Official GIGABYTE Forum

P55 Mobos CPU sensor problem

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #45 on: February 16, 2011, 11:55:42 pm »
Errh, DM, yes we CAN reproduce the faults. At least, with those of us who've personally met the CPU temperature problem, we've certainly been able to turn the misread CPU temperature on and off. So, we know that one or more BIOS settings has an influence on this. We've even been able to provide photos of the whacky temperatures, temperatures that quite clearly cannot ever normally arise. I can only summise that the Gigabyte techie either simply hasn't left the PC on for long enough or performed his test with C3/C6/C7 turned off. I personally found that if I set the Vcore immediately to a fairly high level, then irrespective of C3/C6/C7, the reported CPU temperature was sensible, no matter how long you then waited. However, if I started with a low Vcore and then gradually raised it, and also left enough time, the crazy temperature resulted. So, something quite subtle appears to be going on. I told the techie precisely this but I've a feeling he didn't test in quite that way.

As for my PIO problem, I think I may finally have found a lead on the Web, suggesting that possibly the effected optical drive is faulty. Apparently, if a CD/DVD drive is so bad that it repeatedly gives multiple read errors, then Windows automatically drops the drive's original DMA mode downward. If it's really SO bad, it'll drop it right down in speed to PIO mode and it'll stay there. So, it may well transpire that that particular problem is due to a faulty drive. I'm annoyed about that, as it's a brand new optical drive. Obviously, I'll now do some more testing on it. When I first installed WinXP on the new PC, I used the other optical drive, but when I started installing the soundcard driver and then various apps, I just happened to use the second optical drive instead. Therefore, if it IS faulty, then perhaps some of the software I've installed is corrupted. My only recourse may be to get a replacement drive and then reinstall everything from scratch. It's conceivable that this is also linked in some way with the constant HDD accesses problem. However, it wouldn't explain the random FDD accesses.

 

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #46 on: February 17, 2011, 10:31:15 am »
Hmm, well, I've now experimented somewhat with both of the optical drives - physically disconnecting them, setting each to None in the BIOS, and also setting the Transfser mode to None in Device Manager. Not only are the constant HDD accesses still there (and the LED still constantly flashing) but also that particular optical drive is still stuck in PIO mode when reinstated! So, in my view, that Technet forum (the discussion was a year ago) got it totally wrong. Well, let's face it, it'd be completely barmy to expect a HDD accessing problem to be solved by having to permanently disable the optical drive, wouldn't it?! This just shows you how careful you have to be when searching for solutions to bugs and of not being over-confident.

As for the CPU temperature issue, one other possibility I've been toying with is that it may be a function of the pressure with which the CPU heatsink is mated with the CPU. CPU coolers other than the stock Intel one tend to be more columnar and this means that when the PC is standing in its normal position there's a considerable turning moment exerted around the mating point of the cooler and the CPU, resulting in a droop of the cooler. There's therefore a temptation to tighten down the cooler to make it fit square. This must put a considerable mechanical distorting pressure on the zone of the motherboard that's in and around the CPU socket. Having seen how small the matrix of connection lands were, underneath the CPU chip, when fitting the CPU, I can well imagine that it'd take very little movement to create a disconnect on one of those lands. So, one possibility is that one or more of these connections between the CPU and the socket start off from cold as okay but then, as the CPU heats up and everything in that area expands, it goes open-circuit, aided in this by the distorting pressure on the socket and surrounding zone of board. So, as a consequence, either the CPU obtains a daft temperature value from a sensor directly underneath the CPU chip (a temperature-dependent resistor), or an output from the CPU, going across the board to the BIOS chip, goes open-circuit. ........................ Clutching at straws, here, I have to admit.   

Dark Mantis

  • *
  • 18405
  • 414
  • 10typesofpeopleoneswhoknow binaryandoneswhodont
    • Dark Mantis
Re: P55 Mobos CPU sensor problem
« Reply #47 on: February 17, 2011, 03:53:26 pm »
Forumjoe

Yes I know you can see the fault no problem and I am sure reproduce it at will. I was saying the GGTS can't or at least are not trying to. :-\
Gigabyte X58A-UD7
i7 920
Dominators 1600 x6 12GB
6970 2GB
HX850
256GB SSD, Sam 1TB, WDB320GB
Blu-Ray
HAF 932

Gigabyte Z68X-UD5-B3
i7 3770K
Vengeance 1600 16GB
6950 2GB
HCP1200W
Revo Drive x2, 1.5TB WDB RAID0
16x DLRW
StrikeX S7
Full water cooling
3 x 27" Iiy

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #48 on: February 17, 2011, 04:32:02 pm »
From the photos they're sending ME, they're KINDA trying. But I reckon they're just not spending enough time on it and experimenting enough.

Re the 'stuck PIO' problem I've had, where the SATA optical drive has got stuck in that mode rather than UDMA, I think I'm now getting somewhere. I completely uninstalled the controller on the 1st Secondary IDE channel on which it was connected, then rebooted. The drive is now showing in Device Manager as UDMA2. But I suspect it'll revert to PIO when next used. I gather that some brands of drive are sensitive to the precise IDE channel on which they're connected, so I've contacted the manufacturer for advice, in case it's relevant here.

Dark Mantis

  • *
  • 18405
  • 414
  • 10typesofpeopleoneswhoknow binaryandoneswhodont
    • Dark Mantis
Re: P55 Mobos CPU sensor problem
« Reply #49 on: February 17, 2011, 05:25:19 pm »
Yes the more info you can get on the particular hardware devices the better.
Gigabyte X58A-UD7
i7 920
Dominators 1600 x6 12GB
6970 2GB
HX850
256GB SSD, Sam 1TB, WDB320GB
Blu-Ray
HAF 932

Gigabyte Z68X-UD5-B3
i7 3770K
Vengeance 1600 16GB
6950 2GB
HCP1200W
Revo Drive x2, 1.5TB WDB RAID0
16x DLRW
StrikeX S7
Full water cooling
3 x 27" Iiy

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #50 on: February 19, 2011, 12:02:24 pm »
GGTS have now given me their final word on this misreported CPU temperature issue - they can't reproduce the problem, so therefore the problem doesn't exist!

This is a ludicrous attitude for them to strike, particularly as several of us with different P55s have sent them photos of the crazy temperature values in the BIOS. GGTS's conclusion defies all human logic.

I suspect that the reason for this is that they've realised that they'll only reproduce the problem if they spend more time investigating it, and with all the other bug reports on their various mobos flooding in from all over the world, I guess they need to give those other issues some attention instead. Let's face it, it takes time to set up a PC with a particular CPU, particular memory, etc. Frankly, I wish that they'd not spent as much time on the hardware with this particular issue and instead went straight to investigating the BIOS software instead, as the problem is most defintely BIOS-related. The sheer fact that we're able to turn the misreporting on and off using the BIOS's C3/C6/C7 setting is evidence of that.

It's very noticeable that GGTS avoids actually getting into any sort of discussion with you. They won't even answer very straightforward questions, like 'Is the Microsoft UAA driver required only for onboard sound?' Or, 'Here's the sumcheck value of the newly-flashed BIOS - is it correct?' Instead, they simply take your reported symptoms and try to reproduce them. The lack of any discussion is a huge barrier to solving problems, but it might simply be that they feel that time is better spent elsewhere.

As for the other problems I've been having with my new PC, well, they can be summarised thus:

1) Random FDD accesses.
2) Constant HDD accessing (no HDD idle periods).
3) Permanent PIO mode of one optical drive.
4) Highly-distorted third-party sound (but not always).

At first, I thought that these were also due to either a bad BIOS or a bad mobo driver but I'm not so sure now. I've made some headway with Problem 3, in that I appear to have cured it by uninstalling the controller of the affected optical drive in Device Manager and then rebooting. Mind you, a stuck PIO mode is usually a sign that the OS has detected many read errors from the drive, so I'm not holding my breath over it just yet. It might just be that that nice new drive is faulty.

1 and 2 remain unsolved.

2 could be due to a faulty HDD but that can now only be verified by sending it back to the manufacturer and getting a replacement. The reader will realise the loss of software and data involved, to say nothing of the time required to do it all and to return the PC to normal later on, so I'm reluctant to proceed without more compelling evidence.

The distorted sound happens only about 50% of the time, in that sometimes when I run a music file the sound is fine, right from the outset, but then if I stop the file running and re-run it later, it'll be grossly distorted. It's also noticeable that the constant HDD accesses have a regularity to them (1 - 2 accesses per second) and that regularity modulates the distorted sound to some degree. So, I'm fairly sure the two are linked in some way, eg. clashing on the PCI bus.

As is found with so many obscure problems that PC users discover with their PCs these days, if you google for them, you inevitably find a plethora of 'red herring' solutions from well-meaning but often misguided people. Honestly, you have to laugh at the sheer variety of them. Different respondents on forums offer their '100% guaranteed' solutions which invariably are not borne out when tried. Indeed, some of them are just plain daft and can, if you're not careful, get you into even deeper trouble or can cause corruption of the OS.

I'm of the view that certain problems can be caused by the order in which Windows Updates are applied, or not applied as the case may be, especially as is the case when you've done a reformat of the hard drive or have just finished building a new machine and you therefore need to download and install a great wadge of OS updates. Curiously, some updates from the WU website require subsequent manual changes in the Windows Registry, something you won't realise if you opt for automatic Windows updates each month. Even that I find really weird, in that surely the requisite Registry mods should be done as part and parcel of the automatic download/installs, shouldn't they? Apparently, not.


Re: P55 Mobos CPU sensor problem
« Reply #51 on: May 07, 2011, 05:17:13 am »
stumbled across this thread via google-foo today while researching the crazy temperature problem that has suddenly sprung up with my GA-P55A-UD3. i turned my computer on this evening to find my CPU fan completely stopped. while trying to figure out why, i noticed my CPU temp as reported by my BIOS as -10C.

oddly, each time i entered BIOS while enabling/disabling different settings trying to clear up the condition, the temp would decrease by 10C. i ended up at -90C before i gave up, and shut the PC down and went to have dinner. when i came back, the problem had miraculously fixed itself.

while the problem was occurring, realtemp and CPU-tweaker both reported normal-ish temps (for passive cooling anyway), which would suggest the CPU temp sensor is fine. ET6 claimed the CPU was pegged at 100C.

i have an older BIOS build, F6. im hesitant to flash the BIOS unless i know it will fix the problem. i find it odd that the problem suddenly started without me changing anything.

does gigabyte or anyone else have any new information regarding this problem?

Dark Mantis

  • *
  • 18405
  • 414
  • 10typesofpeopleoneswhoknow binaryandoneswhodont
    • Dark Mantis
Re: P55 Mobos CPU sensor problem
« Reply #52 on: May 07, 2011, 11:23:45 am »
No new info I am afraid but I would suggest try changing the motherboard battery from what you say as it has to be something simple like that if everything else is working. The other thing to check would be the PSU voltages when the problem is occuring.
« Last Edit: May 08, 2011, 06:34:54 pm by Dark Mantis »
Gigabyte X58A-UD7
i7 920
Dominators 1600 x6 12GB
6970 2GB
HX850
256GB SSD, Sam 1TB, WDB320GB
Blu-Ray
HAF 932

Gigabyte Z68X-UD5-B3
i7 3770K
Vengeance 1600 16GB
6950 2GB
HCP1200W
Revo Drive x2, 1.5TB WDB RAID0
16x DLRW
StrikeX S7
Full water cooling
3 x 27" Iiy

Re: P55 Mobos CPU sensor problem
« Reply #53 on: May 08, 2011, 04:23:53 pm »
stumbled across this thread via google-foo today while researching the crazy temperature problem that has suddenly sprung up with my GA-P55A-UD3. i turned my computer on this evening to find my CPU fan completely stopped. while trying to figure out why, i noticed my CPU temp as reported by my BIOS as -10C.

Hi, I have had exactly the same problem with a P55-USB3 the fan completely stopped and when I went into the bios the temperature was showing -37 C, after I waited for about 3 minutes I turned the pc back on and it was fixed. Concerning the bios upgrade, one day when I turned the system on I heard a loud noise all the way into windows and I didn't know what it was, sure enough it was just a very long beep and the temperature in the BIOS was reading -37C again, I upgraded the BIOS from F8 to F9 and I still  had the same reading in the BIOS and the same noise the only "fix" was completely shutting down the PC and waiting a few minutes. Funny thing is that every time this erroneous reading happened I always checked the temperature with RealTemp and it was reporting the correct temperature of the CPU.

I also have random beeps while in windows that stop if I turn off "Fan Fail Warning" in the BIOS but I don't want to do that because I had it turned off when my fan completely stopped and I didn't notice it until the PC became very sluggish. This random beeping in the desktop goes away if the CPU is always at full speed by disabling C states in the BIOS or if I watch a video, listen to some music or play a video game indicating that it only happens randomly(?) if the CPU is completely idle.

Before the random beeping in the windows environment I always heard a beep when I first turned on the system (no not the POST beep, it happened after the POST one) and I noticed that the fan always stopped when that beep happened and then worked again. I really don't know what made it change to making random beeps in Windows instead of making them after the POST but I'm grateful at least that my fan didn't stop working in 4 or 5 months. I bought my PC at the end of July last year and for about 4 months I didn't have these random beeps and the only fan problem was that once it wouldn't want to go above 945 RPM or so and when I checked the BIOS the temperature was -37C again, this thing happened two times, once in August and once in November, after that the whole beeping problem started.

Another thing about the random beeps in windows, whenever they happen I can hear my fan going almost full throttle sometimes for about 1 second after the beep and then going back to normal speeds.

My cpu is an i5 760 and everything is powered by a Corsair TX650W
« Last Edit: May 08, 2011, 07:57:51 pm by reigorn »

Dark Mantis

  • *
  • 18405
  • 414
  • 10typesofpeopleoneswhoknow binaryandoneswhodont
    • Dark Mantis
Re: P55 Mobos CPU sensor problem
« Reply #54 on: May 08, 2011, 05:28:58 pm »
Hi and welcome to the Gigabyte Forum.

It would seem to be a BIOS issue. The trouble is until they release another BIOS update with it fixed there is little you can do apart from report it to GGTS.

Just enter your email address and click on the language of choice.
GGTS   http://ggts.gigabyte.com/
« Last Edit: May 08, 2011, 06:35:18 pm by Dark Mantis »
Gigabyte X58A-UD7
i7 920
Dominators 1600 x6 12GB
6970 2GB
HX850
256GB SSD, Sam 1TB, WDB320GB
Blu-Ray
HAF 932

Gigabyte Z68X-UD5-B3
i7 3770K
Vengeance 1600 16GB
6950 2GB
HCP1200W
Revo Drive x2, 1.5TB WDB RAID0
16x DLRW
StrikeX S7
Full water cooling
3 x 27" Iiy

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #55 on: May 08, 2011, 06:13:27 pm »
GGTS have been in complete denial about this for at least the last three months, and possibly well before that. As you'll see from my earlier postings on this bug, I spent many days, on and off, conducting experiments to pin it down and interacting with GGTS. It appears to be a BIOS problem, affecting not only the P55A-UD3 but also several other mobos in the P55 series.

Unfortunately, GGTS took the view that, until they could reproduce the problem themselves, the problem doesn't exist. And, of course, they did their best to avoid setting up the appropriate conditions for the CPU in the BIOS to reproduce the problem, despite instruction from me.

I found that, with 'normal' settings for power-saving, temp alarms, etc. in the BIOS, the negative temperature readings would appear after the machine had been on for around 30 mins. The problem also appeared specifically around voltage settings for the CPU that were slightly below the average recommended operating voltage for the CPU. If you used the BIOS to tweak the voltage to a higher level, eg. if you specifically were trying to overclock the CPU, the problem would reduce. However, increasing the voltage will, of course, cause the CPU to run hotter, anyway, and so keep the temperature more positive.

Fan speeds will drop, or the fans stop completely in some instances, if the BIOS reads that the CPU temperature has apparently dropped to a low value (since obviously cooling is then not required). The reality is that, when this happens, the CPU's temperature will be going up! So, this is risking the CPU seriously overheating.

GGTS's attitude came across to me as one where they were looking for every excuse not to investigate and re-write the BIOS. I even sent them screenshots of my BIOS, showing the negative temperatures, but they still failed to agree that any such problem existed. Perhaps if and when considerably more users complain to GGTS that this is happening, they'll change their tune and admit that there's a very real problem there. Frankly, my personal view is that the core BIOSs used across the P55 series mobos are badly written and contain not only this bug but several others as well.


Re: P55 Mobos CPU sensor problem
« Reply #56 on: May 09, 2011, 08:03:00 am »
Dark Mantis, i have tested my CMOS battery, and am getting 3.1v. its a 3v nominal cell, so im sure its OK. i didnt test my PSU voltages while the problem was occurring, and the problem has not reoccurred since my first post.

i hesitate to contact GGTS about the problem just yet, because, as i said, the problem has not reoccurred. from what im hearing they will likely tell me it was a fluke or my eyes were deceiving me.

FWIW, my system is mildly overclocked and has been running the current settings for about a year:

i5-750 @ 3.8ghz (200x19) @ 1.25v

unfortunately, due to my current fan controller setup, i do not have RPM monitoring on any of my fans (and thus, CPU fan failure warnings are turned OFF). fortunately, i have a chunky enough air cooler that my system will run passively at acceptable temps as long as im not folding or running prime95 or something ludicrous like that.

forumjoe

  • 195
  • 6
Re: P55 Mobos CPU sensor problem
« Reply #57 on: May 09, 2011, 11:58:50 am »
Wicked_Sludge,

I'm bound to say that if you don't report the matter to GGTS (and others not bother either), then GGTS will never really get the message that there is indeed a bug of this kind in the BIOS. So, please do report it to them. Their attitude will only change when significant numbers of users have complained about the problem.

Dark Mantis

  • *
  • 18405
  • 414
  • 10typesofpeopleoneswhoknow binaryandoneswhodont
    • Dark Mantis
Re: P55 Mobos CPU sensor problem
« Reply #58 on: May 09, 2011, 12:10:27 pm »
Yes i would have to agree with forumjoe there. You must report all problems that you can't resolve otherwise they only hear about the odd ones and don't give the issue all the work it requires.
Gigabyte X58A-UD7
i7 920
Dominators 1600 x6 12GB
6970 2GB
HX850
256GB SSD, Sam 1TB, WDB320GB
Blu-Ray
HAF 932

Gigabyte Z68X-UD5-B3
i7 3770K
Vengeance 1600 16GB
6950 2GB
HCP1200W
Revo Drive x2, 1.5TB WDB RAID0
16x DLRW
StrikeX S7
Full water cooling
3 x 27" Iiy

Re: P55 Mobos CPU sensor problem
« Reply #59 on: May 10, 2011, 05:25:12 am »
Ok, done and done.

we shall see what they have to say!