Thermalright XP120 cooler: The new King

Calculon · Post by **Calculon** » Tue Sep 14, 2004 8:50 am

What are Intel's and AMD's specs how heavy heatsinks should be?

I'm a bit worried about that the large heatsink increases the leverage applied to the mount by a heavy fan.

BillyBuerger · Post by **BillyBuerger** » Tue Sep 14, 2004 8:53 am

icancam wrote: Personally, I've never liked the idea of mounting a fan directly onto a heatsink because of vibration and the coverage gap caused by the fan's central hub. A fan mounted a minimum of 3 cm away overcomes the fan hub gap and, naturally, is not physically attached to the heatsink-processor-motherboard.

But wouldn't the heatpipes transfer the heat to the outside edges of the HS where the fan is most effecient therefore bypassing the need to worry about the center hub? I also figure that's why it works well in suck mode.

Likif · Post by **Likif** » Tue Sep 14, 2004 9:04 am

Mine would be rapid burping, continuous cursing, and occasional mouse and keyboard clicking. How male.

I think perhaps one of the cons of the XP-120 would be the price. Here in Denmark it currently stands at about 80$...

Trip · Post by **Trip** » Tue Sep 14, 2004 9:27 am

Calculon wrote:What are Intel's and AMD's specs how heavy heatsinks should be?

I'm a bit worried about that the large heatsink increases the leverage applied to the mount by a heavy fan.

Zalman bracket is an alternative.

I want to say Intel's limit is still 450g and AMD's 300g, even on the newer sockets, but I can't say for sure.

Post by **Ralf Hutter** » Tue Sep 14, 2004 10:10 am

Calculon wrote:What are Intel's specs how heavy heatsinks should be?

While it's not direct from Intel, this Zalman 6500 page quotes 450 grams as the max Intel HS weight.

icancam · Post by **icancam** » Tue Sep 14, 2004 11:00 am

BillyBuerger wrote:
icancam wrote: Personally, I've never liked the idea of mounting a fan directly onto a heatsink because of vibration and the coverage gap caused by the fan's central hub. A fan mounted a minimum of 3 cm away overcomes the fan hub gap and, naturally, is not physically attached to the heatsink-processor-motherboard.
But wouldn't the heatpipes transfer the heat to the outside edges of the HS where the fan is most effecient therefore bypassing the need to worry about the center hub? I also figure that's why it works well in suck mode.

Good point! Thermalrights are well engineered and no doubt the fan hub gap issue has been mitigated but not eliminated by the positioning of the heatpipes in the XP-120 towards the edges. However, I suspect that better performance would still be achieved if the fan was moved a short distance away since that portion of the fins in the center of the heat sink would not be directly below the fan hub. I'd guess that airflow would improve slightly and the fan could therefore rotate more slowly to lower the overall noise level (at the same heat dissipation level).

Admittedly, it is much simpler to simply attach the fan to the heat sink and I'd guess that any improvement would be minor and possibly past a sane level of diminishing returns. A great deal of my fascination with this silencing obsession of ours has been how even minor reductions in the rotational speed of a fan can create a disproportionately larger reduction in apparent noise level as one "tunes" one's system. Vibration (which is the second enemy of electronics after heat) would would also be reduced by a decoupled fan arrangement and that would also subtly lower the noise level.

Edward Ng · Post by **Edward Ng** » Tue Sep 14, 2004 12:43 pm

If your machine is lesser wattage, passive should work fine with XP-120, considering I'm doing fanless on Sigma One with a, "mere," SP-94 (albeit ducted). XP-120 should handily beat it.

In other news, I have received my XP-90 sample; hope to do the testing this weekend. I'll likely shoot the photos of it before the weekend.

-Ed

sorenbro · Post by **sorenbro** » Tue Sep 14, 2004 2:48 pm

Good news from Edward Ng, I do hope it's going to be possible to fairly compare the two heatsinks as they are going to be testet on two different test rigs.

Ralf, have you tried giving the Globe fan fewer volts, to see how low it will start, and see if this reduces noise further.

mki · Post by **mki** » Tue Sep 14, 2004 3:57 pm

Trip wrote:I want to say Intel's limit is still 450g and AMD's 300g, even on the newer sockets, but I can't say for sure.

300g is the maximum for socket A, but not for AMD's newer sockets. AMD Athlon 64 and AMD Opteron Processors Thermal Design Guide specifies that the cooler on socket 754/939/940 must not be heavier than 450g.

That same guide specifies also the limits for cooler dimensions. The amount of force the cooler applies to the socket depends not only on the cooler mass, but also on it's dimensions. The maximum cooler dimensions for socket 754/939/940 are 77/68/60mm (length/width/height).

Those are "the design target specifications that must be met for the AMD Athlon 64 and AMD Opteron processors to operate reliably."

Trip · Post by **Trip** » Tue Sep 14, 2004 8:25 pm

Thanks

jmke · Post by **jmke** » Wed Sep 15, 2004 3:07 am

alleycat wrote:Thanks for the review, Ralf. As a point of reference, would it be possible to see what temps you get running the HS totally passive?.

I tried this on an undervolted and underclocked P4 2.4C

"After these tests I clocked the P4 down to 1.8Ghz (150Mhz FSB) with 1.5v vcore (minimum this motherboard allows) and left the CPUBurn running after I unplugged the fan. The PC was stable for ~5 minutes before it hit critical temperatures and the system shut down"

http://www.madshrimps.be/?action=getart ... rticID=220

Spod · Post by **Spod** » Wed Sep 15, 2004 3:52 am

Edward Ng wrote:In other news, I have received my XP-90 sample; hope to do the testing this weekend. I'll likely shoot the photos of it before the weekend.

Apart from the SPCR standard 80mm fan, will you be able to test it with a range of 92mm fans, giving comparisons of heat and noise, especially as compared to the CNPS-7000?
A standard setup to compare heatsinks on an even keel is fine, but for those of us who may actually use it, we'll want to know how it sounds & performs using 92mm fans.
Thanks! I'm holding off my heatsink purchase until you've reviewed the XP-90, so I'm really looking forward to your findings!

Post by **Ralf Hutter** » Wed Sep 15, 2004 4:50 am

sorenbro wrote: Ralf, have you tried giving the Globe fan fewer volts, to see how low it will start, and see if this reduces noise further.

The slowest I've been able to run it at is around 550rpm. I didn't measure it, but that's probably around 4-4.5V. Yes, it's a tiny bit quieter at that setting but it certainly won't start up there. All the Globes I have need at least 5.5V to start if the thermalcouple is in the circuit.

Edward Ng · Post by **Edward Ng** » Wed Sep 15, 2004 5:26 am

There's no valid reason at all to only test with 80mm; it's as if Ralf tested an XP-120 with only an 80mm. I would never dare short-change Thermalright that way; even in my Hyper 6 review, I tested the SP-94 with 92mm.

I'll be using 92mm L1A and AF92CT, and then 80mm L1A only if it includes the clips for 80mm (didn't open the box yet). Might have to try both, push down and pull up, but that will most certainly push my review back. There's als a 92mm M1B I may try at lower voltage as well, since I found my sample to be outstandingly low noise at lower voltages, and obviously flowing more air than an L1A for the same voltage (and my 92mm L1A sample is mediocre, being far noiser than my 92mm M1B at 5 or 7volts).

Come to think of it, this review may not be done for next week, as that's way too much testing to perform accurately within the span of two days...

-Ed

Trip · Post by **Trip** » Wed Sep 15, 2004 6:37 am

Thermalright's website wrote:Dimension:

L110 x W125 x H63 (mm) Fin only, without fan

Weight:

360g (heat sink only)

Recommended Fan

Maker: Panaflo (Panasonic)
Model: FBL09A12M
Size: 92 x 92 x 25 (mm)
Bearing: Hydro Wave
Voltage: 12V
Speed: 2450 rpm
Air Flow: 48 CFM
Noise Level: 30.0 dBA
Weight: 110g (3.88 oz)

360g + 110g = 470g for the XP-90.

alleycat · Post by **alleycat** » Wed Sep 15, 2004 7:49 am

jmke - you left the fan attached to the HS to test its passive performance? Ummm...

jmke · Post by **jmke** » Wed Sep 15, 2004 8:17 am

uhmm. what difference does it make in that particular setup ?

none indeed

meglamaniac · Post by **meglamaniac** » Wed Sep 15, 2004 9:14 am

jmke wrote:uhmm. what difference does it make in that particular setup ?
none indeed

The fan is arguably a significant obstruction to airflow if it's left attached but disabled. Thats the last thing you want in the already minimal airflow situation of passive cooling - it could be enough to raise temperatures by several degrees.

chrisb71 · Post by **chrisb71** » Wed Sep 15, 2004 9:50 am

DanceMan wrote:Nice work, Ralf. Now for some controversy. From a post at Hardware Central:
Quote:
"And while I'm no expert on heatpipes, my understanding was that they lead the heat _upwards_.

I have an AeroCool VGA heatpipe cooler, and in their documentation, they say they don't just use gravity but rather some sort of induction grooves on the inside, so that it will work horizontally as well as vertically. I've read reviews that said the performance was the same. I suppose they all use the same technology.

All I know is mine looks like it should not work (an upside down U, with one leg on the chip, the other leg has the heatsink) but it does.

jmke · Post by **jmke** » Wed Sep 15, 2004 11:50 am

meglamaniac wrote: The fan is arguably a significant obstruction to airflow

what airflow? read my post , check the link

Edward Ng · Post by **Edward Ng** » Wed Sep 15, 2004 11:56 am

jmke, even passive coolers create convective air flow, just sitting there in open air of a cooler temperature than the heatsink itself. The only way to prevent there being any flow at all is to either set this all up in a dead vacuum, or completely seal off the entire CPU/heatsink assembly so that the little bit of air there cannot possibly move at all (that, or test in an environment where the ambient air is always the same temperature as the heatsink)...

Having the fan there disrupts the convective air flow, thus having an effect on your results. Want to be sure? Remove the fan, and try again.

This is why passive testing is always done with all active cooling related attachments uninstalled, so as to completely unhinder convective currents.

Don't underestimate convective currents either; they can get pretty strong!

-Ed

jmke · Post by **jmke** » Wed Sep 15, 2004 12:14 pm

the test enviromnent (typo somewhere in that word

) was very small; as to qualify for the "ompletely seal off the entire CPU/heatsink assembly so that the little bit of air there cannot possibly move at all"

I agree that testing with the fan is "the way to go" but as you can see, the XP-120 is not able to do it, even if it remained stable for 2-3 minutes longer without the fan, it wouldn't have mattered.

what I am planning on testing is putting the board into a case with a sidepanel fan, followed by a case with a front intake and see how well it does then

don't expect results of that any time soon, have tons of other stuff to finish first

Cehtna · Post by **Cehtna** » Wed Sep 15, 2004 11:55 pm

What do you guys think about using this cooler with no fan right on top, but use a 120mm PSU fan and a 120mm intakefan and a 120mm exhaustfan in a Antec Sonata? Think it will work, or should i use another config?

lenny · Post by **lenny** » Thu Sep 16, 2004 12:24 am

Cehtna wrote:What do you guys think about using this cooler with no fan right on top, but use a 120mm PSU fan and a 120mm intakefan and a 120mm exhaustfan in a Antec Sonata? Think it will work, or should i use another config?

Why not:

1. Empty 120mm fan frame on XP120
2. Duct fan frame to 120mm exhaust

icancam · Post by **icancam** » Thu Sep 16, 2004 12:27 am

Cehtna, what make and model of CPU are you intending to use with the XP-120? If it's not a real scorcher, such as an Intel Prescott, you might be able to get by with ducting the heat sink to the case exhaust fan. The PSU fan would take care of the rest of the system. If you intend to use a PSU with a 120 mm fan, a high efficiency model (such as a Seasonic Super Tornado) would run cooler and more quietly. You would also want to seriously consider improving the airflow into the case by opening up the front intake. Poor ventilation is an oft mentioned weakness of the Sonata design. There are many threads in the Forums concerning CPU ducting, CPU operating temperatures, and modifying the Sonata. Use the Search function and discover what's possible. Enjoy!

DrCR · Post by **DrCR** » Thu Sep 16, 2004 4:11 am

Cehtna wrote:What do you guys think about using this cooler with no fan right on top, but use a 120mm PSU fan and a 120mm intakefan and a 120mm exhaustfan in a Antec Sonata? Think it will work, or should i use another config?

Personally I would use a setup similar to the one you described. Except I would just have a duct from the heatsink to a 120mm case exhaust fan., perhaps another fan in the duct if the airflow wasn't sufficient enough. The fact that this heatsink likes sucking really opens up your options as far as ducting goes. I'll be watercooling my next rig though.

DrCR

___________

GlassMan · Post by **GlassMan** » Sat Sep 18, 2004 7:16 am

The xp 120 is awsome. Just got it installed and my overclocked cpu temps have dropped 10 C while folding. If I drop the fan speed (Vantec Stealth) to 600rpm temps rise to 5 C below the old Thermaltake SilentBoost level, while folding. The cpu is an amd64 3200+ (2200) running at 2500. Fan is in sucking mode
The hsf covers about 25% of the 120mm exhaust fan, and the fin orientation is vertical so it won't run passive while overclocking (w/fan in place) but it might make it stock. I'll try after this WU. If I had a bottom feeding PSU I have no doubt it would cool the cpu at stock speeds. (I dont think there is enough room to duct to the rear exhaust location.) If you could get the fins horizonally oriented,you wouldn't need a duct.
If you can't fit the xp120,and the xp90 is half as good, it will be awsome as well.

ghostboarder · Post by **ghostboarder** » Sat Sep 18, 2004 10:04 pm

Hmm........stupid question, does anyone think it might actually be possible to use the 120 (or possibly even the 90) in an SFF case? Im planning to start building an HTPC soon, and this sounds like the answer to our prayers.

Post by **Ralf Hutter** » Sun Sep 19, 2004 5:06 am

ghostboarder wrote:does anyone think it might actually be possible to use the 120 (or possibly even the 90) in an SFF case? Im planning to start building an HTPC soon, and this sounds like the answer to our prayers.

Jeez, I dunno. The XP-120 is a big mofo. I can't imagine it fitting into any SFF case that I've ever used (only Shuttles, IIRC). One of the major problems that I can see with the XP-120 is the height of the thing. Even with a 25mm thick fan on it, it's pretty darn tall and would probably interfere with the drive bay trays in most SFF's.

acaurora · Post by **acaurora** » Sun Sep 19, 2004 8:04 am

But perhaps you could just rely on the exhaust fan instead of adding additional height to it with a 120mm fan?

Thermalright XP120 cooler: The new King

XP-120 no fan right on top, but 2x120mm exhaust in a Sonata?

Re: XP-120 no fan right on top, but 2x120mm exhaust in a Son

Re: XP-120 no fan right on top, but 2x120mm exhaust in a Son