ASRock Z68M-ITX/HT

Hi everyone,

I have the Asrock board too with a 2500K CPU. I managed to overclock the CPU until 4.7GHz. All the stability tests (LinX, HyperPi), applications (BOINC, etc), games (Supreme Commander 2, Warcraft3, etc) I used showed that it was stable.
Unfortunately I have the same issue as everybody here concerning the CPU throttling... Even with a Noiseblocker S4 blowing on the mosfets, the 2500K still throttles when in full load, which is not ideal when testing stability!

I will try to add radiators on the mosfets and see if I can get rid of the throttling, although I am not very confident...

I hope that Asrock will release a bios giving us control over this "protection" and a few voltage options.

The remote control requiring a specific USB wiring on the board is also another aberration...
 
Hi everyone,

I have the Asrock board too with a 2500K CPU. I managed to overclock the CPU until 4.7GHz. All the stability tests (LinX, HyperPi), applications (BOINC, etc), games (Supreme Commander 2, Warcraft3, etc) I used showed that it was stable.
Unfortunately I have the same issue as everybody here concerning the CPU throttling... Even with a Noiseblocker S4 blowing on the mosfets, the 2500K still throttles when in full load, which is not ideal when testing stability!

I will try to add radiators on the mosfets and see if I can get rid of the throttling, although I am not very confident...

I hope that Asrock will release a bios giving us control over this "protection" and a few voltage options.

The remote control requiring a specific USB wiring on the board is also another aberration...

Are you sure throttling is a direct result of mosfet temps? For all we know mosfet temps have nothing to do with the board throttling. If throttle stop works then that may be the only solution for these boards.
 
I cured my board's throttling with mosfet heatsinks. I was getting throttling at stock speeds with Prime95 on my i7-2600K.

I got two of the Enzotech MOS-C1 kits for a total of 20 heatsinks. I permanently mounted them to the board with nonconductive thermal epoxy. I covered every mosfet and used the leftover heatsinks on any chip that was about the same size.
 
I cured my board's throttling with mosfet heatsinks. I was getting throttling at stock speeds with Prime95 on my i7-2600K.

I got two of the Enzotech MOS-C1 kits for a total of 20 heatsinks. I permanently mounted them to the board with nonconductive thermal epoxy. I covered every mosfet and used the leftover heatsinks on any chip that was about the same size.
So to clarify now you can maintain a solid 3.8Ghz (IE-Turbo) and it never throttles to 3.4GHz?
 
Trevor - the throttling I was getting was all the way down to 16x with all four cores loaded with Prime95.

I am not overclocking or anything so I wasn't expecting it to stick at 3.8. Or is it supposed to??
 
Ahhh...OK. With OE settings they usually won't stay in turbo mode (3.8GHz) for an extended amount of time (AFAIK I think that's due to the UEFI current protection). You should be able to get it to stay at 3.4GHz though.

I asked because I thought it would have been weird if it stayed at 3.8GHz.
 
So, to put the things clear. Apart for stress tests like Prime95 and Furmark, it doesn't throttles at all at day to day aplications (games)??

My first Asrock was defective (some USB's didn't work properly. I've give it a second chanse as it only cost me 5 more euros than the GA-H67n

Wish me luck!
 
Trevor - the throttling I was getting was all the way down to 16x with all four cores loaded with Prime95.

I am not overclocking or anything so I wasn't expecting it to stick at 3.8. Or is it supposed to??

3.8GHz is the Turbo speed for when only one core is active. For four cores active I think Turbo is only a 100MHz bump over the stock speed(AKA 3.5GHz), and that's assuming the CPU is cool enough. What heatsinks do you guys have on the CPUs that are throttling?
 
My CPU is watercooled and doesn't go over 65°C on LinX, so CPU temperature isn't responsible of throttling in my case.

It even throttle with BOINC whereas the CPU temperature is around 55°C...
 
My board just died. In Prime95 I usually run "In-place large FFTs" but last night I decided to run "Blend" to stress the memory and it BSOD'ed after about 4-5 minutes. At the time my 2600k had settled into running at 3.4Ghz (occasionally recovering to 4.3Ghz). When I restarted the comp after the BSOD it wouldn't POST. I disconnected everything and removed the RAM and still no POST. Today I swapped in the old ASRock H67 board (with the Z68's RAM) and it cranked right up.

I've got a new board on the way (Along with an AXP-140 to replace the Scythe Samurai). Not sure if I'm going to be too quick to run the blend test in Prime95 ever again.
 
Wow, it's a pretty long thread. Is it safe to assume that this is the best Z68 ITX board at the moment? Should I be able to get a decent overclock with an Intel i5 2500k?
 
Can someone also tell me what the significance of the CPU throttle that the board does for the overall experience?
 
Also, anyone know the voltage range for the memory? Can it run the low voltage 1.35v stuff, or is it pretty much limited to 1.5v?

Anyone have time to check this? There are some nice 1.35 volt RAM kits out there if the board supports them.
 
It means if it gets over it's specified TDP and power limits it will throttle down to protect the CPU and the Mobo. I guess 4.2Ghz should be doable
 
I'm back up and running with the new Z68 board. I transfered over my old MOSFET coolers and added a few but I couldn't hit two of the CPU power MOSFETS due to the installation fixture of the new AXP-140 cooler. If I was at home I would have modded the heatsinks to make em shorter but I'm out on the road. The new board came with UEFI v1.20 already installed. After the 2nd boot I went straight to the "Turbo 4.3GHz" (1.32v peak) setting and 2133MHz (1.54v) on the RAM using the built in XMP profile.

The AXP-140 was a bear to install in my SG07 case but it cools much better than the Scythe Samurai. I installed a Noctua 140mm on it. It's not a PWM fan so I started at full speed (1250-ish) and you could hear the fan outside the case. I installed the first adapter to cut the speed to 890 and it's virtually silent now and the temp went up like 1C under load so no biggie. I think it's a hair louder than the Scythe Samurai (PWM) at idle but it's definitely quieter than the Samurai at full tilt. It idles at 27-29C and peaks out at 56-60C vs 29-32C & 67-71C before.

Can someone also tell me what the significance of the CPU throttle that the board does for the overall experience?
For some people their "experience" includes knowing that their setup can ace Prime95 while overclocked...if that's you then you won't be pleased with this board (and you probably need to step up to something bigger than Mini-ITX as of right now). If you're really after performance in games & real-world applications then you should be happy.
 
Ok, thanks. Those are some weird values, they're all just a bit off from what the DIMMs usually are.
 
For some people their "experience" includes knowing that their setup can ace Prime95 while overclocked...if that's you then you won't be pleased with this board (and you probably need to step up to something bigger than Mini-ITX as of right now). If you're really after performance in games & real-world applications then you should be happy.

If the motherboard throttles under those circumstances, then how will I know if the overclock is successful or not? I won't be able to test for stability.
 
I'm back up and running with the new Z68 board. I transfered over my old MOSFET coolers and added a few but I couldn't hit two of the CPU power MOSFETS due to the installation fixture of the new AXP-140 cooler. If I was at home I would have modded the heatsinks to make em shorter but I'm out on the road. The new board came with UEFI v1.20 already installed. After the 2nd boot I went straight to the "Turbo 4.3GHz" (1.32v peak) setting and 2133MHz (1.54v) on the RAM using the built in XMP profile.
So, no throttle now @4.3GHz, with mosfet heatsinks and axp-140? Did you tried any higher, maybe 4.5GHz?
 
If the motherboard throttles under those circumstances, then how will I know if the overclock is successful or not? I won't be able to test for stability.
Well...if it's a desktop doing things that aren't critical just put it thru it's paces using normal programs/games for a few days and see if it has any issues (or if it throttles). That's the way we used to do it before they came up with all of the stability/burn-in programs. If you're using it to control a nuclear power plant's cooling system while overclocked I suggest you buy a different board. :D What stinks about the "overclock and just use it" method is it takes a LONG time to ensure that each setting is good before you can move ahead and push it harder.
So, no throttle now @4.3GHz, with mosfet heatsinks and axp-140? Did you tried any higher, maybe 4.5GHz?
I haven't tried higher. I've been running alot of real world apps/games that lean on the CPU pretty hard without it throttling from 4.3 to 3.4 (or having one single issue) but I'm thinking there isn't a whole lot of overhead when it comes to the motherboard. I may be conservative but I bet there's a reason ASRock chose not include a "Turbo 4.4Ghz" setting in the UEFI. That's not to say I'm not curious though...I may give it a shot later.
 
Just for more info for everyone, I added the Enzotech MOSFET coolers to my VRM on the CPU only. It roughly tripled my time until it throttles back at 4.4GHZ (about 90 seconds) I am using the Antec Kuhler 620 in push/pull in an SG-05 with a 2500k.
 
The only way you guys are going to stop the throttling is with software like throttle stop. You can throw a hundred pounds of heatsinks at the mosfets and you're only gonna buy a few seconds. I don't know why nobody will atleast give it a try.
 
I didn't read this thread, but my XFX 6970 does not work in my asrock z68 mini-itx, but works in all other computers.

I am stumped at this point.
 
The only way you guys are going to stop the throttling is with software like throttle stop. You can throw a hundred pounds of heatsinks at the mosfets and you're only gonna buy a few seconds. I don't know why nobody will atleast give it a try.

Doesn't work. I'm so pissed that the last generation Intel chips have better overclocking ITX boards. I'm stuck with my H61 ITX board because there isn't a Z68 ITX worth buying.
 
I have installed a radiator on the mosfets (it used to cool an old Radeon mosfets). The radiator itself is cooled by a Noiseblocker S4 fan. As expected the CPU still throttles although a little bit less than before. I have sent a mail to Asrock to know if they plan to do something about it (new bios for example). I hope I will have an answer soon.
 
I wonder where the actual temperature sensor is that is causing this throttling. Maybe a heatsink on that will help.
 
Well, I can also confirm that the AXP-140 works on this board with a discreet video card. Pretty damn close though. :eek:

On a related note, has anyone gotten the automatic CPU fan control working with a three pin fan? The 140mm the AXP-140 comes with screams along like a hair dryer at 1500RPM, so I changed the fan control from "Max on" to "Automatic" with a target temperature of 50c. The CPU is idleing at about 38c, but the fan is still running at full tilt. Also tried turning the maximum allowed fan speed from setting 9(full speed) down to setting 5(which must be half-ish). Still no adjusting of the fan as far I can tell. I'm hoping it doesn't require a PWM fan to work. :rolleyes:
 
Well, I can also confirm that the AXP-140 works on this board with a discreet video card. Pretty damn close though. :eek:

On a related note, has anyone gotten the automatic CPU fan control working with a three pin fan? The 140mm the AXP-140 comes with screams along like a hair dryer at 1500RPM, so I changed the fan control from "Max on" to "Automatic" with a target temperature of 50c. The CPU is idleing at about 38c, but the fan is still running at full tilt. Also tried turning the maximum allowed fan speed from setting 9(full speed) down to setting 5(which must be half-ish). Still no adjusting of the fan as far I can tell. I'm hoping it doesn't require a PWM fan to work. :rolleyes:

It needs a PWM fan.
That's correct!
I'm using this:
matx07.jpg


matx12.jpg
 
Damn, I was afraid of that. :( The only reason I had any hope for the three pin CPU fan being controllable is because the board offers the exact same controls for the chassis fan, and the chassis fan header is only a three pin.

I was actually thinking about getting that PWM Scythe fan if this didn't pan out. :) They have a reputation for decent sound quality. Did you have to move the fan clips to make the 120mm mounting holes work? I'd literally have to rip this thing to pieces to get them out again, it's just that tight in a SG07.
 
Damn, I was afraid of that. :( The only reason I had any hope for the three pin CPU fan being controllable is because the board offers the exact same controls for the chassis fan, and the chassis fan header is only a three pin.

I was actually thinking about getting that PWM Scythe fan if this didn't pan out. :) They have a reputation for decent sound quality. Did you have to move the fan clips to make the 120mm mounting holes work? I'd literally have to rip this thing to pieces to get them out again, it's just that tight in a SG07.
The good of these 14cm PWM fans from Scythe is their efficiency:

  1. Good air flow at low RPM
  2. When at max, better than may others at the same RPM, with a lot of air
No problem with the clips, they fit perfectly!
Just keep in mind that the fan has very long cables:

  • 1 with the 4-pin PWM
  • 1 with the plate for PWM/VR switch, to be inserted in the back of the case (in a PCI slot)
I got them very cheap ($11.99) from AntaresPro.
 
Doesn't work. I'm so pissed that the last generation Intel chips have better overclocking ITX boards. I'm stuck with my H61 ITX board because there isn't a Z68 ITX worth buying.

What's the point of a Z68 board that doesn't overclock? Ssd caching? My Amd Thuban mitx rig is most likely faster in multithreaded programs than these sandy bridgez68 mitx systems and that's really weak considering how long ago I built it.. I have more respect for gigabyte and Asus for not releasing mitx Z-68 motherboards altogether instead of releasing an overpriced and gimped z68 mitx motherboard that can't sustain an oc for more than a few minutes. I really doubt adding heatsinks and bigger coolers will remedy the issue long term.
 
Last edited:
Mine overclocks just fine and never throttles in games, real world apps, and "normal" benchmarks like 3DMark and such. It only throttles when 100% loaded on all 4 cores/8 threads for extended periods. Nothing like using a system the size of shoebox to play BF3 with everything but MSAA on & maxed out at 2560x1600.

I've got to say though...if your idea of a good time is watching intently as Prime95 runs then this is probably not the board for you.
 
My post wasn't meant as a flame, if you saw it tthat way then I apolagize. I' m just dissapointed that there isn't a sb mitx motherboard that can oc and that sucks because I really wanted one.
 
I have been waiting for a decent overclocking itx board forever. It seems like even with the throttling that this is still the best choice. I am really tempted to buy it and make a custom water block for the mosfets. in the name of science of course. Lol
 
To be clear - I am disappointed that the board is pretty much incapabable of supporting a 100% loaded overclocked 2600k for a significant amount of time. Even though I regularly log the multiplier while gaming (and never see an issue after the fact) it's always in the back of my head. Is it throttlling? Is that all she's got captain? Kind of lame.

But at the end of the day it is the fastest 1155 Mini-ITX board you can get when it comes to actual/normal use. I am amazed at how fast this little thing is & I wouldn't trade it for anything else in this form factor that's on the market right now. But I can tell you that I won't hesitate to replace it when a new option is released.
 
Back
Top