The CES Swiftech waterblock Challenge/ XS Party

[Philly_Boy]

Ok. . .I have been fighting a 102-103 degree fever for three days. I feel a bit better today so I wanted to chronicle what happened last Tuesday night in at the Swiftech Challenge.

The Attendees:
Gabe Rouchon, the CEO of Swiftech along with his thermal engineer, Stephen. OPPAINTER, one of our own and now a tester working for Swiftech ran the challenge. Also attending were Danny Salandra, CEO of D-Tek, Free Cable Guy, Tony Leach from OCZ, Jinu117, Michael Schuette of Lost Circuits, [cTx]Sgt Vipor, Brad (one of the cTx clan members) and me. Charles, FUGGER, the owner of Xtreme Systems hosted the event.

A pic of the two CEO's present: Gabe from Swiftech (left) and Danny from D-Tek:



The Measuring:
We set out to do one 15 minute run for each block. We ran FUGGER's QX6700 quad core CPU up to 3.6ghz per core with 1.425 volts v-core. We then opened up 4 instances of CPU Burn 5 and took measurements after the 15 minute mark. We used Core Temp to monitor the temps and a temp meter to monitor ambient temperature. Once we measured the temps after the 15 minute mark, we averaged the temps on all four cores to get an average loaded temp. We subtracted the ambient temp from the average to get a result we dubbed the delta T or difference in temperature between ambient and the avg loaded temp of the CPU. The block with the lowest delta T won the Challenge.

First up we built the test bed and accompanying water loop:



The test bed's water loop consisted of a Swiftech Micro Res feeding a MCP655 pump which in turn fed a MCR320 radwhich fed the blocks used in the challenge. We used 7/16" ID 5/8" OD Tygon R3603 tubing throughout the loop.

The first (and only) challenger: The D-Tek FuZion
FCG brought a block from D-Tek called the FuZion. We found out at the last minute that Danny Salandra, the head guy at D-Tek was in town for CES and invited him along. This block has already been the topic of much discussion elsewhere in this forum.

It has a cast/machined copper base and an injection molded delrin top. The base has many pins with a depression in the middle of the pins about the size of a dime. The Vortex base is said to swirl the water around and up thru four exit holes where they pool to an exit barb. It has a slightly bowed base (on purpose) due to the way it is assembled. We thought it was the one block that had a decent chance to beat Swiftech's new Apogee GT. Tonight we would find out if D-Tek's David had what it took to fell Swiftech's Goliath.

A pic of the FuZion:



Here's the base (borrowed from another thread):



First Kris. . .err. . .FCG mounted the block into the test bed. He used the mounting kit provided by D-Tek and tightened the top nuts until we saw the Asus motherboard begin to bow. Once done we fired up the test bed and water loop to bleed the system. The FuZion had so much flow we never really got all the small bubbles out of the loop. It also caused quite a turbulence in the Micro Res. Anyway, we ran it for the full 15 minutes and took the temps.

FuZion in the test bed:



FuZion at idle:



Initially loaded:



After a few minutes:



Almost done the run:



When we finally called it at the 15 minute mark the results were:

C0-69c C1-58c C2-63c C3-58c = 62c avg less ambient temp of 22.9c yielded a delta T of 39.1c.

Not bad. . .not bad at all. But was it good enough to beat the new Apogee GT?

The Challenged - The Swiftech Apogee GT
Swiftech was founded in 1994 but in recent years it has emerged as one of the premier forces in the watercooling industry. It's products have become the favorite many computer enthusiasts and the Storm G4 and Apogee blocks have become two of the the premier CPU waterblocks in the industry.

So when Swiftech decided to come out with a new block and give us a peek during CES, we were excited to accomodate that introduction. The new Apogee GT is a bit different than a regular Apogee in that the diamond pins are smaller and there are many more of them. It will also come with two different sized o-ring gaskets to seal the base to the delrin top. One 2.6mm diameter gasket will yield a flat base (machine lapped to +/-.003" flat) and the other gasket is thicker, 2.9mm in diameter, and will yield a slightly bowed base that Swiftech has found to increase the performance of the waterbolck.

The Apogee GT:

New retail packaging - can anyone say - available soon at Fry's?


Here's the base and hold down plate:


A closer look at the base:


The famous bow in the base:





OK, enough hype over the new Apoge GT. Eric, aka OPPAINTER, mounted the Apogee GT in the test bed and we prepared for a run with the Apogee.

The mounting tool:



Mounted in the test bed:



We did manage to get in one run of 15 minutes. FCG was trying to confirm temps by using MBM5. Shortly after he customized the readout for the quad core Kentsfield CPU, the rig locked up and spontaneoudly rebooted.

Apogee GT idle temps:



Apogee GT initial loaded temps:



After 15 minutes



The final run temps were:

C0-67c C1-63c C2-65c C3-63c = avg loaded temp of 64.5c less ambient of 22.4c yielded a delta T of 42.1c.

Well, well, well!! It would seem that the Apogee GT got pwned! After the system locked, OPP, Gabe and Stephen from Swiftech set about to see if everything was ok with the test bed. Each time they remounted and rebooted the test bed the system would lock after they loaded up the CPU.

Refitting the Apogee: Gabe (L) and Stephen refit the Apogee while Danny from D-Tek looks on:



Apogee GT 2nd run loaded before lock up:


We decided to refit the FuZion to see if there was instability with the test bed that could've caused the system to lock.

FuZion remounted:


FuZion on 2nd Run loaded:


So, as you can see the FuZion performed consistantly well. A second run yielded a delta T of 37.25c and a third run yielded a delta T of 38.25c EDIT: The FuZion did lock up during 2 runs as well, leaving us to believe that there was some instability with the mother board. Perhaps when we torqued the Apogee GT down on one of the mounts we bowed the mobo to the point of breaking a trace. We ran the Apogee GT two or three more times and the same problem with locking up occurred. [/EDIT] Swiftech then ran a version of the Apogee GT block that they think will eventually replace the Storm. I am not permitted to show photo's or the test results, but that block did complete 2 successful runs of 15 minutes before the board began to show instability.

In the end it was almost 2am. We had been testing these blocks at this point for over 6 hours. We declared the D-Tek FuZion the winner. It's best run of 37.25 delta T bested the Apogee GT's only run by 4.85C. FCG got to claim the Antec P180 Quiet Case (A P180 case pre loaded with a complete Swiftech water loop) as the prize.

And so ended the Swiftech Challenge.