A real gaming system with Z68 and mATX form factor

Mousemagician

Limp Gawd
Joined
Feb 13, 2010
Messages
326
Hi guys,
I finally made the jump!:p
The story: after a (VERY) disappointing (VERY) short experience with the Zotac Z68 ITX board:mad:, I decided to switch to micro-ATX.:D
Everything began a few months ago, when I upgraded my old friend gaming system (Asus P6T DeLuxe, Intel LGA1366 i7-920, 2x fanless Gigabyte HD 4850 CrossfireX, in a Antec P182 case) to a small form factor with Gigabyte GA-H67N-USB3-B3, Intel i7 2600K, XFX HD6950 2GB, in a LianLi PC-Q08.
Here the thread of this build.

I got the H67 ITX from Gigabyte just to hanging on until a good P67 or Z68 ITX board would have come out.
Here the thread of expectations and communications with motherboard manufacturers on this matter.

All of you know the recent sad situation: no decent Z68 ITX boards, o at least, nothing that makes the difference with a good H67 ITX board, for now…:(:(:(
Was the H67 ITX solution disappointing?

NOT AT ALL!!!

Why, thus, looking for some beefier?

BECAUSE I DEEPLY NEEDED TO FEEL THE ELECTRIFYING SENSATION OF A ROBUST OVERCLOCK (I know that you understand what I mean, I know that!)

Here what I decided to do:
FROM THIS
external03.jpg


TO THESE
comparison01.jpg

In other words, I got a REAL SFF gaming system and a nice file server.
Components list for the micro-ATX
CASE: Thermaltake Armor A30 (Microcenter $90)
MoBo: Asus Maximus IV Gene-Z (Newegg $169.99+$7.87 shipping)
CPU: Intel core i7-2600K (recycled from my ITX system, it was from Microcenter $294)
CPU heatsink: Thermalright AXP-140 RT (
recycled from my ITX system, it was from Heatsink Factory $69.95)
GPU: XFX HD-695A-CNFC (HD6950 2GB) (recycled from my ITX system, it was from Newegg $269)
GPU heatsink: Scythe Setsugen 2 (recycled from my ITX system, it was from Newegg $44.99)
RAM: Corsair Vengeance 8GB (2X4GB), DDR3 1600, model CMZ8GX3M2A1600C9B (recycled from my ITX system, it was from Newegg $80)
PSU: Cooler Master Silent Pro M700 (recycled from my old socket 1356 system)
HD: 128GB Crucial RealSSD C300, model CTFDDAC128MAG-1G1 (recycled from my ITX system, it was from Newegg $254.99)
RAID 0: 2x Hitachi Deskstar 7K1000.C SATA 3.0Gb/s 1TB 7200 RPM (recycled from my ITX system)

The build story is gonna come after...
Just let's know that I got an easy and stable

cpuocd.png


4.7GHz (with turbo boost still functioning)
 
Last edited:
Nice build, how is the air flow with that case, i hear people bitchin and moaning about it, getting good temp with the OC?
 
Yea I know what you mean about fiending for a robust overclock. It reminds me of my first case, a Lanbox Lite, which was the predecessor to the Armor A30.
 
The decision about the case has been the most difficult for this build.:confused::confused::confused:
I was sick and tired of the absence of a sliding MoBo tray as on the LiaLi PC-Q08.:mad:
I liked a lot the concept of the LanBox Lite and I wanted to stay "SMALL"...:rolleyes:
Initially, I was going to take the LianLi PC-V354 because I wanted USB3.0 in the front of the case, a great style (I love my PC-Q08), good airflow, moderately small footprint; but it didn't have a sliding MoBo tray, damn... I was taking it, anyways...
BUT... I realized that the PSU would have been exactly on top of the VGA, reducing significantly the space for my Scythe Setsugen 2!!!:eek::eek:
Change of plans, sigh!
I ended up with the Armor A30, scared about all these bad impressions about air flow, positioning of PSU, reduced space for a good aftermarket PSU heatsink, those ridiculous 6cm fans in the back...
A few months ago I saw the A30 at my local Microcenter, and I was very well impressed, though!
For the modularity most of all.
Here a few pics:
The phase of organ transplants, from the PC-Q08 to the Armor A30
comparison03.jpg


building03.jpg


MoBo loaded on the tray...
genez09.jpg


Sliding it in!
matx01.jpg


matx02.jpg


matx03.jpg


Just enough room between the AXP-140 fan and the PSU, but the fan retentions must be bent to have the MoBo sliding in and out.
matx06.jpg


First, REQUIRED, update: a new fan for the AXP-140
matx07.jpg

A FANTASTIC Scythe Slip Stream 140x25mm Fan PWM Adjustable VR, code SM1425SL12HPVC-V
matx08.jpg

I have no enough words to describe the goodies of this fan:

  1. 14CM PWM
  2. Switch VR/PWM
  3. Long, long, long cable (did I say long?)
  4. EFFICIENT! An example: with the original Thermalright fan (3pin) it was constantly at max speed, noisy, giving 47C at idle. With the Scythe I got 28C at 800RPM, silent... incredible!
You must consider, though, that a free slot is necessary to install the commands (no problem with my case).
Now, I set the fan tuned by the MoBo:
fanxpertcpudefault.png


It fits better in the case:
matx28.jpg


matx36.jpg


matx29.jpg

And I love it!!!:p:p

Second upgrade was a PWM 12cm fan for the Setsugen 2 on the GPU. I got another Scythe: the SY1225SL12LM-P 120mm, code SY1225SL12LM-P. I couldn't use this in my PC-Q08 because it takes 3 slots with the videocard. Now, I can!:D:D
matx09.jpg


The old and the new side by side:
matx10.jpg


Thanks to this update, I can plug the PWM fan to the header in the videocard and let the Catalyst Control Center to tune it.
matx17.jpg


matx13.jpg


matx12.jpg


Third upgrade: another 12CM PWM fan blowing on the MoBo, placed in front of the HD cage, just in between the front intake blue led fan and the SATA headers in the board:
matx20.jpg


matx21.jpg


I know... too much spagetti wires.. I know... I'll take care of it.
The fan has been zipped to the side of the case.
matx24.jpg


The final result is impressive: low temps for the MoBo, a good help for the AXP-140, a breeze to the videocard, all managed by the MoBo (in other words: silent)
fanxpertchassisdefault.png


That is it, for now...
Next step: the motherboard
ASUS Maximus IV Gene-Z

A dream...
 
Last edited:
What about the asrock itx z68 board? Are there problems with that unit? I really wanted to roll an itx setup with my current stuff
 
What about the asrock itx z68 board? Are there problems with that unit? I really wanted to roll an itx setup with my current stuff
I never tried it, but you cannot get the 4.9 GHz that I see in your sig.
You cannot have the 4.7 GHz as well.
And neither 4.5 GHz.
VERY likely no 4.4 as well, or at least not stable and durable...
What you CAN get?
The best Z68 ITX out there (the Zotac is unpresentable).
Unfortunately, ASRock forgot to put enough phase power and heatsinks in the mosfets!
I've been really frustrated thanks to that...:mad::(
 
Time to talk about the backbone of this build:
the Asus Maximus IV Gene-Z
After the BIG disappointment of what the Z68 ITX can do (or, better, cannot do), the decision to move to m-ATX was like admitting to fail...:(
I was so craving a monstrous ITX system, really.:(
Anyways, now I'm thinking, instead, how come it took so long to upgrade to it!:eek::eek:
First of all, the final system is not that bigger compared to the PC-Q08; it's located in the same spot than the ITX, it's less tall, a little larger and deeper, just that!
In change, I have a full monty motherboard, with all the bells to have something really powerful and stable (I mean OVERCLOCK).
And, also, the potential for a CrosfireX or SLI setting!!!:cool::cool::cool:
But, what at the end convinced me the most to make the change was... the support for up to 5 PWM fans.
That meant that I get rid of the hard to set fan controller from Crystalfontz (of course, I'll install it anyway because is still an excellent and cool PC console), and monitor the fans via software.
Here a few pics:
Cool boxing
genez01.jpg


genez02.jpg


Here, my beauty:
genez04.jpg


Ready to get the AXP-140 on:
genez06.jpg


That's on:
genez07.jpg



genez08.jpg


Now, the memory:
matx04.jpg


These BIOS buttons and display are in the perfect spot to be displayed in the window of the case:
matx22.jpg


And, here, the auto overclock set by the AI Suite II, just for fun!
extremeautotuning.png

...with the simple touch of a button (the left click on the mouse):p;)

To be continued...
Manual OVERCLOCKING:cool::cool::cool:
 
Last edited:
Yea I know what you mean about fiending for a robust overclock. It reminds me of my first case, a Lanbox Lite, which was the predecessor to the Armor A30.
That's exactly what I was talking about!:p
It's nice to see that I'm not the only one to fell these... necessities.;)
 
To overclock with the Asus Maximum IV Gene-Z I referred to a reliable publisher:
Tom's Hardware
in the article
Overcoming Overclocking Ordeals: Asus’ Maximus IV Gene-Z
At page 8 they show an easy and detailed overclock to 4.7 GHz from the UEFI.
That's how I made it:
bios01.jpg


The RAM was originally recognized as 1333 MHz, so I set to 1600 MHz.Please, disregard for the memory RAM the 1.6V in the pic below, I set 1.5V, as indicated by Corsair.
bios02.jpg


Then, the DIGI+ VRM sub-menu:
bios03m.jpg


Apply settings, restart, and that's it!
cpuocd.png


100311cpuzram.png


I, then, launched the Prime95 64bit:
I got 87C in the CPU and 31C in MoBo immediately after I loaded at 100% the CPU, and the fans maxed out automatically:
tempspk.png


When I sopped the test, temps in the CPU went down instantaneously to 48C and to 38C after 60 seconds:
tempsdown.png


Now I understand why no ITX can stand such overclock... I don't have VRM temps, but there is no way that these ITX boards can dissipate the heat, most of all the ASRock, that doesn't have any thermal protection on them.
After this test, I recalled that the SPCR team recommends to set the PSU with the fan aspirating from the CPU area in order to direct the warm air outside; my setting was, instead, with the PSU fan aspirating from the top of the case, what a waste!
matx38.jpg


Thus, I inverted the PSU orientation.
I bet the warm air aspiration from the CPU area, now, will be even more efficient.;)

I'll make benchmarks pretty soon.
To be continued with...
Air flux inside the Armor A30
something that many people, me included, imagined problematic, but it's not, really!
 
Okay, a surprise that just came out:
even this system throttles when on Prime95 at 100% load:eek::eek::eek:
Before checking these pics, consider that the power consumes include the 36W of the monitor. So, make your calculations.
Look here:
from 4.7 GHz...
temps06.png


to 3.5 GHz :rolleyes::rolleyes::rolleyes:
temps07.png


to 3.4 GHz :(:(:(
temps08.png


to 3.0 GHz :mad::mad::mad:
temps09.png


and then, after stopping the test, back to idle (1600 GHz) :confused::confused::confused:
temps10.png


Yes, I was wrong.
The throttling isn't a matter of ITX Z68 but also of this bigger sister m-ATX...:rolleyes::rolleyes::rolleyes:
If you see, the power is constant during the test and temps are the same as well; it looks like an automatic protection loop.:eek:

But there is something interesting as well:
Under Furmark the CPU is maxed out at 4.7 GHz without throttling
furmark10.png


Using MSI afterburner I set the fan monitor and I was also able to unlock voltages!:D
furmark11.png

This is a great news; I couldn't change voltages with the older versions of MSI afterburner...
New horizons are opening about VGA overclocking!:cool::cool::cool:

About the significant throttling, I'm not concerned: I'll never get the loads that I got with Prime95.:p:p
Instead, the 4.7 GHz is nicely popping out from the 1.6 GHz idling whenever I open an application or transfer files; moreover, CPU frequencies were stable during Furmark:cool::D:p;).
GOOD!
 
Is that 87C at 3.0Ghz and higher? That's not good. With a AXP-140 on a 2600k I'm seeing 49-51C at 3.4GHz and 59-61C at 4.3GHz in CPUID Hardware Monitor. That's with a non-PWM Noctua 140mm fan on the middle speed (900rpm).

Maybe try applying the thermal compund again?
 
Isn't 87° C a little high? I'd be worried beyond belief if my CPU reached those temp levels.

My second question is, has anyone gotten an discrepancies between what AI Suite and Core Temp/Real Temp/HWMonitor shows? AI Suite shows a temp almost 10° C cooler than the other three programs. AI Suite also matches the UEFI, leading me to believe all the non-ASUS apps are incorrect.
 
Is that 87C at 3.0Ghz and higher? That's not good. With a AXP-140 on a 2600k I'm seeing 49-51C at 3.4GHz and 59-61C at 4.3GHz in CPUID Hardware Monitor. That's with a non-PWM Noctua 140mm fan on the middle speed (900rpm).

Maybe try applying the thermal compund again?

Isn't 87° C a little high? I'd be worried beyond belief if my CPU reached those temp levels.

My second question is, has anyone gotten an discrepancies between what AI Suite and Core Temp/Real Temp/HWMonitor shows? AI Suite shows a temp almost 10° C cooler than the other three programs. AI Suite also matches the UEFI, leading me to believe all the non-ASUS apps are incorrect.
So, the important question here is:
Do I have a bad CPU?:eek:
Let's face it scientifically:eek:
I set back to MoBo to factory settings, and here what I got:
It starts at 53C...
temps12.png


...and achieves 65C!
temps14.png


After a 30 seconds, the CPU fan blows at 100%, thus cools 2C down:
temps15.png


For comparison, the core temps are 6-11C higher, confirming what said by JohnnyGrey86
temps16.png


After stopping, everything goes back to normal:
temps17.png


...and core temps as well:
temps19.png


Well, my CPU is just fine and temps were high because of the overclocking.:p

So, Trevor, I suppose that the temps were at 87C with frequencies of 3.0 GHz because still all the 4 cores were at 100% load. I want to think that it never happens in the real usage, even under load.:rolleyes:
I went back to 4.7 GHz, with 1.375V. No more!
Great questions, anyway, thanks!
 
I still think that's running too hot. Perhaps try reapplying thermal compound?

I've only been a Sandy Bridge user for three days, so I may have no room to talk, but I have a stable 4.4ghz overclock at 1.320v, and AI Suite shows a max around 55° C. This is using a Cooler Master Hyper 212+, with an Akasa Viper fan inside a TJ08E case.

Can anyone else with this, or other ASUS boards confirm the difference between AI Suite/UEFI and Core Temp? This 10° C difference is scary, especially when you don't know which one's right! If Core Temp is indeed accurate, that means you reached a temp of at Least 95° C Mouse, high enough to fry the chip!

I assume all the third party programs are INACCURATE due to AI Suite being the same as in the UEFI (BIOS), but it's still scary nonetheless.
 
First of all the way you applied the TIM is NOT the most effective application method. The best way is to put a small pea sized blob in the middle of the cpu, install the hsf, and let the heat and pressure even out the TIM underneath. Trust me on this one.

And what TIM are you using?
 
geez dude such an effort on such an ugly case.. well, at least it looks like youre having some fun :)
 
Also thought of something, flipping the PSU so it sucks air in from right above the CPU cooler might not be the best idea. It is so close to the fan on the cooler that it might lower the cooler's airflow, due to them both fighting for air.

I think the Armor A30 is am ugly case, and they tried tackling the airflow problem of the LanBoy's by smacking a fan on top. The LanBoy's looked amazing, however :)
 
First of all the way you applied the TIM is NOT the most effective application method. The best way is to put a small pea sized blob in the middle of the cpu, install the hsf, and let the heat and pressure even out the TIM underneath. Trust me on this one.

And what TIM are you using?
Thank you for pointing it out.
I read and thought about the best way to apply thermal compound...
This review from Hardware secrets
What is the Best Way to Apply Thermal Grease?
clearly shows 1C difference between one single dot and spread:
thermalcompound.gif


I'm not really concerned about this anyway, as you can see here, this method gave my a nice paste spreading:
011yjq.jpg

011xxh.jpg


I used Arctic Silver 5.

 
Last edited:
I still think that's running too hot. Perhaps try reapplying thermal compound?
Thanks for the advise, I'll try it as latest attempt, I feel comfortable with my paste spreading.

I've only been a Sandy Bridge user for three days, so I may have no room to talk, but I have a stable 4.4ghz overclock at 1.320v, and AI Suite shows a max around 55° C. This is using a Cooler Master Hyper 212+, with an Akasa Viper fan inside a TJ08E case.
It can be for 2 reasons:

  1. Lower voltage than mine (I went back to 1.350 though)
  2. Better case cooling (this Armor A30 is really challenging in this context)

Can anyone else with this, or other ASUS boards confirm the difference between AI Suite/UEFI and Core Temp? This 10° C difference is scary, especially when you don't know which one's right! If Core Temp is indeed accurate, that means you reached a temp of at Least 95° C Mouse, high enough to fry the chip!

I assume all the third party programs are INACCURATE due to AI Suite being the same as in the UEFI (BIOS), but it's still scary nonetheless.
I tend to agree with you.
I read some reports stating that Core Temp overestimates temps and that single core measurements aren't important as CPU temp.
 
Last edited:
geez dude such an effort on such an ugly case.. well, at least it looks like youre having some fun :)

Also thought of something, flipping the PSU so it sucks air in from right above the CPU cooler might not be the best idea. It is so close to the fan on the cooler that it might lower the cooler's airflow, due to them both fighting for air.

I think the Armor A30 is am ugly case, and they tried tackling the airflow problem of the LanBoy's by smacking a fan on top. The LanBoy's looked amazing, however :)
I accept both of your opinions, and I thank you for it, but I don't like very much the "ugly" adjective...:mad:
It's quite offensive...

Anyways, I flipped back to PSU so it doesn't suck air from the CPU area, I agree with you here.
 
Last edited:
After all the concerns that arose for the temps, I decided to face seriously the cooling with the Armor.
I was able to transform the scaring 87C with massive throttling...
temps251345v.png


...to 77C with no throttling at all!!!
temps28newfans.png


How?:confused:
I got rid of the ridiculous 6cm fans (silent, sure, but useless, no air moved)
009ntc.jpg


...with these MONSTERS:
007xi.jpg


These are a couple of SUNON PMD1206PMB1-A, from the Egg.
They make an incredible 56.5 CFM.
The only issue?
RPM: 8000 RPM
Noise Level:56 dBA

Fortunately, they are connected to a MoBo fan header that lows them down at 40% speed, still noisy though, but better than the hairdryer effect at max RPM:eek:

Just for comparison:

014rvg.jpg


Compare them mounted:

013uug.jpg


The final result:
018ody.jpg

 
Last edited:
I also changed the other USELESS Thermal Take fan that was in the front as intake:
008jdz.jpg


...with another 14cm PWM Scythe fan, connected as second CPU fan in the MoBo, so it will ramp up along with the CPU fan:
016fez.jpg


Thus, I also got rid of the 12cm PWM Scyte that was in the middle and of the HD cage as well (I mean, there was something wrong having the HD cage in front of the intake in this way):
017pi.jpg


The installation of this fan needed some conciliation with the mounting holes...
018uks.jpg


I was able also to mount the upper drive cage, though, after using the Dremel to have the fan fitting in...
Here a pic of the trim (before smoothing it):
029hwz.jpg


In this way, I put the Crystalfontz controller, one 1TB HD in the side and another suspended with two rubber bands (the SPCR way).
 
Last edited:
Finally, I changed the big 20cm top fan (that mysteriously is indicated as 23cm in the TT site) with a XIGMATEK XLF-F2006 blue:
001gob.jpg


It went as intake, instead of the original exhaust:
027avl.jpg


The reason for the intake is that the 6cm fans, now, are more than enough for exhaust, so this intake could also help the VGA cooler.

To be honest, I was disappointed by this fan since it moves the exact quantity of air than the original one (they are identical!); but I can at least use the 3-pin header to control it via the MoBo.
 
...with these MONSTERS:
007xi.jpg

I could see what these things would sound like just from reading the label. 10.6W! Good thing they move a lot of air because they may well be heating some of it on the way! :D

Cases are personal things. Some people think a Hummer is ugly, others find it inspiring.
 
I could see what these things would sound like just from reading the label. 10.6W! Good thing they move a lot of air because they may well be heating some of it on the way! :D

Cases are personal things. Some people think a Hummer is ugly, others find it inspiring.
You bet... these fans are insane!:cool:
I'm looking for a good fan controller that can handle the wattage, best if mounted in a PCI spot.
I was thinking about the Sunbeam 20W Rheosmart PCI Slot Smart Fan Controller - (PL-RS-PCI).
 
and i thought the delta black label 60mm fans were crazy loud but this is just insane.


Jen
 
and i thought the delta black label 60mm fans were crazy loud but this is just insane.


Jen
Jen, you are right. There is a jet taking off every time I boot the computer, then the MoBo slows them down to 40% RPM.
I'm looking for a controller that can set lower than 40% because they are still pretty noisy and that can handle more than 10.6W per channel; the Sunbeam PL-RS-PCI is not the one... it gives 6V minimum and "only" 20W total.:mad:
I'll look for something different (a 0-100% V regulator).
Any advise?
 
Last edited:
Anyways, these little fans are a FUN!
I like them, they keep the CPU cool during heavy loads, so I'll keep them!
 
A 15 ohm 10 watt power resistor should drop the fans to 40 or 50%. Just make sure its not placed in the airflow as it will heat up.
 
A 15 ohm 10 watt power resistor should drop the fans to 40 or 50%. Just make sure its not placed in the airflow as it will heat up.

or do a 5 volt mod on the fan to reduce voltage as well , that may make fan quiet and still enough air. i have a 7 volt resistor that i purchased i think from svc that does well for this 92mm fan that is insanely loud when at 12 volt

i have one fan controller that does go to 0 volts dont know what it max amp rate is due to it being no brand name. bought it used several years ago


Jen
 
Yes that would also work. Attach fan ground to the 5V rail (via a 4 pin molex) and the positive to the 12V rail should give about a 50% reduction (7V) in fan speed. The problem with both of these solutions is that some fans won't spin up unless the starting voltage is quite high. 7V should be OK but 5 or 6 might not make it.
 
just curious, any estimates on how much $$ in fans you have spent? Seriously...wow.
 
just curious, any estimates on how much $$ in fans you have spent? Seriously...wow.
AH, AH, AH!!!:D
You are right.
I am a FANIAC!:p
You cannot imagine how many fans I have sitting around unused (or stored in my closet).:D
I just cannot help myself...
I can tell you, though, that I get the best prices out there: I got the 14cm PWM Scythe, as an example, for $11; the 12cm PWM Scythe for $7; the 20cm Xigmatek for $20 with $10 MIR...
Anyways, a bunch of many...:(
 
A 15 ohm 10 watt power resistor should drop the fans to 40 or 50%. Just make sure its not placed in the airflow as it will heat up.

or do a 5 volt mod on the fan to reduce voltage as well , that may make fan quiet and still enough air. i have a 7 volt resistor that i purchased i think from svc that does well for this 92mm fan that is insanely loud when at 12 volt

i have one fan controller that does go to 0 volts dont know what it max amp rate is due to it being no brand name. bought it used several years ago


Jen

Yes that would also work. Attach fan ground to the 5V rail (via a 4 pin molex) and the positive to the 12V rail should give about a 50% reduction (7V) in fan speed. The problem with both of these solutions is that some fans won't spin up unless the starting voltage is quite high. 7V should be OK but 5 or 6 might not make it.
I already have the fans at ~40%, but their 3800RPM are still too noisy to me.:(
I also want to keep the potentiality to max them out in case of necessity; maybe it will never happen, but it would be a shame to have purchased these tornadoes and be forced to don't use them at max if necessary.

Thus, I'm going to try a Lamptron FC-FC4-B:
4 channels
20W/channel
I'll keep you posted!;)
 
Back
Top