are just as I reported many months ago: 49-54 idle. 75-80 at maximum load (two instances of Orthos). Notably, when I remove the side cover, the temps drop considerably: 45-48 idle, 69-73 at maximum load. This suggests that the poor cooling of the Wavemaster is the primary culprit.

The Wavemaster is notorious for having poor intake in particular. I think one reasonable solution would be to purchase a modded side panel with fan holes (performance pcs sells these). Would it make more sense to have dual 80mm fans on the side or a single 120 mm?

I can also look into PCI slot fans, as I have plenty of slots to spare. Any recommendations?

Also, my two front fans are plugged directly into the power supply. I believe this means that they are always running at highest speed. Would it make sense to replace these fans with more powerful (higher rpm) ones then?

To remind everyone, the Coolermaster Wavemaster has two 80mm front intake fans and one 80mm rear exhaust fan.

Here's my complete system:

Intel Core 2 Quad Q6600 Kentsfield 2.4GHz 2 x 4MB L2 Cache LGA 775
GIGABYTE GA-P35C-DS3R LGA 775 Intel P35 ATX Intel Motherboard
2 x CORSAIR TWIN2X2048-8500C5D
eVGA NVIDIA GeForce 8600GTS 256MB GDDR3
Samsung SpinPoint HD501LJ 500GB
ASUS DRW-1814BLT (Black 18X DVD+R ?)
Windows Vista Home Premium 64-bit (OEM)
CoolerMaster Black Wave Master (TAC-T01-EK)
CORSAIR CMPSU-520HX ATX12V v2.2 and EPS12V 2.91 520W Power Supply 100

 

[BonzaiDuck]

I had the WaveMaster -- a beautiful midtower case, by the way. I chose not to mod it, because my dentist wanted to buy it. He wasn't over-clocking, was running a "cool" Intel P4 Northwood 2.4 Ghz processor (you know . . . . you remember . . . "it was 'THE ONE'").

But I made an assessment of how to mod it for improved air intake without destroying the aesthetics, and saw an example of similar modding by someone who'd installed water-cooling in the case.

You need, with attention to straight, smooth cuts, to dremel a hole in the case lower-front behind the aluminum "monolith" bar that hides the vent holes. Then you need to cut a piece of modders' mesh to fit the hole, and shape it so that it presents a continuous surface with the rest of the curved aluminum plate behind the monolith.

There are several ways you can secure the modders mesh: you can drill holes and use pop-rivets, but you'll want to hide the pop-rivets from view; you can tap 6-32 screw-holes on the back side of the original aluminum surface which do not go all the way through the metal -- or if they do -- you can shear off and grind the ends of the 6-32 screws so they are flush with the front surface once they've been secured and tightened from the rear to hold the modders' mesh on there.

For shaping the modders' mesh, you only need a good eye, a ball-peen hammer, and something to use as an anvil. Before you fit the mesh to the case, but after it's been shaped, you'll want to sand it thoroughly, primer it with an etching primer (green stuff in a can), primer it with black sandable primer, and paint it with an aluminum-silver metallic paint.

I think this is a better idea than using the side-panel for vents. It allows better airflow over the hard disks, and may actually assist in keeping the noise level a tad lower.

Also, I believe you can replace thoses 80x25mm fans with one or more 92mm or even 120mm fan(s). I think a single 120mm fan is a better idea. And you want to duct the fan intake to the aperture covered by the modders mesh so that only exterior air is pulled into the fan's intake side. You can do this with foam art-board or sheet metal. The art-board version takes very little time, and once it's been glued together and made rigid, you can then primer it and paint it silver to match the WaveMaster's aluminum.

You may have to remove the lower drive cage temporarily to make the work easier, but it's held in there with pop-rivets if I'm not mistaken. So you'd drill them out, and then replace them when you reassemble the case.

 

[BonzaiDuck]

Posted in other threads, I've made detailed observatiions about cooling enhancements on "air" with the Q6600 pushed as high as 3.2 Ghz.

Right now, under "Blending," I project with good accuracy peak load temperatures of 65C with room ambient at 81 to 83F.

It's possible, I think, to (a) get a "bad" processor, or (b) a motherboard that doesn't record the temperatures properly and needs "calibration -- you can comment about the feasibilty of these speculations.

But I'd look into a several things regarding your problems here.

Case intake restrictions, and case exhaust bottlenecks.

Graphics card cooling.

Heatsink mounts defective or improperly/incompletely installed.

A mismatch between the convex base of the Ultra-120-Extreme and the way it's mounted on the CPU.

[Therefore, lap off the convexity, making sure that it grinds down evenly to maintain perpendicularity between heatsink base surface and the cooler's vertical tower of heatpipes and fins.]

Insufficient thermal paste; too much thermal paste; bad thermal paste.

Warped surface of the IHS or processor cap.

[Therefore, lap off the nickel-plate to bare copper and a flat surface. The copper IHS is about 1/16"+ thick, and the nickel-plate is thousandths of an inch. The nickel provides an opportunity similar to expert paint application for auto-bodies or computer cases: the so called "reference primer-coat.]

A defect in the soldering of heatpipes to the Ultra-120-Extreme. Here, I don't know how you would test it unless you have another computer handy, and you take control-sample measurements of the processor temperatures with the stock cooler. You do not want to lap the U-120's base before you RMA it. The reason they void the warranty for lapping is the risk to TR that some noobie grabbed the fins and heatpipes to lap the base, and did the damage to the solder-joints that way.

Without my motherboard ducting improvements, I'd imagine you would get load temperatures of 70 to 72C using the other enhancements I've mentioned. [Actually, it could be better. The ducting provides a bigger advantage to chipset cooling with use of an efficient cooler like the TR-Ultra-120, where ducting probably provides a minimal improvement.] So I'm guessing you would show something in the 65 to 70C range for the improvements sans ducting. And that would occur with room ambients close to 80F.

I could be wrong about all or some of this, but I'm using a motherbaord that's a power-hog, and I've pushed the B3-stepping Q6600 to 1.42V VCORE. I've measured the improvements with statistical accuracy. So I know all these things help.

 

[moforles]

For those who are interested, here's a look at the inside of my oven...I mean case.

http://wavemaster.wetpaint.com/