Stren's 2012 CPU Water Block Roundup

[stren]

Hi everyone and welcome to my 2012 CPU Waterblock Roundup. First off this is in progress, I'm waiting for final results using Indigo Xtreme before I publish the first round of data. This is just a placeholder for now!

First off let's thank the sponsors who made this happen, if you like reviews like this please support those who sponsor the tests:

Performance PC's were generous enough to sponsor six blocks - so big love to them:

Thanks to my other sponsors (in alphabetical order):

DT Waterblocks for sponsoring the 5Noz and Sniper blocks

Danger Den for sponsoring the M6 Block

EK Waterblocks for sponsoring the Supremacy Block

Indigo Xtreme for sponsoring the TIM

Koolance for sponsoring the CPU-370 Block

MIPS for sponsoring the IceForce HF Block (and Frozen/Aquatuning for getting me it)

Sidewinder Computers for sponsoring the Bitspower Summit HF

Spotswood Computer Cases for sponsoring the tech station

Swiftech for sponsoring the Apogee HD Block and a MCP35X2 pump to enable the measurements vs flow rate!

Blocks tested:

Alphacool NexXxos XP3 Light - Acetal Edition
Bitspower Summit HF - Acetal/Nickel
Danger Den M6 (aka CPU600)
DT 5Noz
DT Sniper
DT Direct
EK Supreme HF Rev 4 (Full Nickel)
EK Supremacy (Nicekl/Plexi)
Koolance CPU-370 Rev 1.1
Koolance CPU-380
MIPS IceForce HF
Swiftech Apogee HD
Phobya UC1-LT
Watercool Heatkiller 3.0
XSPC Rasa
XSPC Raystorm
XSPC Raystorm Full Copper

Pics:

Test Setup:

Rampage IV Extreme + i7-3930K @ 4.8GHz, 1.45VCore
Pump: MCP35X2 - PWM controlled by CF-635
Reservoir: - EK 400 Basic
Radiator: XSPC EX 560 with Push/Pull 2150rpm Gentle Typhoons using Bgears 120->140mm adapters
Quick Disconnects: Koolance VL4N series
Temperature Probes: Dallas DS18B20 fitted into a bitspower brass barb
Flow Meter: King Instruments 7520 series 10" 0-3.5gpm scale
Data logging and PWM control: CrystalFontz CF-635 with SCAB module

You can see more details here: http://extremerigs.net/2012/07/building-temperature-sensors.html

Results

Here

 

[stren]

Alphacool NexXxos XP3 Light - Acetal Edition

Exterior Photos

Spoiler: Warning: Spoiler!

Both mount plates are included:

Here we see the block with the intel mount, it loosely slots on top of the block:

Protective film was stuck on the block:

However the finish still looks damaged and not that flat from the reflection:

Also are those copper screws or just plated/painted?

Interior Photos

To Come!

Performance Results

 

[stren]

Bitspower Summit EF

Exterior Photos

Spoiler: Warning: Spoiler!

Box in a box to contain the instructions and mounting screws

Everything laid out:

Minimalist style

Internal Photos

Spoiler: Warning: Spoiler!

Base + Jetplate:

The channels are pretty thick and might explain the poor performance. There's the rest of the block. No o-ring for the jet plate unlike most other blocks:

The total area of the 33 channels is fairly large however (33mmx33mm):

Video

To Come

Perfomance Results

Much more to come!

 

[stren]

Danger Den M6

Exterior Photos

Spoiler: Warning: Spoiler!

I was also sent some shims and some backplates:

No protective bag for the block, it seemed kinda dirty in there too:

The block has a notch on one side - not sure why yet - maybe we'll find out soon!

Underside was taped to protect it:

The underside is bizarrely copper - most nickel blocks are full nickel or nickel base + plexi/acetal top:

The tape leaves some residue that needs to be cleaned off:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

DT 5Noz

The 5Noz was DT's first block - DT sent me two a while back for use in my build (Project Thief). The block is now discontinued in favor of the sniper block. So these photos of the box aren't the exact unboxing- they are from the sniper. However the packaging was basically the same:

External Photos

Spoiler: Warning: Spoiler!

Some nice surface texture:

Large base:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

DT Sniper

Exterior Photos

Spoiler: Warning: Spoiler!

Spare o-ring is a nice touch

Internal Photos

Spoiler: Warning: Spoiler!

Here's the white version of the sniper:

This was the rev2 as opposed to the rev1 which I received in black. The performance difference was very slight due to a slight finetuning of bow.

Four allen key screws to remove the base, leaves the jetplate and o-ring clearly visible. The jetplate can only fit in one way.

Removing the jetplate to see the other side:

The channel copper base with some staining as this is after testing was complete:

The 36 channels measure an area 28mm wide by 29mm long:

Video

To Come

Performance Results

 

[stren]

EK Supreme HF Rev 4

External Photos

Spoiler: Warning: Spoiler!

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

EK Supremacy

Exterior Photos

Spoiler: Warning: Spoiler!

Unboxing:

Nice touch on the text on the side of the block:

One final close up of the base:

Internal Photos

Spoiler: Warning: Spoiler!

As we are testing on socket 2011 we need to change from the generic "good for everyone" jetplate to the specific 2011 jetplate, the allen key is provided in the box:

After unscrewing the 4 screws you can take the block apart:

Swapping the jetplate is quick and it fits nice and snug:

Put the mount plate on, then the o-ring:

Then seal it up.

Video

To Come

Performance Results

 

[stren]

Koolance CPU-370 Rev 1.1

Exterior Photos

Spoiler: Warning: Spoiler!

Unboxing:

Chunky mounting plate - I like that it attaches to the main block and isn't loose like some others:

Surface looks uniform unlike the alphacool:

Internal Photos

Spoiler: Warning: Spoiler!

These are after testing, there is some staining, however coolant wasn't changed between blocks, so it can't necessarily be attributed to Koolance.

Jetplate and top:

Removing jetplate with it's small o-ring:

Staining is close to the o-rings:

Base shows very fine channels:

Approximately 53 channels 34mm long x 29mm wide:

Video

To Come

Performance Results

 

[stren]

Koolance CPU-380

Exterior Photos

Spoiler: Warning: Spoiler!

As we unbox we see instructions, various backplates, TIM and the tools required to install:

The block itself is fairly heavy, mainly because of the hefty nickel plated mounting plate
that goes around the block. This is fairly typical for Koolance and a similar weight was used
on the CPU-370:

Koolance have a nice sticker that protects the base which is of higher quality than most of
their competition:

The base is a perfect mirror and the plating is of excellent quality:

From the reflection we can pinpoint some non-linearities in the base which are presumably
related to a jetplate:

The other nice touch is that Koolance have updated their nut design. Instead of a circular thumb nut like the 370 block, the nut now has slightly rounded off wrench flats so that you can use a torque wrench to install if desired. Although Koolance recommended 0.6kgm*cm (0.04lb*ft) to me, it's not included in the instructions. As I follow the principle of installing blocks for test like an end user would then I did not measure the torque and instead followed my nose as usual. Indigo Xtreme will be used at a later date to take out any mounting error effects.

Block Internal Photos

Spoiler: Warning: Spoiler!

I only look at the internals after testing. So any marks/staining is most likely because the
block is now used and not necessarily related to manufacturing quality. 4 Torx head screws
hold the 3 main parts of the block together and are removed using the proved allen key:

After removing the mounting plate the "core" of the block remains:

This comes apart easily showing the acetal top with 2 o-rings and the base with jetplate:

A close up of the acetal top, some manufacturers like a 2nd o-ring for the jet-plate and some
don't. I haven't particularly noticed a performance difference as a result though:

A close up of the jetplate which looks like most others:

The base has some staining around the o-ring location:

The machined channels in the base are much deeper and wider than on the previous CPU-370
block. The CPU-370 favored very many small and long channels.

The channel area measures about 28mm wide:

By about 30mm wide:

It would seem that the total channel area of the CPU-380 (45 channels in an area of 28x30mm)
is less than the CPU-370 (53 channels in an area of 28mmx35mm) but because of the depth of the
channel cut that the surface area may still be higher.

Videos

To Come

Performance Results

 

[stren]

MIPS IceForce HF

External Photos

Spoiler: Warning: Spoiler!

Unboxing:

No protective wrapper for the block, PK-1 is included though which is nice:

The mount plate is not attached, it also doubles as a throwing star!

No protective plastic sticker on the base, it's also the smallest base:

No visible distortion of the base:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

Phobya UC1-LT

External Photos

Spoiler: Warning: Spoiler!

Unboxing

One of the lightest blocks in the group:

The copper base and warning look identical to alphacool's. I wonder if they share that part. The jetplate and top is definitely different though.

Again the same distortion as alphacool:

Less wear and tear though:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

Swiftech Apogee HD

Exterior Photos

Spoiler: Warning: Spoiler!

Unboxing:

The block comes attached to a backplate:

I suppose this is a nice demonstration for those whose first time it is. However for socket 2011 we'll need to take the backplate off and change the thumbscrews. THis is also the only block where thumbscrews are held in with circlips. This is nice but the downside is that you need needle nose pliers to change them.

Internal Photos

Spoiler: Warning: Spoiler!

Undo the 8 hex screws and you can remove the mounting plate from the block core:

Mounting Plate:

Interior of the block top:

Copper base:

Close up of the block top - note the mounting peg:

Close up of the copper base:

Measuring the base:

Video

To Come

Full Performance Results Here...

Individual Block Performance Results In Detail

Spoiler: Warning: Spoiler!

Here is the performance vs TIM/Mount - IX did not do as well vs MX2 as I hoped here.

Recommended IX Reflow Procedure

Spoiler: Warning: Spoiler!

- save your overclock setting in the bios then set to stock settings and shut down
- remove existing TIM, clean CPU thoroughly and apply ETI
- make sure ETI is sealed to the edge of the CPU
- Attach block and fully tighten all mounting screws
- boot at stock cpu clock settings
- monitor temps with realtemp and speedfan set to chart mode
- unplug pumps after booting
- launch a burn with prime or other
- watch for the temperature dip and wait until cores come back up to maximum temps and throttling occurs before shutting off prime
- let the cpu cool to it's ambient with the pump off (<1 minute)
- turn the pump on
- reboot and load your old overclocked settings from the bios

 

[stren]

Watercool Heatkiller 3.0

Exterior Photos

Spoiler: Warning: Spoiler!

Unboxing:

One of the bags of screws burst during transit:

Heatkiller know how to make two metals of different colors work together. Compare to the Raystorm full copper for a bad example:

No mirror finish but the tooling has left a light pattern, I doubt this will affect anything, it almost loooks quite nice to me:

No visible distortion of the reflection:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

XSPC Rasa

Exterior Photos

Spoiler: Warning: Spoiler!

Well I had this block for so long that I lost the box, I also don't have a specific 2011 mount for it, so I'm reusing the one from the XSPC Raystorm. I included it in the test because I already had it, and I know some people still consider buying the budget XSPC rasa kits.

I believe the rasa has a mirror finish - this has been stained and sitting in the cupboard so long that I gave it a quick sand with 1000 grit before testing:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

XSPC Raystorm

Exterior Photos

Spoiler: Warning: Spoiler!

I had this block almost since it first came out, but I figured again that I'd include, to see if there is much of a difference with the full copper block (I expect little).

"Faux" unboxing as this block had been unboxed for a long time, see the full copper unboxing in the next post if you want to see what it would really look like:

One thing that annoys me about the raystorm is not just that the mount is not attached to the block, but that the metal cover that should be glued to the plastic comes off very quickly, and now you have two pieces that can flap around when the block isn't installed:

Some water staining on the base from a previous install, the base itself is mirrored but the photp doesn't show that:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

XSPC Raystorm Full Copper

External Photos

Spoiler: Warning: Spoiler!

Unlike the regular raystorm that I owned for a while this one is brand new courtesy of ppcs

Unboxing:

Someone at XSPC was having a good old joke with that backplate design:

Again the black insert on the mount plate was loose. I really wish both of those could be fixed down.

Without mount plate:

You can see some of the bow on the base:

And again here:

Internal Photos

To Come

Video

To Come

Performance Results

 

[stren]

For the full results write up click here!

First off another big thanks to the sponsors because without them this never would have happened. If you like reviews like this make sure to support them so that they will support future reviews!

Performance PC's were generous enough to sponsor six blocks!

Results in Brief - far more info can be found on my website:

Average of all 5 MX-2 mounts

This data assumes all mounts are equally valid and that the average of all 5 is a good choice for performance metric:

Flow

Performance vs Pump Setting with Indigo Xtreme

Mount "goodness" (Delta between MX-2 and IX)

All Results at one pump speed:

Gold Award

The Koolance CPU-380 performed exceptionally besting every block on MX2 (nearly all of them by 1+C) while still being at the top of the heap with Indigo Xtreme. The quality and feel of the block is up to scratch too. Mount variability is good and the flow is decent too and I would heartily recommend it.

Silver Awards

The following blocks also performed very well and have their own niche uses:
- the Phobya UC1 performs very well and is also priced very well
&#8211; Swiftech Apogee HD because it performs very well and has a very nice mounting system
- MIPS IceForce performed very well and has very high flow
- the XSPC raystorm performs well and wins the value for money award!
- the Watercool Heatkiller performed very well with IX

Until Alphacool has fixed the single port threading issue I can not recommend the block!

 

[stren]

Reserved 4

 

[MaadDaawg]

The Swiftech Apogee HD rocks.

I've had all their earlier models and this one puts them all to shame. I run my 980x based gamer at 4.4Ghz with an HD on the CPU. Temps idle under 30C on all six cores.

I do not have it hooked up in parallel like it is designed for, but it's a great block!

 

[stren]

Reserved again

 

[stren]

All unboxing pics done; Still to do:

Interior pics
Finish testing
Present results

Please be patient guys the testing will take a while to complete and let me know anything else you'd like to see!

 

[skizzled]

All unboxing pics done; Still to do:

Interior pics
Finish testing
Present results

Please be patient guys the testing will take a while to complete and let me know anything else you'd like to see!

Awesome job so far stren. Looking forward to the results.

 

[Skripka]

All unboxing pics done; Still to do:

Interior pics
Finish testing
Present results

Please be patient guys the testing will take a while to complete and let me know anything else you'd like to see!

Are you testing just temps or are you doing restriction/flow as well?

 

[stren]

Awesome job so far stren. Looking forward to the results.

Thanks

Are you testing just temps or are you doing restriction/flow as well?

I'll be measuring temps vs flow/pump power. I'll plot flow at a fixed pump power, but I don't have a differential manometer (yet) to do real flow/pressure curves

 

[Neo Zuko]

Why not the Apogee Drive II as well?

 

[Jorona]

Why not the Apogee Drive II as well?

That Apogee Drive II is an Apogee HD with a MCP35X stuck onto it. Its performance should be very siliar to the HDs.

 

[Neo Zuko]

I'm trying to decide between the Apogee HD and the Raystorm Copper

 

[stren]

Yeah Swiftech may send the APD2 too, we'll see what happens with that.

 

[omegatotal]

watching this one! planning to buy some stuff in the future :-P

 

[owcraftsman]

Subscribed: I have two of these blocks in use. The Raystorm and Swiftech Apogee HD. I'm interested to see how it all comes out.

 

[t_ski]

Any updates? I have the funds to buy a new block if you have something here that will beat my Heatkiller 3.0.

 

[SonDa5]

Any updates? I have the funds to buy a new block if you have something here that will beat my Heatkiller 3.0.

I think the top performers will be very close in performance. I think yours will be in the top. I just sold my Apogee HD to buy a DT SNIPER. I think the DT SNIPER will be close to the top of the best in thermal performance but besides performance I think the DT SNIPER looks great and I felt good buying a block made in the USA.

 

[Pufhead]

any updates soon?

 

[SonDa5]

any updates soon?

Stren is almost done with the testing according to what he has stated over at overclock.net

http://www.overclock.net/t/1287684/strens-2012-cpu-water-block-roundup/300#post_18113899

Also the Socket 2011 DT SNIPER direct contact block was added to the review.

I'm looking forward to the final results as well.

 

[Pufhead]

Stren is almost done with the testing according to what he has stated over at overclock.net

http://www.overclock.net/t/1287684/strens-2012-cpu-water-block-roundup/300#post_18113899

Also the Socket 2011 DT SNIPER direct contact block was added to the review.

I'm looking forward to the final results as well.

aa didn't know he where posting on overclock.net to ^^ thanks
ya cant wait for the final results gonna buy a new block soon so just holding back until i see some results

 

[stren]

Results on the first page now, I'll be on vacation for two weeks so apologies for not being able to answer questions til I return!

 

[loki_racer]

I know this isn't your fault, but I have to say. As one of the 1/8th of all US males that is colorblind, that CPU Waterblock Flow Rate vs Pump Setting is impossible to read. I know one of the triangle lines won, just not sure which.

 

[SonDa5]

Thanks for the Epic review.

 

[Seventyfive]

I know this isn't your fault, but I have to say. As one of the 1/8th of all US males that is colorblind, that CPU Waterblock Flow Rate vs Pump Setting is impossible to read. I know one of the triangle lines won, just not sure which.

Another interesting fact is that 100% of all color based charts discriminate against color blind people

The MIPS won the flow rate challenge.