ASUS P5W DH - Problems + Fixes Thread

[Polynikes]

OK, I knew there had to be something wrong there. I've never heard of temps that low on any watercooling kit. My own kit has some of the best parts on the market right now and I get nowhere near those temps.

Now, if it was 35 F in the room, it might be possible.

[~aoe~]

This is a nice link that gives some useful background info on C2D temps and how to interpret them.

However, my favourite quote is from MaxxxRacer's Water-cooling guide (leak testing):

Final Notes About Your New System
This section is to address one point that seems to escape most users and has been a point of much frustration: onboard temperature sensors.

When many people upgrade to watercooling, the first thing they do is run to smartgaurdian, MBM5, or whatever temperature monitoring software they use, and see if their temps improved compared to their previous air cooling setup. While I can’t stop you from doing this, I can inform you about those so-called temperature sensors. Plain and simple, they are useless. They can be up to 10C off and are NEVER accurate. Yes, I said never. So please do not rely on the onboard sensors for CPU temperature readings. Furthermore, many CPU temperature diodes are affected by air movement around the CPU. The more air movement (ala: Thermaltake TT w/ 220CFM delta fan) the lower the reported CPU temperature, regardless of what the actual CPU temperature is.

At this point you are probably wondering what is causing the inaccurate readings. Your first guess is the software. Dam those software engineers. WRONG.. Your next thought goes to the CPU manufacturer.. Your getting warmer, but Intel/AMD are only part of the problem. Yes, you guessed right now. It’s the motherboard manufacturer. That is indeed correct. Motherboard manufacturers use IC's (integrated circuits) from various companies, to monitor temperature and voltage. Depending on which IC they use, it may have absolutely horrid accuracy, or high accuracy. However, there is another factor. While a particular IC may be of high quality and have the capability to display highly accurate voltages, the IC temps/voltages still MUST be calibrated. This is where the system starts to crumble. To my knowledge there has never been an accurate temp/voltage reading ever reported from a motherboard. This is because the motherboard manufacturers simply will NOT take the time to calibrate the IC's they use with a known standard.

So the next time you look at your temperature monitoring software, please remember what I have said, and all but disregard the useless numbers it reports to you.

After so much frustration trying to resolve my own temps with my current system, the above post made me feel so much better

[aledsav1]

OK, I knew there had to be something wrong there. I've never heard of temps that low on any watercooling kit. My own kit has some of the best parts on the market right now and I get nowhere near those temps.

Now, if it was 35 F in the room, it might be possible.

This is a nice link that gives some useful background info on C2D temps and how to interpret them.

However, my favourite quote is from MaxxxRacer's Water-cooling guide (leak testing):

After so much frustration trying to resolve my own temps with my current system, the above post made me feel so much better

right now we know where we stand???? so any reading could and most probably is meaningless. still as i have know found out that i add 15c to my (pointless) temp readings due to tj being 100 and coretemp reporting tj85 does this mean i can have temps higher than i would normally not want to go near?

[aledsav1]

Oh just to add wasn't going to lap e6750 due to fantastic low temps ha! looks like i will be losing a bit of skin now.

[gwolfman]

I'll have a go at trying to explain, but not sure I'm qualified...

Your skin emits liquid (sweat) which, when air passes over it, creates a cooling effect (due to evaporation of the liquid). This is why the air coming into the case 'feels' cooler (it's moving and passing over your skin).

The radiator fins don't emit a liquid (hopefully), and therefore does not cool by evaporation.

The reason the tubes feel cool is that they are cooler than your body temp (body temp ~37C).

It is possible in humid conditions that water condenses on the radiator, which would then assist the cooling effect (by evaporation), possibly allowing your temps to fall slightly below ambient. However, we're only talking 1-2C here at most.

Well put!

[Polynikes]

Don't worry, aledsav1, your temps are fine. Even if they were waaaay off, which I'm sure they're not, you've got to be fine because you're watercooling. (That is, provided your pump is still running. ;D)

[~aoe~]

Well, I spent a lot of time trying to resolve the inconsistencies between my temps after reading the guide linked above. Then I recently read the quote above, and, as I say, started to feel much better about things.

Based on everything I have read so far, my take is this:

The temps reported by all these software tools are useful as an indication only, but we have to accept that they are, in all probability, not accurate, as the sensors themselves have probably not been calibrated to provide an accurate measurement. However, poor calibration shouldn't (in theory) make the readings totally unreliable (there must be a tolerance that the suppliers of the sensors work to, and that tolerance cannot/should not be excessive - be nice to know what those tolerances are )

At the end of the day, system stability is the most important (only?) factor in determining whether the OC is acceptable. Saying that, as many of us are pushing our systems to the limits, then extreme temps could severely reduce the lifespan of the CPU and other components.

Also taking into account that it is very difficult to get a reliably-calibrated temp sensor into a suitable position to record the accurate temps, and also considering that the CPU specifications are again based on tolerances - well, there are so many variables that it's difficult to provide an informed opinion.

At the end of the day, I would say that, if you're prepared to OC your system to the max, then component lifespan is not a major issue - just system stability. If you can maintain a stable system, then it's more likely that you will replace the components before they fail.

[~aoe~]

Don't worry, aledsav1, your temps are fine. Even if they were waaaay off, which I'm sure they're not, you've got to be fine because you're watercooling. (That is, provided your pump is still running. ;D)

I agree

[Polynikes]

At the end of the day, I would say that, if you're prepared to OC your system to the max, then component lifespan is not a major issue - just system stability. If you can maintain a stable system, then it's more likely that you will replace the components before they fail.

Considering the fact that modern computer hardware is theoretically capable of lasting for decades, overclocking, even heavily, will not reduce the lifespan of the hardware long enough for anyone to actually notice. Most people do major upgrades (mobo/cpu) every 3 years or less, so the chances of overclocking killing a component in that timeframe are almost nil. But hey, I may be wrong. I've been overclocking for a few system build cycles and have never had a part fail me.

[jkresh]

I was planning on running my go q6600 on a gigabyte 680i board, but as that has been a complete disaster (could not get it too boot consistently and it killed some of my ocz ram, without overclocking). I am going to run it in my p5wdh at least until x38 boards are out.

What bios is recommended for q6600 overclocking? I have done a search and have not seen an overall recommended bios, more of a couple of people having success with some and nothing terribly consistent.

[holwill]

Just a quick question for the guys using watercooling and pushing 400fsb or above, how many of you are using the small Asus fan that clips on the mosfet cooler? Only reason I ask this is because, I have been for the last 12 months, but now find myself needing to add a PA120.1 to the intake fan of my case(v2000) so it will obscure that area meaning I can no longer use it. Should I be worried and add a small 40mm fan somehow or just forget about it, as it will be fine with nothing added?

[BulldogPO]

I do´nt have that little fan added. Everything works well.
Those mosfets do´nt even get warm.

[splicer225]

Hi there,

I've been using a P5W DH for a year or so and I've never really had any issues with it until recently. I've got an E6420 which was running no problems for quite a few months. For the past few weeks I've been running it with a gentle overclock - I just upped the FSB to 333 (2.66Ghz) and the memory (4 sticks of 1 GB Corsair XMS2 PC5400) was running at 670Mhz or so. Again, everything running absolutely fine. I added an X-fi card and everything was working fine and then I soon started getting frequent lockups during the POST. They aren't lock ups per se, just extremely slow posting. It used to take around a minute to POST (yes, I know that's too long as it is) but now it appears to be taking longer and longer. After it says 4096MB OK it just hangs there from anywhere between 2 and 6 minutes before it continues on to loading Windows. If I keep restarting the computer, I can sustain the overclock (with the 2+ minute POST) however once I turn off the computer and then switch it on again I get an Overvoltage/Overclock error. I don't think there is an actual problem with the overclock - I can play games/run Orthos for extended periods absolutely fine - it is just getting past the POST which is the issue. Even if I can boot with 2.7Ghz, then Windows and everything on it runs perfectly normal.

I've tried BIOS 2206 and have now downgraded to 2004 after reading this thread, but am still getting the hangs. I've tried messing around with voltages, but none of them seem to work and I get the same issues. The problem is quite intermittent - it may successfully boot on some occasions, in other occasions I'll need to restart, in other occasions I'll need to unplug and replug the machine. One thing I was thinking is could it possibly be a power supply issue? I have an Enermax 500W and a Radeon 2900XT which I understand is fairly power hungry (but it was working perfectly fine previously). Although, the slow POST still exists at stock speeds/voltages. I think (well hoping) that I'm doing something fundamentally wrong here in the BIOS settings but I can't for the life of me figure out what it could possibly be. I have Vcore set to 1.35v (tried everything up to 1.37), memory voltage I've tried 1.8 - 1.95v, and I have Vmch and the ICH voltage (I could be getting these names wrong) at 1.05v and 1.20v if my memory serves me correctly. After clearing the CMOS as soon as I do anything in regards to memory timings, the slow post begins again. I've disabled HyperPath (problems exists with it on and off) and also all fan control options in the BIOS. Is there anything else I could try? Or is my board slowly dying do you think? And if it is, will I even be able to get an RMA for it since the problem is intermittent and unpredictable dependent on the settings for my devices (the e-tailer I bought it from are notoriously bad and inconsistent when it comes to support - I didn't know this when I bought the board!).

Sorry if this sounds all very incoherent, I'm not very technically minded when it comes to overclocking or anything. But I would really appreciate any help someone could throw at me. Many, many thanks in advance! Cheers.

[~aoe~]

One possible cause for slow POSTing is problems detecting peripherals (specifically HDDs in my experience).

If you have multiple HDDs, try disconnecting each one and booting to POST. You could also try connecting to different ports.

If using IDE, also check connections and jumper settings on the devices.

[palito]

wats the best Bios ?

[happ]

From the first page of this thread:

Extremely slow POST - Disable everything to do with fans in "Hardware Monitor". [Credit - OBR]

[splicer225]

I've already tried to disable all fan control options and have added each device one by one and seen if it has some sort of effect on the POST speed . The only thing that increases the POST speed is when overclocking (and it does so dramatically). Even by just manually setting my RAM timings to 4-4-4-12 (which is the speed they are supposed to run at). Without overclocking anything else, or changing any other settings - that will lead to slow POST.

[aledsav1]

wats the best Bios ?

Just to ask same question....from users experience what has been best bios...most overclockable? ta

[MarcusTaz]

running stock 2004 bios fine, stable and I yield 3.4GHz on a Q6600 B3

[gwolfman]

I've already tried to disable all fan control options and have added each device one by one and seen if it has some sort of effect on the POST speed . The only thing that increases the POST speed is when overclocking (and it does so dramatically). Even by just manually setting my RAM timings to 4-4-4-12 (which is the speed they are supposed to run at). Without overclocking anything else, or changing any other settings - that will lead to slow POST.

Maybe it's your memory then...? You have any other sitting around you can try? If not, then pull all but one stick out and try booting, see how it goes. Try and eliminate possibilities one by one.

[Sans]

I am running bios 2205. My Problem is one that many Vista users are suffering from right now. I get the atikmdag error when running 3D intensive games. This error causes black or flickering screens followed by a bubble in the task tray telling you that a display driver has stopped responding and has been recovered.

TDR Explained

This problem isn't just ati or nvidia it's both:
ATI
Nvidia


Some people have had success when they sync or link their ram. I didn't see that option in the bios anywhere. Is running the fsb and ram in a 1:1 ratio the same thing as syncing?(And if so what options do I go into to make that happen for DDR2 800 Ram?)

[splicer225]

I think I've fixed my slow POST problems, thanks very much guys. Carefully reinstalled the RAM modules and it seems to be a lot speeder/stable now. Cheers!

In regards to the ATIKMDAG error I don't get that in any games, but I get it whenever I try to run Windows Dreamscene (a moving background - literally just playing a video). I never used to get it with my old ATI card (X1600) but I get it now each time I try to run it with my new card (2900XT).

[trc]

hey guys. i've recently built my first pc with the p5w, and everything has been fine up to this point. when i came to install all the drivers, i haven't been able to do it for the ethernet, wifi and remote drivers. all 3 come up with the error message "error:5 - cannot execute..." with the location of the setup executable file for the corresponding driver. this is probably a really simple issue, but i can't for the life of me figure out what's wrong. any help is much appreciated.

tom

[gwolfman]

Maybe it's your memory then...? You have any other sitting around you can try? If not, then pull all but one stick out and try booting, see how it goes. Try and eliminate possibilities one by one.

I think I've fixed my slow POST problems, thanks very much guys. Carefully reinstalled the RAM modules and it seems to be a lot speeder/stable now. Cheers!

Ahh, so I was right. Cool, glad it's fixed!

[rafaeldb]

The 4 gb question again

No just kidding but its about the 4gb, I know from reading this huge thread that only XP or Vista 64 can see more than 3072 but I was planning on using Vista 64 for when I need to work with more ram and keep using Vista 32 until the 64v gets more support. My question is there any problem or harm for my ram when I'm on vista 32 (for most of the time)?

Thx