Help with Intel Bad Axe 2 and e6600 overclock

S

slappynutz

n00b
Joined
Jan 30, 2007
Messages
61
Sorry if I'm asking something that's been answered a million times before. I've done tons of searches and I must be using the wrong keywords.

I bought an Intel Bad Axe 2 and C2D e6600 never really intending to overclock much, but now that I've got everything up and running I'd like to see what the chip can do. The only problem is that a lot of the steps suggested in the e6600 overclock guides seem to be missing in my Bad Axe 2 bios ... things like Intel Speed Step and the like.

So if anyone has the same combo as I do, and has successfully overclocked it, I'd really appreciate some help.

My setup:

Intel D975XBX2 Bad Axe 2
Intel Core 2 Duo e6600
4 Sticks of OCZ Platinum Revision 2 DDR2 800 (PC2 6400)
eVGA 8800 GTX
Scythe Ninja
Corsair 620W P/S

Everything runs rock-solid with the following Bios Settings, which are pretty much default except for the Memory timings, which I set manually as SPD didn't register the correct values automatically. With that setup in my apartment with an ambient temperature of 26C, I have idle CPU and MB temps in the low 40s that spike up to the low 50s under load. I've heard you can push the e6600 without tweaking the voltage significantly so I don't expect heat to be a big factor. I can always just turn the low Nexus fans I have up higher.

Processor:

CPU Voltage: 1.3250V
Enhanced Power Slope: Disabled
MCH/ICH Volt Override: 1.525
Host Clock Frequency: 266

Memory:
Reference Frequency: Default
Memory Frequency: 800mhz
tCL: 4
tRCD: 4
tRP: 4
tRASmin: 15
Memory Voltage: 1.96

Bus:
PCI Express Burn-In Mode Percentage: Default
PCI Latency Timer: 32
HPET: Disabled


If there's a Bad Axe 2 overclock guide somewhere that I've missed, or some hidden settings as described in the existing 6600 guides that I haven't found, I'd really appreciate the help.

Thanks a lot.
 
Set the RAM frequency to 533mzh so that it runs in par with the FSB frequency.

Make sure that the ram voltage is set to the recommended setting (manufacturer's suggested voltage) Also use the recommended timings.

Disable c1e, speedstep and thermal control (if available)

Enable CPU power slope

Basically, your CPU has a default multiplier of 9. Multiply that by your host frequency (266mhz) and you have a 2.4ghz E6600 chip.

To overclock, simply change the Host frequency to 300mhz and that X9 = 2.7ghz

Your RAM should run 2X as much as your host frequency (300 X 2) = 600mhz.

Next, try 350 host frequency for a 3.15ghz overclock.

If it works, great, if not, try to change your CPU Voltage to 1.4V or so.

Don't exceed 1.55V on air, or 60c during load.
 
Thanks for the reply.

I did exactly what you suggested, and even with a minor overclock of 300mhz my system fails to post. I suspect it's in the c1e/speedstep/thermal control ... which might be called something different in the Bad Axe 2. Must be missing some strange setting somewhere.

One thing I notice is that no matter what I set the ram frequency to it's still 800mhz when I return from the crashed post. Really odd.
 
Lot of info in that thread. I was hoping to do some overclocking without ramping the voltages that high. From what I understood on other guides I could ramp up the processor mhz without messing with the voltages. Going to try loosening the memory timings (leaving them at the 5-5-5-15 default and upping the voltage a bit) ... I just can't get anything to post at all on stock voltages.
 
Here's what I try.

First thing is I enable the CPU Power Slope. Save and Reboot.

Then I set the Ram Voltage to 2.04, up from 1.84 (which is the default and honestly seems like an undervolt ... my OCZ memory is rated from 1.9 to 2.2), decrease the Reference Frequency to 266mhz and the Memory Frequency to 566Mhz. Save and reboot. Loads to Vista fine and reports the correct numbers.

Then I go back in and, not touching any of the CPU voltages, up the Host Clock Frequency to 275. Post fail, and when I go back into the BIOS the Memory Frequency has defaulted back to 800Mhz.

So I start from scratch again. This time I bump up all the processor voltages (cpu, FSB and MCH) a single notch. I try 275 again, and it posts and loads to windows. Go back in and try to push to 285, and I get another failed post.

Pretty sure all my problems are knowing what voltages to set from the start. I've just read so many posts from people claiming to get 300mhz from stock voltages I'm surprised I can't even get 275.

Does anyone have the D975XBX2 and an e6600 who can tell me the precise settings they're using? The linked thread suggests a cpu vcore of 1.5 to start and I'm just not comfortable going that high, especially when people claim being able to clock higher without even touching the voltage.
 
Are you going into bios with the recovery jumper to the recovery position and turning off the c1e and eist (one or both can only be found in recovery, my memory sucks). Turn on the enhanced power slope. Use the 266 strap (find the memory calculator in the thread posted, that helps tons to figure out what you are doing...aww heck, I'll link it... here.

Manual override the CPU voltage, even if you are not raising it. My 6600 will do 3.4 on the stock voltage, but I think mine is a good one from what I have seen. You should probably get at least to 3.0 or 3.2 on stock volts. Beyond 3.4 I needed to up the CPU to 1.3875 where it is now for 3.6 Orthos stable. Also, beyond 3.4 the board needs the FSB and MCH upped to 1.3 and 1.6 respectively, or at least my board did. I can do 3.7 at 1.41ish, but thats going beyond the TAT temps I like, so I stopped at 3.6.

Now if you have one of those "F" processors (or is it D?, memory sucks again) as in LxxxFxxx, that might be a problem. I've been hearing those are real pains to OC if at all possible. Hopefully your cpu is LxxxAxxx or LxxxBxxx.

Good luck.
 
Thanks for the tip about the recovery jumper ... I'd been looking for the c1e setting everywhere. Went and disabled it and hoped everything else would fall into place.

Unfortunately, still, nothing will overclock. My settings in recovery console:

Intel Enhanced Debug: Disabled
C1E: Disabled
Failsafe Watchdog: Enabled

Then I also disabled all the fan controls, as well as EIST in the main bios.

I set my memory voltage to 2.04 and left the timings 5-5-5-15, then clocked it down to 266/533. Rebooted. Everything was fine.

I then set my CPU at 1.35V, my FSB at 1.25, and my MCH at 1.525. Rebooted. Everything fine.

So then I set the Host clock at 280, up from the stock 266.

Post fail. I checked my C2D box and my Batch # is LXXXAXXX packed on 1/2/7.

Really, really, wish I could figure out what's going wrong. It feels like the memory clock is still linked to the FSB, regardless of whether or not I set it to override to 266/533. As soon as I overclock the FSB the post fails and when I'm kicked back to the BIOS the memory clock is back at 800. Annoying. There must be some other setting I'm missing.
 
Disable the Failsafe Watchdog, it's the main cause of limited overclocking since it's too sensitive.
 
Thanks. I had a feeling I needed to disable the Watchdog but didn't want to change settings without knowing exactly what they are. Will try again after work tonight.
 
Thanks again for the help guys.

Tonight I set my ram to the OCZ-recommended timings (4-4-4-15 at 2.08V), ran some tests which passed, then dropped them to 1:1 at 266/533. Disabled EIST, C1T, and most importantly the Failsafe Watchdog setting, which is a new "feature" in the Bad Axe 2 maintenance bios. Ran some more tests, and it seemed fine.

So, at stock voltages, I upped mhz to 300 from 266. Booted straight into Windows. Running OCCT right now, with my ultimate goal of a modest overclock of my e6600 to 3.0ghz.

Also solved the strange Bios freezing issue ... seems the Bad Axe 2, at least with the current Bios, has problems with a couple of its USB ports. Found a solution on another forum where I switched the keyboard to the lower USB port under the ethernet port ... no lockups. Have had problems with both my Dell 2407FP's usb hub and the keyboard on some specific ports.

Thanks again. Seems like I'm on my way.
 
Not to completely spam the overclocking boards, but just wanted to report back that I'm running Vista 64 stable at 3.0ghz on stock voltages. So I'm happy. The silly Windows Experience Index jumped from 5.3 to 5.7 and my memory went from 5.7 to 5.9.

I've got:

Processor 5.7
Memory 5.9
Graphics 5.9
Gaming 5.9
Hard Disk 5.6

Guess Vista isn't too impressed by my old single-drive 74GB Raptor. Oh well. :)

Thanks again.
 
I have no idea what the Windows experience is...Vista scares me, lol

But why stop at 3.0? You have a big cooler and hopefully it is flat and hopefully your IHS is flat too (reseach lapping CPU's and heatsinks), see where it can go on stock volts! 3.2? 3.4? Even faster? These C2D's are truly amazing. Just make sure to download a copy of Orthos to load your CPU, and download the Intel Thermal Analysis Tool to monitor CPU temps. Intel Desktop Control Center is good too to play with case/CPU fan settings by temperature too. Enjoy!
 
I like Vista so far, but most of those Intel apps don't work with it yet. I use Everest to check temps, which seems to play nice so far. Haven't found a version of Orthos which runs in Vista 64 yet, so I've been using OCCT ... almost done with hour two at 3.0ghz ... seems rock solid so far.

I'll probably push it a little bit further after I've given it a couple days at 3.0. Now that I've got it running well at stock I don't want to have to crank the voltages much at all. Quiet is more important to me in the long run, so I don't want to generate any extra heat, beyond what I'm already adding by disabling speed step.
 
Went to 3.15 stock and everything was solid ... jumped up to 3.3 and blue-screened on my way into windows. Might be a ram timing/voltage issue at that point, or the FSB needs a bit extra juice. Dunno. I have every voltage set at either default or the lowest possible when I don't know the default. Could probably hit 3.3 with some minor tweaking but will settle with 3.2 for the time being. At 3.2 my "Windows Experience Index" jumps up to 5.8 of a possible 5.9, and runs stock at the lowest settings across the board.

Would kind of like to hit 3.3 now but I'm content with 3.2. It's more than I thought I'd be able to get anyway. :)
 
So you got to 3.1ghz with no Voltage changes?

Very lucky. Looks like you have a good CPU. Which means that you have to push it more:D

Set the CPU Voltage to 1.4V and your FSB to 1.4V or so.

Try a 401 FSB for a 3.6ghz overclock ans see if your PC can take it!

Let us know!
 
After tons of testing I've actually got an E6600 that's stable on stock voltages across the board at 3.2ghz. I thought there was a crash happening in Half-Life 2 due to OC'ing, but it was just the Vista 64 AA bug ... random hl2.exe crashes if AA is enabled. Disabling it makes everything stable at 3.2. OCCT has passed 6 times, twice on 1-hour cycles, so I'm pretty sure it's good.

The only way I can get the chip to report a problem is at 3.3ghz at 1.3250V stock. It loads into windows fine and passes Memtest86+, but OCCT finds a crash during its 30min test. My problem before was that I'd set the Core Voltage to default, and the FSB and MCH to the lowest possible. Once I reset the bios to default and realized they should be at 1.250 and 1.525 respectively, 3.3 was a lot more stable. Only OCCT could bring it down.

Weird thing happened with Everest in Vista 64 though. I had it open to monitor temps and found it wildly inaccurate. It would register low 30s during regular testing, but spiked up to 61 during an OCCT stress test. It freaked me out and I stopped OCCT immediately ... which made Everest drop its temp reports down to 31. 30 Degree drop in 3 seconds. Not really an expert in thermodynamics, but I think that's a mistake.

Dropped straight to the bios, which reported a CPU temp of 48c. Pretty sure my build idles in the high 30s, and hits the high forties/low fifties under heavy load. Until a reliable temp monitor for Vista 64 is available, I'm just going to have to do the drop-to-bios-check to look at temps. For a Scythe Ninja operating fanless in a room with an 80 degree ambient temp, that's pretty good.

Going to see if I can hit 3.3 by simply bumping the core voltage to 1.3375. Pretty sure it will go a lot higher, but I'm taking it slow. Besides, with my 8800GTX on a 24 inch LCD that I always activate v-sync on, the difference between 3.2 and 3.4/3.6 isn't really going to be major. Good to know I'll be able to crank more power out of this thing down the road.
 
Hope no one minds if I ask one more question.

Was waiting for CPU-Z to be released for Vista 64 before I really started to tweak my system further, and now that it has I'm having some strange readings. Despite disabling C1E, Failsafe Watchdog and EIST, it seems as though speedstep is still in effect.

No matter whether I leave the Core Voltage at default or 1.4, CPU-Z reports the core voltage as 1.213, with occasional drops to 1.050. The Core Speed jumps between 2200 and 3300 (which I set in Bios). OCCT passes all tests at 3.2 ghz, but as soon as I jump up to 3.3, regardless of increased voltage, it fails. I'm assuming it's because (somehow) speedstep is still kicking in.

Considering updating to the newest Bad Axe 2 bios to see if that changes anything. Suggestions?
 
Don't worry about CPU-Z reading the voltage wrong, it's a know problem after 1.4V or so.

But I don't understand the speed changes...

You should try the new bios...
 
I installed the new bios and did all the new settings ... plus I installed the new SpeedFan. The plus is it seems to read the correct core voltage ... about .05 below the bios-set voltage. So at least that works. However I still can't manage to get 3.3 even at 1.4v ... must be some other issue like the FSB or ram.

Played some Half-Life 2 and though the CPU temp never spiked above 42, the Internal and Motherboard temps shot all the way up to 50/51. Strange ... must be the bridge. Haven't even tweaked the voltage to them but they must be running hot with the 3.2ghz overclock. Half-Life 2 seems to be having problems with the 8800 ... save game icons eventually corrupt and the game will crash when I've got AA enabled. Hope it's just a driver issue.

Kind of worried about the M/B temp ... it averages about 8c higher than the CPU idle, and about 10c higher under load. Might need to add a fan or something.
 
Hey slappynutz and guys that helped him. I have the same mobo and cpu and I applied alot of the things that you said for him to do and I have a couple questions. I set the host freq to 300 and the ram freq to 533 and I disabled almost everything except for speedstop and the failsafe watchdog which I couldnt find. Does anyone know exactly what those are called or where they are on the bad axe 2? Now the first time it booted up it hung at the "welcome to windows" or whatever when its loading to the desktop for like 10-15 secs and when it eventually got to the desktop i had no internet. So I restarted my pc and and it didnt hang for as long but a little more than normal and I have internet now. Does nayone have any suggestions for me from what I've said? BTW, I am at 2700 mhz right now and I'm about to start ORTHOS.
 
Not to change the subject, but I find that Orthos doesn't really stress my QX6700 enough. When running Orthos, WindowsXP Task Manager shows only a 100% load on two out of the four cores.

The program that really stress out all four cores is OCCT. I'm able to boot into WindowsXP running the QX6700 at 3.5 to 3.6GHz, but OCCT fails after 30seconds. I don't believe I get any failure with Orthos at 3.5~3.6GHz.
In addition, my CPU with 100% on all four cores jumps to 80C.

Now, I'm just running at 3.15GHz with the lowest voltage settings in the BIOS for everything under the Processor menu.
 
JimmyNeutron said:
Not to change the subject, but I find that Orthos doesn't really stress my QX6700 enough. When running Orthos, WindowsXP Task Manager shows only a 100% load on two out of the four cores.

The program that really stress out all four cores is OCCT. I'm able to boot into WindowsXP running the QX6700 at 3.5 to 3.6GHz, but OCCT fails after 30seconds. I don't believe I get any failure with Orthos at 3.5~3.6GHz.
In addition, my CPU with 100% on all four cores jumps to 80C.

Now, I'm just running at 3.15GHz with the lowest voltage settings in the BIOS for everything under the Processor menu.
Click to expand...

Well I dont have 4 cores, just 2, and it always has both stressed to 100%, so I guess its good for dual core cpus.

Anyway, does anyone ahve any tips or suggestions for what I said before?
 
JimmyNeutron said:
Not to change the subject, but I find that Orthos doesn't really stress my QX6700 enough. When running Orthos, WindowsXP Task Manager shows only a 100% load on two out of the four cores.

The program that really stress out all four cores is OCCT. I'm able to boot into WindowsXP running the QX6700 at 3.5 to 3.6GHz, but OCCT fails after 30seconds. I don't believe I get any failure with Orthos at 3.5~3.6GHz.
In addition, my CPU with 100% on all four cores jumps to 80C.

Now, I'm just running at 3.15GHz with the lowest voltage settings in the BIOS for everything under the Processor menu.
Click to expand...

jimmyneutron over at the conroe overclock database thread, there is a user that ran what i saw in his screen shot 4 SP2004? looks like orthos though on his QX6700

http://www.hardforum.com/showthread.php?t=1138241&page=2

his name is JethroXP, and the reason why you get those high temps is you have 4 cores so that actually increases ALOT of heat compared to just 2 cores..
 
devil_trigger said:
Hey slappynutz and guys that helped him. I have the same mobo and cpu and I applied alot of the things that you said for him to do and I have a couple questions. I set the host freq to 300 and the ram freq to 533 and I disabled almost everything except for speedstop and the failsafe watchdog which I couldnt find. Does anyone know exactly what those are called or where they are on the bad axe 2? Now the first time it booted up it hung at the "welcome to windows" or whatever when its loading to the desktop for like 10-15 secs and when it eventually got to the desktop i had no internet. So I restarted my pc and and it didnt hang for as long but a little more than normal and I have internet now. Does nayone have any suggestions for me from what I've said? BTW, I am at 2700 mhz right now and I'm about to start ORTHOS.
Click to expand...


In order to disable the Failsafe Watchdog you need to change the maintenance jumper on the motherboard and boot. It will go right to the maintenance bios menu. Just check your motherboard manual for the location of the jumper. It's near the IDE port and has a tab for easy pullin'.

Depending on your ram you might be setting the host freq too high. I've always set it at 266/533, as that's what most recommend for this overclock.

And though I've always disabled C1E and EIST, after further testing it's clear that there's a bios bug where even if you disable them in bios the multiplier and voltage will still drop at idle in Windows. Which means, until Intel fixes it, we'll all be overclocking with C1E and EIST essentially enabled anyway. After the last bios update I cleared the bios settings and disabled only the Failsafe Watchdog, the Marvell Sata controller, and the floppy controller, and enabled Powerslope. Everything other setting I ran stock, and I'm overclocking fine. Maybe when they get around to actually disabling C1E we'll see if it makes a tremendous difference with an overclock.

I'm running completely stable at an fsb of 360, with every voltage at the motherboard stock (vCore reports at 1.27 in Windows with vDroop), and the Speedstep settings are still enabled. Passes Orthos and OCCT at 3.24ghz and from all indications is rock solid.

If I bump the fsb to 366, I pass 24 hours of Orthos and fail OCCT. Strangely, even if I bump the vCore from 1.3250 all the way to 1.4000 it still fails OCCT at 366. Boots into Windows, passes every test I throw at it and games fine, but fails OCCT. Leaving it at stock voltage and 360fsb until I get some more time to tinker. Or until Intel actually allows us to disable Speedstep.
 
Thats the thing though, I've looked in both bios's over and over again, even the maintenance, and couldnt find the failsafe watchdog anywhere. Its ok though, I got to my goal of 3.2(see sig:) )
 
devil_trigger said:
Thats the thing though, I've looked in both bios's over and over again, even the maintenance, and couldnt find the failsafe watchdog anywhere. Its ok though, I got to my goal of 3.2(see sig:) )
Click to expand...

You must be using an old bios. Something like the 2333 or earlier. The Failsafe Watchdog was introduced in 2333 and has been twitchy in every iteration since.

*edit* The Failsafe Watchdog was introduced AFTER bios 2333. Just re-read my post and realized the error.
 
slappynutz said:
You must be using an old bios. Something like the 2333 or earlier. The Failsafe Watchdog was introduced in 2333 and has been twitchy in every iteration since.

*edit* The Failsafe Watchdog was introduced AFTER bios 2333. Just re-read my post and realized the error.
Click to expand...

Yeah I think 2333 is the one I have. The one where the HDD activity light stays solid when you disable the marvell sata controller. I never saw a need to update the BIOS though so oh well.:p
 
I've been reading this thread with interest lately as I recently purchased an Intel DX975XBX2. I thought I would share a bit about my experience...

Using an Intel QX6700 Core 2 Extreme I was able to achieve a FSB of 340 MHz, with the default multiplier of 10 and default voltage setings for Vcore, FSB and MCH.

I’m using 2GB of Mushkin DDR2 667 (Part No. 991512). The FSB and memory are running 1:1 so the memory is being over clocked by ~2.5% at 680MHz, which the result I was trying to achieve for this RAM. I was still able to lower memory timings to 3 3 3 10, which is the specification for the Mushkin part used – one of the reasons I selected it - it has tight timings. Memory voltage was set to 2.2v which is the low end of its voltage specification.

One of the things I struggled with when performing this overclock was it’s inability to pass Prime 95 torture testing starting at FSB ~333 – evident within the first 10 minutes. It would however pass the test using ~318 MHz – at least for the 30-60 minutes I tested it. Raising Vcore, MCH and FSB individually or in combinations did not result in being able to consistently pass Prime 95 for any substantial length of time.

I finally gave up on Prime 95 as a benchmark for my efforts because at FSB 340 MHz, the machine would boot fine, pass SiSoft and 3DMark06 testing, and be used for extensive gaming without issue. This result really made me question utility of Prime 95 as benchmarking metric. More research is needed here I think.

Once I got the FSB where I wanted, I attempted to change the CPU multiplier to 11. This time the machine would not boot, even after trying several higher voltage settings. This brings up the value of buying a CPU with an unlocked multiplier....

Temps turned out to be ~32c idle and ~44 loaded. The highest I saw was ~47c.

A couple of other observations:

- I was unable to make Vcore rise above 1.306 to 1.318v no matter what manual setting was used. The appications used to monitor voltage and temperature were Intel’s Desktop Utility program and Speedfan v4.32.

- While idle the processor will drop CPU multiplier to 6, apparently in an effort to reduce power. This was the result even when the c1e and EIST were disabled. Apparently this is a know issue with this board. However as soon as any load is been placed on the CPU the FSB changes back to the default of 10.

- When idle V-core will drop to 1.22v. Again, when a load is placed on the CPU, Vcore will rise to 1.306 to 1.318.

- Enhanced Power Slope appeared to average out the peaks/valleys of Vcore, but very little. I left it on, however, I did not see a significant difference in vCore range when off.
 
You must log in or register to reply here.
Back
Top