• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

getting 75C on Load

BassBomb

Diamond Member
Makes the room temp go up quite a bit

Stock cooler, stock volts

425FSB x 8 = 3.4GHZ

Stable however, but too hot for me.

Ambient is about 30C, probably more down where my case is

using coretemp to measure + orthos blend stress
 
Amazing, that, as you say (and I have no reason to doubt) -- the system is stable. I'm guessing that you dropped the multiplier for the processor a notch -- so you could get the CPU FSB up to 425 Mhz. Wait a minute . . . . let me check the processor specs on that baby . . . . Nope . . . . it's the stock muliplier -- 333 x 8 = 2.67 Ghz.

You definitely should replace the stock cooler. I've continued to use the stock cooler and fan on two builds using the E2640 and E2680 processors OC'd 33% above stock at the stock voltage. With those, and room-ambient at near 80F, they seldom exceed 54C under PRIME95 load temperature.

Others may differ from me on this, and especially, you're not clear as to whether your citing the TCase or TJunction temperature(s). If it's TCase, then your temperature is way too high -- a needless risk to your investment. If it's Tjunction, then TCase could be somewhere between 60C and 65C -- and that's too close (IMHO) to the throttling threshold.

Let me start with a disclaimer: The most recent aftermarket coolers seem to show inconsistent results between reviews published by different sources. I just purchased a:

OCZ Vendetta 2

In this review, it doesn't fare that well, but the fan packaged with the cooler has a "low" speed of 800 rpm, and my idea of "low" is more like 1,800 rpm. I will replace the fan.

In another review, the OCZ trumped just about everything but the SunbeamTech Core-Contact Freezer -- which wasn't reviewed. In the above review, you see the Freezer doing top-notch. The OCZ was $10 more, with the Freezer at CrazyPC selling for $40.

Then there's the ThermalRight IFX-14. The IFX-14 had a dearth of reviews, even months after its release in spring, 2007. But reviews last fall show it as a high-performer:

ThermalRight IFX-14

Notice how the reviewer controlled for different fan speeds -- and therefore (roughly but linearly) the CFM air-flow.

Last time I looked, the IFX-14 was billed at around $70 at HeatsinkFactory.

But for the E6750 and that overclock setting -- especially if you don't want to use a lot of case-space-mobo-real-estate with a large cooler, the ThermalRight Ultima 90 should be just fine. I don't care with others will say: the Ultima provides for both 120mm and 92mm fans, and just exactly fits with a 92mm fan. I currently have one fitted with a 92mm high-performance DELTA fan -- bound to be noisy at the top end, but can be controlled to its midrange through a motherboard plug and BIOS setting -- you could allow it to vary to the top end for temperatures indicative of the processor's full load.

The Ultima 90 may be $47 at HeatsinkFactory, but you might be able to find it for less elsewhere.

THINK I'M FINISHED HERE?! NOPE!!

Here's what you can do with the stock cooler-fan-assembly. So far, your options have had a price tag ranging from $40 to $70.

1) Remove the HSF and lap the extruded round base (hopefully down to bare copper. Figure this gets you a 2+C degree improvement.
2) Remove the processor, and lap the processor cap down to bare copper. This should net an additional 2+C degree improvement.

NOTE on (1) and (2) Some people have reported a 10C reduction in temperatures from lapping the IHS cap and heatsink base, and I think (with the E22x0 builds), I obtained around a 5C improvement by lapping the processor caps alone.

3) Go to HeatSinkFactory and buy a $5 tube of IC Diamond thermal paste. I am assuming that you used either the brown s*** that comes already spread on the stock heatsink base, or you cleaned it off and used Arctic Silver 5. In either case, you will see an improvement. With AS5, and a similar over-clock thermal profile, I measured something around 3C degrees improvement.

4) Examine your computer case ventilation and cooling strategy. For example, Coolermaster's Centurion case was offered briefly last January (NewEgg, I think) for an after-rebate price of $0.00, so that you paid just shipping and tax. Ordinarily, the Centurion can be had for between $40 and $60. But it nevertheless offers some fundamental potential for cooling -- with 120mm intake and exhaust fans, and there is potential for cutting case-bottom holes for additional 120's if you want to put casters on the case and lift it off the floor.

Some cases allow too much the mixture of fresh ambient air with air warmed by the mobo and CPU before exhausting the mix. These things can be remedied, but inefficient exhaust and airflow can increase processor temperature by a few degrees -- maybe more.
 
tj max said 100C
Ambient is what the temp of my house was (by the home thermometer)
I moved the case out of the cupboard and it hit 72 load now

Reducing to stock speeds I hit 62 load

NB, RAM, PSU are all hot to the touch (not unbearable)

I am using a thermaltake tsunami w/4 yate loon 120mm. HR03 GT mounted up with yate pulling air through it and pushing upwards, side window is modded to fit 120mm pulling through 92mm hole

used to have 55 load in winter at 3.2ghz which was acceptable
 
My personal view is conservative: If you can't keep the TCase below 60C or the TJunction (what coretemp reads) below 70C, you're either not cooling well enough or you've OC'd too far.

You probably know this. You'd set your system to a stock VCore setting. The temperature of the processor increases by a square of the voltage increase; it increases linearly with an increase in bus-speed.

With the HR-03, the only question I'd have is: How well the exhaust from the GPU is being forced from the case? Or does it mix with the case air before exhausting? In my setup, I've ducted the graphics card and HR-03 to pull air over the back-side of the card and down on the chipset -- then across the motherboard and out the case through a 120mm fan exclusively devoted to the motherboard ducting.

Also -- especially -- when I was doing the E26x0 builds and using the stock HSF's, I'd noted a lot of material posted by people concerning the possibility that the push-pin assembly might APPEAR secure -- when in fact, it wasn't. Check those suckers.

But I must emphasize: with the E6750 at stock VCore, you should be able to keep the core temperatures (TJunction for each core) to 60C or below under load with an aftermarket cooler such as those I've described. Further improvements are forthcoming with the lapping and diamond paste, even for high-performance heatpipe coolers.

I had an E6600 originally socketed in my ducted rig OC'd to 3.3 Ghz (up from 2.4 stock). With a ThermalRight Ultra 120 Extreme lapped flat to bare copper -- no lapping of the IHS -- and diamond paste, I don't believe the TCASE temperature (which I was measuring naively at that time) exceeded about 53C at full load and room ambient of 78F. At most, that meant that my TJunction CoreTemp values were hitting around 68C. The E6600 was a hotter-running Conroe than your E6750.

72C may be acceptable -- and you should know better than to restrict airflow around a case in a cupboard. If your room-ambient is 30C, then that's pushing 86F. My thermal benchies involve room-ambients of no more than 83F, and I've never gone over 67C once I'd refined the ducting improvements. And these are all processors (E6600 and Q6600) of the B3 stepping.

I'd seen a lot of people reporting coretemp values just above 70C, and satisified with that. Like I said, the TCASE is supposed to be 15C lower than the TJunction core temperatures, but I've seen the spread come closer to 18C. It may vary. And actually, with the dual-core processors, I think it was supposed to be more like 10C difference.

Others may chime in here -- and perfectly welcome -- but I thought that those G0 processors were topping out in the high 50C range OC'd to respectable levels and running with room ambients around -- probably -- 75F. I can only say -- if you're room temperature is around 85F, then high-60s Celsius for the processor cores may be what you could expect. And remember -- the "stories" I'm telling about other forum posters and their observations involved use of aftermarket coolers -- like the Ultra 120's or the Ultima 90.
 
I see no Tcase or anything like that, just TMAX and my two core temps. TMAX always says 100C.

Ill draw up a picture of how my case internal looks

http://pics.bbzzdd.com/users/funkbro2/untitled.JPG

2x intake (side fan blows down on motherboard)
1x exhaust to left of cpu

cpu fan I assume is blowing down on the board
gpu fan pulling air through hr03 and into cpu area

That said, even if I were to get a CPU cooler. The RAM, NB and PSU all do get as hot
 
BassBomb -- mah man!!

I took a look at the Tsunami on the NewEgg site, where the pictures are ample, large and clear.

That case is only slightly better for airflow than my old CoolerMaster WaveMaster.

I don't want to compel you to do what you might not be inclined to do, but after the WaveMaster, I stopped buying new cases and made significant mods to old SECC-steel tower cases from computers build in the 1990s. The reason for using the old cases: if you cut up the steel and make a mistake, the case had been junk anyway.

I got much better results by overpowering the case with intake air. I don't care what anyone else here says -- pressurizing the case has positive benefits -- especially if you can sequester the hot parts of the assembly and channel the already-pressurized air across those parts through narrow apertures -- which then exhaust the air IM-MED-iately from the case exhaust ports.

The problem with the Tsunami is -- first of all -- the door which restricts airflow to what is probably an anemic (but quiet) intake fan on the case front. I can say this -- at least you have the side-panel fan.

IF I HAD YOUR CASE, HERE'S WHAT I WOULD DO . . . (you don't have to do it, but . . . )

1) Look into the possibility of widening the vents or slots on the case door. Carefully. Carefully with a dremel cut-off wheel, a steady hand, and a fine file to smooth the cuts.

2) Eliminate restrictive mesh fan-filters from the front fan; replace with fiber-glass AC filters (about $1 per 18"x24" panel at Home Depot) -- cut to size (and do what is necessary to keep stray fiberglass strands from fouling the intake fan.)

3) Consider the use of double-wheeled casters bolted to the case bottom -- to lift the case off the floor -- in the event you wish to pursue (4) below:

4) Remove all electronic parts, drives, and PSU from the case. Cut 120mm fan holes NEATLY in the case-bottom (one -- even two may fit) -- mindful that the fans need to clear the motherboard and allow access.

After that, there are probably steps (5) through (10) -- whatever. You can build duct boxes with foam-art-board, or you can either shape or glue (Poly-Zap) Lexan plates cut to size to:
-- duct air across the component side, then up over the top and down past the HR-03 fins onto the chipset and other components.
-- duct air across the motherboard -- using components like memory modules to channel the air. Around the hot components, build the duct-plate so that air flows through narrow apertures around those components.
-- duct the air from under the mobo duct -- and if possible separately from the CPU cooler exhaust -- directly to the exhaust fans. You will have to deploy the fans to keep the airflow moving across the mobo and through the CPU cooler.

Also, I would investigate some fan replacements. If you want my opinion, I prefer fans with beefier motors -- even with high dBA noise ratings at their top-end speeds. But the idea is to control these fans from the mobo -- and the wise choice of fans would mean that the lower speeds will mean a lot less noise.

For the case-bottom fans I'd stick with either 120x25mm (even with LED lights), and the lower-case-front intake may only allow for a fan that's 25mm thick. HOWEVER, if you can fit a 38mm-wide fan in there, you have more options.

Even better would be 140x20mm or 140x25mm fans in the case-bottom. Modder's mesh would allow for filtering the air. More time and trouble: U-channel aluminum pieces pop-riveted to the case bottom would provide a way to insert more blue-fiberglass AC filter material.

Now -- I'm only guessing from your sig-configuration that you're using an Antec Earthwatts 500W PSU -- which should be fine. It's rated in excess of 80% efficiency -- which is also fine. Why is it heating up? I would only venture a guess that it's capturing all the warm air from you VGA and other components -- which is mixing with ambient internal case air.

If exhaust CFMs exceed intake CFMs, the case interior air becomes both rarified and warm -- I think. The heat on components has nowhere to go -- fewer molecules bouncing against those parts. Thus -- pressurization plus ducting equals better cooling.

Is that what you have? An Earthwatts EA500?
 
Sure -- that may be -- but like I said -- I tend to be conservative on cooling objectives. Well -- conservative about what is "acceptable" -- radical about attempts to keep it below those limits.

BassBomb had noted that he's reading CoreTemp values. Those are the "TJUnction" temperatures.

TCase is reported by other utilities. On boards with the nVidia chipsets, the nVidia Monitor program reports TCase.

If his CoreTemp values are at 75C with a room ambient around 85F, then TCase is between 60 and 65C -- which, I believe, is acceptable to the Intel "chapter 5.1" spec for his processor.

But . . . . cooler is better . . . .

Also -- there had been reports by some here that CoreTemp in earlier versions was not reporting those TJunction temperatures adequately or accurately, and I thought someone had pointed to newer versions of the program. And I think this problem surfaced with the Intel chipsets -- especially the newer ones . . . . X48, etc. But his is a P35 Intel chipset . . . .

EDIT:

Here's a heatsink comparison review that I might better have posted in the thread I wrote today about interpreting heatsink reviews:

Best CPU Cooler Performance, Q1, 2008 -- BenchMark Reviews -- April, 2008

They note that they controlled the room ambient to 1-degree variation for all tests; they used the same fans for all tests; same thermal compound. They used a Q6600 G0 processor, over-clocked from 2.4 to 3.3 Ghz -- with a thermal profile that would be hotter than BassBomb's E6750 -- also a G0 stepping. they do not report what they chose as a room-ambient, although I would guess it was below 80F and probably closer to 75F.

The stock Intel cooler and fan show load core temperatures (TJunction) at between 69.5C and 70.0C. There is no mistake that they are reading the core values as opposed to the TCase value:

"EVEREST Ultimate Engineer Version 4.20.1170 was then utilized to create core loads and measure each individual CPU core temperature. "

So it looks as though BassBomb's processor is really not running too hot if his room ambient is 30C / 86F degrees. He could reduce these temperatures, though, by as much as 10C and an expenditure of <= $10 for some wet-or-dry sandpaper and Diamond thermal paste; by more if he's willing to spend some ducats on another cooler. Case mods are a lot more effort, requiring tools and time -- as with building ducts.

But these review results point out the limitations of the stock cooler and fan.

ONE MORE EDIT:

In the linked review, the room ambient was almost chilly at an average of 20.2C. That's about 68.4F degrees. Figure a difference with BassBomb's ambient of just less than 10C. Assume a hotter TDP profile on the over-clocked Q6600 versus BassBomb's E6750, and the benchtest results on the Intel cooler are almost within 5C of the E6750's.

So: (a) his temperatures are normal, like we said, and (b) the temperatures can be improved considerably with some money and effort.
 
ok, well after having AC on in the house for the night and day
my load temps now at 3.4ghz is ~70 (case still on desk, first door open)
 
I just got a PM from a member who wanted to show off his transparent ducting mod. I won't post it -- that's his privilege -- but it's "ree-ul purty." And -- effective. He's running a Q6600 G0-stepping. Your E6750 at that over-clock setting should generate less thermal wattage than a Q6600. I just PM'd him back -- maybe he'll post a link to the "BLUE_LED" JPG photo.

There's no reason for the fellow to lie: his core temperatures were hovering around 50C.

Without the ducting, you could certainly get those temperatures to come down for an E6750 G0 by improving intake CFMs and doing some of the other minor stuff I mentioned. Right now, $40 or $50 seems like something I myself don't want to let go of -- it's been a bad year. But I WILL say -- you'd really notice the difference with a decent heatpipe cooler on that puppy.
 
Back
Top