Naked X2 4200 Temps

I took my X2 4200 (IHS-less) out of my passive Zalman loop last week, and I'm wondering if my temps are abnormally high. Here's my setup:

DD D-5 ----> Zalman ZM-WB3 ----> ThermalTake M2 bay radiator w/ Silverstone 120mm @ 1900 rpm ----> DD bay res ----> back to the pump

At stock speed / voltage, I idle at 35 - 38 (~20 ambient).

At the current speed / voltage, I'm idling at ~44 (just reseated an applied AS5 as per instructions on the web site).

Currently at:



My concern is that because the Zalman block is scratched (ever so slightly) I'm not getting good contact. However, the Zalman block fits great on an IHS-less processor. I'm going to re-mount the block and see if temps lower at all.

I'm also wondering if the Thermaltake rad really is adequate. I'm willing to go external if necessary.

I would like to replace the Zalman block with a high-flow block that will work with an IHS-less X2. I've seen some threads indicating that the Apogee will work - is this the case?

BIOS idle temps

With the side of my case removed, I'm seeing BIOS temps of 40 system (case) and 44 CPU. This is after a 5-minute cool-down and reboot straight to the BIOS. Something seems amiss here, considering I'm in Maine and my ambient room is well under 20 degrees right now.

My MB is an MSI Neo4 SLI, and I'm using the 3.9 BIOS (Feb 06). I searched and didn't see any oddities regarding this board and temperature readings.

I have a probe-type temp monitor attached to an LCD readout. Where would I place the probe on an IHS-less proc?

Ignore the BIOS readings. They're meaningless. Use a program like MBM or Speedfan.

you would put the probe as close to the core as physically possible.
failing that bore a hole in whatever metal is directly above it, and realize it will just be close to, but not actual, core temp.

Moto7451 said:

Ignore the BIOS readings. They're meaningless. Use a program like MBM or Speedfan.

Click to expand...

Since they all get their input from the same sensors, why do you think MBM or SpeedFan would be more accurate than BIOS?

Moto7451 said:

Ignore the BIOS readings. They're meaningless. Use a program like MBM or Speedfan.

Click to expand...

As clocker2 said, programs like MBM and Speedfan probe the same sensors the BIOS does to report temps (in my case on the MSI, some Winbond variant). I've heard that the indirect temperature calculation done by such sensors is not very accurate, which is why I may switch to a physical probe.

I've reseated my Zalman block yet again, and I'm idling at 32 right now after 5 minutes in Windows. I've reaseated this block no less than a half-dozen times on this naked X2, and I think I finally have good contact. It's necessary to spread AS5 very thin - as in, use a razor blade to gently pull a paper-thin layer over the core.

(5 minutes later, and my idle is now at 34. Can I blame this on the TT M2 rad?)

have you tried flushing and/or blasting the dust out of its fins?
also, if you slap some generic HUGE car part rad in there temporarily and
see temp gains, then theres your answer

hope this helps

IHS for naked X2?

Spade said:

have you tried flushing and/or blasting the dust out of its fins?
also, if you slap some generic HUGE car part rad in there temporarily and
see temp gains, then theres your answer

hope this helps

Click to expand...

No dust whatsoever. However, my temps were ~7 degrees lower (load) with the Resorator (and that's with two 7800 GTs in the loop, albeit with an Eheim 1000 pump replacement). My old loop:

Resorator ---> CPU ---> Northbridge ---> GPU1 ---> GPU2 ---> Aquabay Rad ---> back to Resorator

At this point, I have to assume that either the TT rad has air in it, or that the rad itself isn't up to the task. Also, the block may be scratched and not getting good contact with the core.

I'm going to move to an external rad (2 x120, BlackIce I think) and try a new block. Any suggestions on a block for a naked X2?

Is the rad very warm to the touch? See if you can figure out a way to take at least a crude measurement of the coolant temperature, perhaps by dangling a fish thermometer into the res.

The radiator isn't warm to the touch, but the resevoir is (which is directly after the rad in my CPU-only loop).

And now, even when I'm running my X2 at stock speed / voltage, the idle temperature remains the same as when I up the speed / voltage. I'm hovering around 40 with an ambient of just under 20.

Running S&M (100%) all the way through, load temps max at 54 (with the overclocked settings in my original post). It seems as though I've reached a cooling minimum with my setup, although I'm not very pleased with it.

The res and rad should be at about the same temperature. You need to find out how hot your coolant is getting, then you'll know if your problem is with transferring heat to the CPU or transferring heat to the air.

I'll measure the water temp this evening and post my results.

Well you are using a pretty tiny straight-through rad to cool TWO cpu cores...those temperatures are actually pretty consistant with the amount of wattage that an X2 puts out and I'm suprised your temps are as low as they are. An external rad isn't required, you could just as easily squeeze a blackice pro single 120mm radiator into a vacant area of your case, and it will dominate the thermaltake tube rad.

The slight scratches are not affecting your temperatures at all, unless the underside of your block looks like a CD that someone stomped on and twisted their foot.

With your high system temps, those are normal too. You see, on most boards, the temperature sensing diodes or themsistors are located pretty close to the main processor part of the board, where heat from the MOSFETS and bridge chips and memory spreads out over the board where the sensors are located. Also, watercooling a processor takes away that nice flow of air over top of the board that performs a useful cooling function

Also, those half multipliers are a bad idea, since your board probably rounds down the bus speed so the memory controller works properly
(reference: http://www.hexus.net/content/item.php?item=1002&redirect=yes )

Pro*Banshee said:

An external rad isn't required, you could just as easily squeeze a blackice pro single 120mm radiator into a vacant area of your case, and it will dominate the thermaltake tube rad.

Click to expand...

I figured the rad is the source of the problem, though I've yet to measure my water temp. I'm going to pick up a BlackIce Xtreme next week, as well as a Strorm G4. I imagine I'll see lower temps as a result.

You say the TT rad I have is a "straight-through" rad. How does this differ from rads like the BlackIce series?

Pro*Banshee said:

The slight scratches are not affecting your temperatures at all, unless the underside of your block looks like a CD that someone stomped on and twisted their foot.

Click to expand...

There are only a couple of slight scratches, and I think my application of AS5 is appropriate to handle the depressions.

Pro*Banshee said:

Also, those half multipliers are a bad idea, since your board probably rounds down the bus speed so the memory controller works properly
(reference: http://www.hexus.net/content/item.php?item=1002&redirect=yes )

Click to expand...

Thanks for the link. I figured I was doing something odd with the memory frequencies when I set the 9.5 multiplier, so I'll up the FSB and drop the multiplier to 9. I'm in need of some better memory though, this generic stuff I have is holding me back.

aaronjb said:

You say the TT rad I have is a "straight-through" rad. How does this differ from rads like the BlackIce series?

Click to expand...

I have not used a BIX or any of the black ice rads, but i believe they are a double-pass rad like any other heatercore, unlike the thermaltake, which is only a single-pass.

Moto7451 said:

Ignore the BIOS readings. They're meaningless. Use a program like MBM or Speedfan.

Click to expand...

MBM doesnt come with sensory (its on your mobo) so...MBM would read the same as your BIOS.

**edit..Ditto

I think the problem is simply the "bayRad" whatever that is. Im not familiar with it, but a rad that fits in 1 or 2 bays will not disspiate the waters temp to something comparable to a DIY's kit_(mayority of respondents here)