Official Water Cooling Thread of the Stock Maximus SE ROG NB Club:

[Talonman]

What I would like to establish some kind of idea of what 'typical' NB temps members are getting, when they do run the stock ROG block.

Asus'es official stance:
Fusion Block System:
H2O ready: The next step in thermal solution.
The Fusion Block System is a more efficient thermal solution compared to competing followers with complicated looks. It is a hybrid thermal design that combines the ROG´s renowned heatpipe design with the additional ability to connect to a water cooling system. By taking the entire integrated solution into design considerations, the user can enjoy exceptional thermal improvement to north bridge and south bridge with a single connection. The Fusion Block System is the most versatile, efficient, and the most advanced thermal system there is on a motherboard.

The Maximus SE boxes official stance:
Fusion Block System:
H2O Ready: The next step in thermal solution. The Fusion block system combines the ROG's renowned heatpipe design with the ability to conduct to a water cooling system for a 47% increase in thermal performance compared to other heatpipe designs.

Here is our collection of LOADED NB temperatures reported my various members. I would like to see more aftermarket loaded NB temperatures hit the chart too.
Please feel free to submit your data, and this chart will be updated.
We are looking for the hottest temperature, your NB can hit, under max load!
We are giving both the ROG stock cooling, and the aftermarket blocks full credit, for being able to keep a NB happy at idle by default.
That is why I felt recording the NB temperature at idle was not worth it. I was really more interested in the failing point.
This log's purpose was to simply record how high our NB temps were able to climb on the Maximus SE, using our current cooling solutions.

Some lessons I think we have learned about the Maximus SE:

1) If you are running the stock ROG cooling, and decide to remount it using a higher quality TIM, you may see as much as a 10°C drop in NB temp under max load.

2) If you decide to go with an aftermarket water block, you can possibly see another additional 10°C drop in your NB temp under max load.

3) Raju found that loaded NB temps over 47°C have been prone to failures, during longer Prime torture test runs. This is not an exact number, nor should it be considered as such. It is simply ment to help give you a better idea, of what NB temp you might wan't to avoid, for the best system stability. I don't believe it is a hard line drawn in the sand that can't be crossed, but rather use it as a guide myself, and was grateful for the information posted in his findings.
Raju's and Bingo13's Maximus SE review: http://www.anandtech.com/mb/showdoc.aspx?i=3149

Members reported loaded NB temp's using the stock 'Republic of Gamers' cooling:
ROG Member ----- NB volts -- NB Loaded -------------------------- CPU ------------------
bill d ------------- 1.30v ------ 34°C ------------------------------ Q6600 @ 2.4GHz
TheSamuri -------- 1.54v ------ 34°C -------------------------- Q6600 GO @ 3.9GHz (Reseated W/MX-2, NB was 48°C)
Renegade5399 ---- 1.65v ------ 34°C --------------------------- E6850 GO @ 4GHz
ragge86 ---------- 1.65v ------ 36°C ------------------------------ Q6600 @ 3.8GHz
.OCX ------------- 1.57v ------ 38°C ---Orthos Beta for 10 hours --- E6750 @ 3.4GHz
initialised --------- 1.52v ------ 39°C --- After 10mins of SP2004 ---- E4500 @ 3.2GHz
NysoO ------------ 1.60v ------ 39°C -------------------------- Q6600 GO @ 3.6GHz
raju -------------- 1.65v ------ 39°C --- Prime95 Torture Temp ---- Q6600 @ 3.6GHz (Reseated W/new TIM, NB was 49°C)
DerekFSE --------- 1.50v ------ 40°C -------- Running Orthos ------ E6400 @ 3.6GHz
BulldogPO -------- 1.65v ------ 40°C - 24/7 QMC crunching clocks -- Q6600 @ 3.8GHz
giorgos th. ------- 1.47v ------ 41°C ------------------------------ Q6600 @ 3.60GHz
neo rtr ----------- 1.70v ------ 42°C ----------------------------- QX6700 @ 3.6GHz
Aldy402 ---------- 1.63v ------ 43°C -------------------------- Q6600 GO @ 3.825GHz (NB was 62°C-65°C, loaded on air)
weescott --------- 1.74v ------ 43°C ----------------------------- Q9550 @ 3.8GHz
j0nnyr0773n ------ 1.54v ------ 44°C -------------------------- Q6600 GO @ 3.6GHz
Clump ------------ 1.62v ------ 46°C ---- After 1HR of Prime 95 ---- Q6600 @ 3.6GHz

Members reported loaded NB temp's in the 'Aftermarket Block Club':
ABC Member ---- NB volts -- NB Loaded ------------------- CPU --------------- NB Block
Creek Tha Gray ---- 1.61v ------ 24°C -- occt at 30min -- Q6700 @ 4.2GHz ------ EK S-MAX (Vapor Phase Change user)
Nikhsub1 ---------- 1.52v ------ 25°C ------------------ Q6600 @ 3.7GHz --------- MCW30
Talonman --------- 1.57v ------ 27°C --1.5hr Prime95 -- Q6600 GO @ 3.7GHz ------ EK S-MAX Acetal
WFO -------------- 1.46v ------ 30°C -Folding 48Hrs - Q6600 GO @ 3.7GHz ------ EK S-MAX Acetal
Clunk ------------- 1.74v ------ 35°C - Prime over night - Q6600 @ 3.6GHz ------ DangerDen Maze4

I am currently giving the block full credit for being copper:

Today I port matched the Fusions block inlet/outlet to the barbs i.d. Got copper shavings in the process. The block itself is copper "not" aluminum.

I stole these quotes from the "ASUS Maximus Formula SE " thread in the INTEL section of the board (Page 58 and 59):
It made me feel pritty good that my OC'ing performance might not take a big hit running the stock ROG block.

OK, I am highly suggesting plumbing the Fusion block into your loop if you can. Dropped NB temp from 47°C to 34°C under load. NB volts at 1.65. It also helps anything on the heatpipe lose a degree or two, my SB temp is down some too. It didn't kill my flow too bad, dropped it some, but not bad. All I had was a CPU block in the loop anyways. And the MCP655 has a lot of push power.

I haven't tried without the fusion block, but i have the same temps with it connected. 33°C idle, 36°C load on NB. Around 32°C both idle/load on SB. ~1,65v on NB.
I have a SB X-Fi Gamer card in my last PCI port and i've got no problems at all with it.
Run my Q6600 on 3,8ghz (423*9) 1,5v vcore 100% Stable
btw. Best board i've ever owned for overclocking

Thanks for the post's boys!

Is there any other water cooling members using the ROG's NB block? If so I would love to hear a report on how it is doing for you.

Post up some temps. If she is a good block, or good enouigh to amply cool the NB, let's make it official.

One reference shot of my loop, and my progressing build details are located here: http://www.xtremesystems.org/forums/...d.php?t=160606

[bill_d]

q6600 cpu 27c nb 32c sb 38c running all auto no oc

[Exahertz]

thats great news hearing that the block is indeed copper!

Ill give you full temps when i get my NB on a loop.

[Talonman]

I was thrilled about the block being real live copper too.

I look foward to your report Exahertz!

It is looking like our range of temps is about from 32C idle, and up to around 38C under max load.

That can't be holding us back any correct?

We still need more data!

If you have and doubt about how small the performance hit will be with a 3/8" section in your loop, read this:
http://www.xtremesystems.org/forums/...d.php?t=145388
The boys 'in the know' feel less than 1C between running 1/2" and 3/8" hose, which our NB block is 3/8".
I don't believe running a 3/8" block will cost us any OCing performance, and it is handy to use the stock ROG coolers.
I really was not up for the battle of removing the stock cooling, unless I had to for performance limitation reasons.

This chart will make you feel much better about using a 3/8" section in your otherwise 1/2" loop.
http://www.xtremesystems.org/forums/...d.php?t=147767
At least it did me...

I now think our biggest consirn should have been if the NB block was copper, and will it be enouigh to keep our NB happy.
My bet is that it will. Go ROG!

[-=j0nnyr0773n=-]

I have 1.54v to nb and it idles and loads at 44c.... I think something's weird about that. It's on water..

[Talonman]

Thanks for the post.

We shall see if 44C at idle and load is odd.....

It sounds kinda bad sensorish to me?

I'm new at the game, is your pump a strong one?

Throw up a picture of your entire loop if it's handy.

I would like to see it.

[-=j0nnyr0773n=-]

Thanks for the post.

We shall see if 44C at idle and load is odd.....

It sounds kinda bad sensorish to me?

I'm new at the game, is your pump a strong one?

Throw up a picture of your entire loop if it's handy.

I would like to see it.

Here is the pump I am using:

http://www.dangerden.com/store/produ...5&cat=0&page=1

Also, I will try to post a pic later, but my good camera is not available at the moment so it will be a crappy one form my phone.

[Talonman]

OK. Looks like a decent pump to me.

How many blocks total in your loop?

[-=j0nnyr0773n=-]

Just the cpu and nb. I have the pump turned up on full speed with 1/2" inner diameter tubing except for where it drops down for the nb.

[Talonman]

Then yeha, that is a full blown mystery.

More data, more data!

[bill_d]

i have 3 blocks cpu maximus to nb to cpu to my commando 2 d5's in series and 3 XtremeIII rads in parallel nb 1.30v 32c idle 34 load room temp 24c water temp 26.2c and about 1.7 gpm 3/8 tube
you can try lower nb volts
but i dont oc

[ElEctric_EyE]

What I really liked after flipping mine end over end was that there was a serious backplate to the NB WC block, whether that too had to be made of copper I dunno... Asus has got my allegiance. It is a high quality board. I'll be able to make more contributions when I get my MF*^%$&*$%$8800GT which is on B/order till 11/9

[Aldy402]

when I first got the motherboard, my tubing wasn't good enough to go from fuzion to NB, it was just too damn close together... the slight kink bothered me..

Im waiting for my MCW60 to arrive then i'll do CPU->GFX->NorthBridge
i'll post my results soon

[Talonman]

The next question is at what operating temp, will a NB, start to go South?

[Raja@ASUS]

We found that water cooling has to be used with air cooling for good results, kinda defeats the object a little I guess..

There are 2 other issues too,

1) The Fusion block is spot epoxied to the NB heatsink, there is poor contact between the 2 surfaces, no thermal paste is used by Asus to fill the gaps this bonding process leaves behind. Our block came off, and we reattached it using thermal paste and epoxy. Thermal paste in the middle, with a good quality epoxy glue around the perimeter of the block to hold it in place, it is now rock solid.

2) The supplied thermal pad interface between the chipset IHS and the heatsink often make poor contact, we had around 60% contact, with no contact in the middle. Replaced with good TIM.

Loaded temps during Prime95 Torture Temp @ 460 FSB were around 39 degrees @1.65VNB and 2x 120mm fans. Before the above 'mods', loaded temps were around 49 degrees.


Review should be up on Anandtech in couple of days..


regards
Raja

[Talonman]

Thanks for the post, and the info.

I look foward to your review.

Too bad about ASUS not attaching the NB in a highly heat conducting manor.

That is not what I wanted to hear, nor did I want to remove all my stock ROG cooling. That is disappointing news to me.

Mabey our temps will still be good enouigh?

I still think we need a few more temps reported, water cooling the stock NB block, before we can draw any big conclusions..

If we get more reported in our current range: (About from 32C idle, and up to around 38C under max load), we would still be fine.

Raju, when you say your block came off, do you mean fell off, or off with heat and slow twisting? I am wondering with your board hitting 49C under load, and your poor block mounting, if that might turn out to be a telling thing....

I kind-a hope it was just a production freak.

Currently, you are reporting the highest temp we have heard yet, running the Maximus SE stock NB ROG block on water at 49C Max load.

So far:
Renegade5399 - 34°C under load
bill d ----------- 34°C under load
ragge86 ------- 36°C under load
j0nnyr0773n --- 44°C idle and load
raju ------------ 49°C under load (After re-mounting, max went down to 39°C)

What a range!!

[jonny_ftm]

Too many posts with stock settings. In my opinion they are useless. Also, what do you mean by load.

If we want good readings and reproductible results, some more rigour methodology is needed.

- NB voltage
- A quad core or dual core (quads will produce so much more heat)
- What OC setting people are running and CPU/SB vcore: it will increase mosfets and SB temperatures, and through the heatpipes, the NB will be less cool

In my opinion, you should OC your quadcore at 1.5vcore and your NB to 1.6-1.7v and see how it does. Anything stock or a C2D won't need watercooling the NB with the massive heatsinks. As soon as you extreme OC: that passive stuff will be limiting the NB.

I'm also daring about the mosfets, for sure passive cooling won't be enough.

That's my opinion, I don't have an X38, but hey, it is hotter than a P35 that already needed active cooling on high quad overclocks

[Clump]

OK
Q6600 at 400*9 for 3.6 GHZ.
Vcore is 1.46 (per probe, don't remember what BIOS setting is LOL). It's prime stable for 4 hours +.

All other voltage settings on auto.
Per probe:
NB - 1.62V
SB - 1.07V
SB2 - 1.52V

CPU block is a D-Tek Fuzion. Rad is MCR220. Running rad fans at 80% using Qfan in BIOS.

Idle:
CPU - 42C (Lowest core is 39, high 44)
NB - 39C
SB - 40C

Prime 95 (large FFT) loaded for about 1 hour:
CPU - 60C (Lowest 57, high 62)
NB - 46C
SB - 46C

I have not run with the NB air cooledd, so can't compare. Seeing the 10 degree improvement from remounting looks like it might worthwhile though!

[Talonman]

Too many posts with stock settings. In my opinion they are useless. Also, what do you mean by load.
If we want good readings and reproductible results, some more rigour methodology is needed.

- NB voltage
- A quad core or dual core (quads will produce so much more heat)
- What OC setting people are running and CPU/SB vcore: it will increase mosfets and SB temperatures, and through the heatpipes, the NB will be less cool
In my opinion, you should OC your quadcore at 1.5vcore and your NB to 1.6-1.7v and see how it does. Anything stock or a C2D won't need watercooling the NB with the massive heatsinks. As soon as you extreme OC: that passive stuff will be limiting the NB. I'm also daring about the mosfets, for sure passive cooling won't be enough. That's my opinion, I don't have an X38, but hey, it is hotter than a P35 that already needed active cooling on high quad overclocks

Thanks for shaking the stick at me. Chart data level has been raised.
Good idea!

Thanks for the post Clump!
Outstanding data to mull over...

Prime 95 (large FFT) loaded for about 1 hour:
CPU - 60C (Lowest 57, high 62)
NB - 46C
SB - 46C

So after running Prime 59 for a full hour, using the stock ROG cooler, you managed to get your NB temp up to 46C.

That is a decent temp for full load correct, or is 46C on the warning track for NB temps at full load? Does INTEL have an official stand?

So far:
ROG Member ----- NB volts -- NB Loaded --------------- CPU ------------------
Renegade5399 ---- 1.65v ------ 34°C ------------ Intel Core2Duo E6850 G0 @ 4ghz
bill d ------------- 1.30v ------ 34°C ---------------------------- Q6600 Stock
ragge86 ---------- 1.65v ------ 36°C ---------------------------- Q6600 @ 3.8ghz
j0nnyr0773n ------ 1.54v ------ 44°C ---------------------------- Q6600 GO
Clump ------------ 1.62v ------ 46C (After 1HR or Prime 95) ------ Q6600 @ 3.6ghz
raju -------------- 1.65v ------ 49°C (Prime95 Torture Temp) ----- ??????
Note: After Raju remounted, max temp went down to 39°C

I realize this is not a real scientific test with all the possible variables taken into consideration involved in water cooling, but the goal is to just get a ball park figure of what temps the average consumer might expect to achieve running the stock ROG cooling. We then need to determine at what point does the NB actually start to loose performance due to excess heat. The root question is: Is the ROG cooling good enouigh to not impede performance in any way, and if not, at what NB voltage or temp do we start to take the performance hit?

[jonny_ftm]

OK

I have not run with the NB air cooledd, so can't compare. Seeing the 10 degree improvement from remounting looks like it might worthwhile though!

@Talonman: if I read well, he didn't use the stock cooler (passive), he watercooled his NB and if I read well he reseated it with a 10°C temp gain.

The X38 runs hotter than a P35 and some say hotter than a P965. I doubt any passive cooling will do it at above 1.6v on high FSB and a quad core. 9x400 MHz can be achieved on many P965 chips (except mine ) at 1.45v for the NB.

ASUS and all other brands had their stock thermal compound always low low grade.

Personally, and if I see the Maximus topic, things are presenting bad: Only 2-3 people got a 3.7-3.8 GHz on their Q6600 G0, and none have posted a Prime95 small FFT and Large FFT screen.

What I advice anyone to do: as Raju said: reseat the stock cooling heatsinks with some AS5 or MX-1.

I'm really interested in your great topic though, would be very interesting if people post detailed data like clump did, with all needed info

Personally, once I order my Maximus (yep, I'm tempted as soon as I see some Q6600 G0 stable at 3.8-3.9 GHz above 1.5vcore to verify how the mosfets and X38 manage it at these heavy loads), I'll remove the stock heatsinks, sit my two HR-09U on the mosfets (with 2 x 60mm silent fans), I'll put my HR-05 IFX on the SB, and I'll cut the heatpies from the NB block to watercool it using the stock fusion blockk

Yep, I'm mad to demolish that beauty circuit, but I'm sure, passive cooling will never manage extreme OC.

Hope to see more data soon on the limits of the stock cooling on this Maximus

Again, great topic

[BulldogPO]

My everyday 24/7 QMC crunching clocks are:
Q6600 G0@3.6GHz 8x450MHz FSB 1.55V Vcore.
NB Volts are 1.6V and NB temp is 40C.

My system is setup following way:
2x 240mm BlackIce 240mm rads -> Swiftech microres -> Swiftech MCP655 pump set to maximum speed -> Asus Fusion block -> Swiftech Apogee GTX CPU block -> back to radiators.
I use 9/13mm silicone hose, and few hoses are 11/15mm silicone.

[bill_d]

seen where 6 or 8 people pulled nb off board or apart trying to reseat if new umpowered up it works most of the
time but i know my luck i'd get a hole where nb was and i don't need a 280.00 square frisbee to give to the dog

[hornet]

My Maximus



Video->http://www.youtube.com/watch?v=gdbkI2GE3_E

Enjoy,

hornet

[jonny_ftm]

seen where 6 or 8 people pulled nb off board or apart trying to reseat if new umpowered up it works most of the
time but i know my luck i'd get a hole where nb was and i don't need a 280.00 square frisbee to give to the dog

Never force the heatsink to remove it. Unlock the plastic pins or any screws, and apply a hair dryer for 20sec on the heatsink you want to unsolder. It should come without any effort. If you still have resistance, don't twist or use any kind of lever tricks as advertised on some forums, just give it some more hot air with the dryer. When the heatsink is well hot to the touch, it will come without effort

[Talonman]

Thanks for the post hornet!!

Sharp looking system.