1 3 4 5 6 7 Previous Next 133 Replies Latest reply on May 7, 2012 4:01 AM by Flying_Kiwi Go to original post
      • 60. Re: DG45ID Irregular fan speed
        Flying_Kiwi

        joew wrote:


        Contacted Intel support. They don't believe it is a fault with the motherboard and as it is a non standard CPU fan can't offer a lot of advice. Tried changing settings etc but with no luck. They are looking into how the PWM signals work and will email them over. I asked whether this could be suggested to be looked at for the next BIOS update and the the agent said he would mention it but as its not standard Intel equipment he said its unlikely.

         

        Could any of you list the fans you are using? Does the Intel Desktop utility show a fan stop warning for you as well?

        My setup is Intel all the way including the standard Intel provided fan on my retail Intel E8500 CPU. I believe what you describe is a different problem however as the locking into high speed mode (whith the fans just running flat out, not stopped) is also accompanied by problems with shutting down or rebooting the PC (as detailed earlier here). If you have a Retail CPU you could put the original cooler back in situ for testing purposes.

         

        yf38 do you still have this issue logged with Intel to fix for their next BIOS revision as although disabling C-State, elliminates these problems for me as well, I don't think that disabling some of the CPU powersaving functionality is a solution, it's merely an interim workaround.

        • 61. Re: DG45ID Irregular fan speed
          joew

          I am not getting high fan speed. The rated speed of my fan should be 900 - 2500 rpm. I am getting quite the opposite the fan speed is too low, ie. zero.

          • 62. Re: DG45ID Irregular fan speed
            yf38

            Hi FK,

            I have no issue logged with Intel because:

            1) The work around disabling cpu C state is fine for me and I made some measurement with a wattmeter not showing visible difference in power consumption at idle state.

            2) I am convinced that this problem is not a good solid bug that can be reproduced the same with every card.

            My 308 rev. has not this problem, my 310 rev got it,

            2) This kind of problem has probably no good 100% firmware fix for all cards, only more or less effective firmware workarounds.

            4) This card is now more or less in end of life status and I suspect that the probability that Intel spends time to strongly investigate the subject is very small.

            5) May be Intel already knows the true problem, unfortunatly asking a hardware fix on the board, and due to the small impact of it they may prefer to do nothing.

             

            For Joew:

            1) If you think that the cpu fan stays at a too low speed, dont panic, Intel QST begins to increase the cpu fan RPM only when core temperature reaches 75 to 80°C (sorry I use °C here) in order to keep the noise as low as possible.

            2) Intel Desktop utility only reads the sensors and provides alerts in accordance with your settings, it does not modify the fans speed. It is possible to disable alerts wif you consider that it is false (as we had at some moment with voltages).

            3) For the fan which stops it may be due to the fan characteristics that do not support the too low speed that the card tries to set, maybe better to come back with Intel fan, which is not so noisy at usual RPM.

            As said before I used Intel E8400 stock fan, and then I switched (no good reason to do it) to Scythe Big Shuriken, which works fine, with lower noise at high RPM and that fits in a Silverstone Sugo SG01 case.

            • 63. Re: DG45ID Irregular fan speed
              joew

              Which bios date is 308 rev?

              • 64. Re: DG45ID Irregular fan speed
                mcudnik

                I recently purchased (2) DG45ID boards and am using the E8400 proc.  I was watching this thread intently waiting for an answer.  Upon upgrading to the 0129 bios, the CPU C-states selections are added to the bios menus.  They are enabled by default.  When enabled, I get the irregular fan speed.  Restarting and disabling just the CPU C-State selection, the irregular fan speed stops.  I upgraded to the 0131 bios on both boards.  The C-states were enabled again and the fan issue returned.  Disabling them in the bios once again corrected it.

                 

                Is it the E8400 and greater line of processors that cause it??  Not sure.  Mine are running great now though.  Awesome board.

                1 of 1 people found this helpful
                • 65. Re: DG45ID Irregular fan speed
                  Flying_Kiwi

                  yf38 wrote:

                   

                  Hi FK,

                  I have no issue logged with Intel because:

                  1) The work around disabling cpu C state is fine for me and I made some measurement with a wattmeter not showing visible difference in power consumption at idle state.

                  2) I am convinced that this problem is not a good solid bug that can be reproduced the same with every card.

                  My 308 rev. has not this problem, my 310 rev got it,

                  2) This kind of problem has probably no good 100% firmware fix for all cards, only more or less effective firmware workarounds.

                  4) This card is now more or less in end of life status and I suspect that the probability that Intel spends time to strongly investigate the subject is very small.

                  5) May be Intel already knows the true problem, unfortunatly asking a hardware fix on the board, and due to the small impact of it they may prefer to do nothing.

                  Thanks for the feedback. The problem definitely can't be motherboard revision specific because I had an AA rev 308 board and now this AA rev 310 board and BOTH displayed the same problem with locking in high RPM state. When I did report the problem, Intel did ask me an extraordinary amount about the CPU (it was as if they knew this was relevant) and as some people without E8500s but using the same motherboard revision don't have the problem, combined with the fact that turning off CPU C-State power saving makes the problem go away, I suspect thats where the problem is. I Googled CPU C-States and it appears that some of the advanced C-States aren't present on E4xxx and E6xxx but are on E7xxx and E8xxx CPUs so that could explain why some are having the problem with faster CPUs and others with slower CPUs don't have a problem.

                   

                  You won't be able to get a fair reading of power consumption because with the problem occurring (ie with C-States enabled), the fan speed and reboot problem crops up so there is no guarantee that power saving is actually working as it should be. With C-States disabled, of course the CPU doesn't even have the chance to enter this power saving mode. The board therefore won't be using all of the powersaving features built into the CPU and for an Intel board running an Intel CPU, for me that simply isn't good enough. I've tested my board without my TV card plugged in and the same problem occurs so it can't be anything other than an Intel problem.

                   

                  As you rightly point out the board will likely soon have any development work halted so it's all the more important that any and all affected by this problem report it and request that a fix be provided - if necessary a BIOS release where the only difference incorporates improved compatibility for the relevant E8xxx series CPUs would suit me. Up to around 80% saving on CPU power consumption is worth going for.

                  • 66. Re: DG45ID Irregular fan speed
                    joew

                    As you rightly point out the board will likely soon have any development work halted so it's all the more important that any and all affected by this problem report it and request that a fix be provided - if necessary a BIOS release where the only difference incorporates improved compatibility for the relevant E8xxx series CPUs would suit me. Up to around 80% saving on CPU power consumption is worth going for.

                    How can we Intel's attention on this?

                    • 67. Re: DG45ID Irregular fan speed
                      Flying_Kiwi

                      joew wrote:


                      How can we Intel's attention on this?

                      By reporting it to their tech support people in your country of residence and asking for a solution. Tell them exactly your setup and the nature of the problem (don't forget to mention the problem that happens with attempting to shut down the pc when it's in this locked high speed fan mode - the fans don't turn off - see earlier in the thread for further details). You can even mention that you are aware it's something affecting others through this thread! Let us know what they say.

                      • 68. Re: DG45ID Irregular fan speed
                        yf38

                        Joew: About my 308 rev. it had almost all the bios from 077 to 0131 (july).

                        At the very beginning (077) there was a pumping fan speed problem which was corrected soon in a subsequent bios. I am not sure that the cause was the same, only Intel knows.

                        However after that it had not fan pumping problem, but I agree with FK remark that some other 308 may have the problem.

                        About your fan, if it is specified starting at 900RPM this is higher than the minimum speed that Intel QST tries to go down to minimize noise. For example, my Intel fans were going down to about 850RPM and Scythe big shuriken goes down to 700RPM because it is more efficient.

                        If At 900RPM your fan is "too" efficient QST will try to slow it down and at some moment the fan will stop because it is to much slow compared to its specs.

                         

                        FK: you observation that 308 card may also have the problem is in line with my opinion that this may be some marginal problem due to timing or noise on the board, and also may depend of how we use them and what is the power supply, and so on.

                        If this happens only (?) with cpu which have the cpu C state capability (and enabled) this is maybe because this feature adds much extra noise and maybe other constraints to the system.

                        About power consumption, yes I was capable to make measurements because with my 310 card the problem rarely went enough far to cause reboot, almost only fan pumping, and the measurement did not show a sufficient difference to be visible with my wattmeter.

                        I dont know what marketing figure gives a 80% power consumption reduction, but it is probably a comparison between a "standard" idle halt consumption and the deep power reduction provided by cpu C state, if it is. However this represent a small amount compared to the typical consumption of the system even in idle state. Mine are somewhere in the 60W range including two sata disks, an optical drive, an USB TV key, keyboard & mouse.

                         

                        mcudnik: thanks to confirm that disabling cpu C state also cancels the problem with your cards. E8400 and the other E8XX E9XX are running on this card for a while without special problems (except this one ?).

                        The reason why Intel introduced cpu C state enabling/disabling in Bios 129/131 has never been very clearly explained, may we consider that it was a potential workaround that we should finaly discover ourself ?

                        • 69. Re: DG45ID Irregular fan speed
                          Flying_Kiwi

                          yf38 wrote:

                           

                           

                          FK: you observation that 308 card may also have the problem is in line with my opinion that this may be some marginal problem due to timing or noise on the board, and also may depend of how we use them and what is the power supply, and so on.

                          If this happens only (?) with cpu which have the cpu C state capability (and enabled) this is maybe because this feature adds much extra noise and maybe other constraints to the system.

                           

                          I would concur except disabling C-States resolves the problem entirely (while my PSU output remains well within the green according to Intel Desktop Utilities). This means the problem must be related to either a bug with the implementation of C-States power saving in the BIOS or a bug in the CPU itself relating to this and I'd be surprised if it's the latter!

                           

                          About power consumption, yes I was capable to make measurements because with my 310 card the problem rarely went enough far to cause reboot, almost only fan pumping, and the measurement did not show a sufficient difference to be visible with my wattmeter.

                          I dont know what marketing figure gives a 80% power consumption reduction, but it is probably a comparison between a "standard" idle halt consumption and the deep power reduction provided by cpu C state, if it is. However this represent a small amount compared to the typical consumption of the system even in idle state. Mine are somewhere in the 60W range including two sata disks, an optical drive, an USB TV key, keyboard & mouse.

                          My point about measuring accurately any power savings is how do you know C-State Power Saving was actually occurring? There's clearly something buggy going on and in all probability the required level of C-State power saving was never actually reached - hence no reduction in consumption. If you Google for CPU C-State power saving, the reference I used is the first one shown (specifically relating to the more advanced states later on in the article). Right at the end it refers to an example CPU and states the level of power savings possible (although mine isn't the same CPU, it's clear that these more advanced C-States enable very significant power savings to be made).

                           

                          mcudnik: thanks to confirm that disabling cpu C state also cancels the problem with your cards. E8400 and the other E8XX E9XX are running on this card for a while without special problems (except this one ?).

                          The reason why Intel introduced cpu C state enabling/disabling in Bios 129/131 has never been very clearly explained, may we consider that it was a potential workaround that we should finaly discover ourself ?

                          I hope that mcudnic contacts Intel to report this bug as the more people that report the problem, the more likely they'll be to fix it. Clearly the symptoms are cancelled but the problem is still there, hence this additional powersaving mode cant be correctly utilised. I doubt the incorporation of C-States into newer BIOS versions was to let us discover the workaround for ourselves and in fact every BIOS from 113 on has presented this problem for me (111 was fine in all regards). If they wanted a workaround, they could have had the BIOS sence the CPU version and not offer up C-States power saving as an option at all. The idea that flashing the latest BIOS on a recent AA revision board produces these problems at default settings is not acceptable and most people wouldn't know how to go into the BIOS and disable this feature, let alone thinking to try that. It was only a few comments from earlier in this thread that 'inspired' me to try this and have a eureka moment.

                          • 70. Re: DG45ID Irregular fan speed
                            Curious592

                            Do all of the problems correspond with onboard fan connections to PS, case or other fans?  When changing the heatsink/fan as I did, the detection may stall preventing power up.  So rather than a crimped wire being a problem, the detection simply needs to pass, correctly assigning fans.  The TV card issue still seems related to scheduling.  Changing from 12AM to 12AM to a 6 hour period may have helped hibernation outside of that period.  Another complication is automatic updates.  After the bulk of them have been applied and the system stabilizes, even the keyboard seems be working correctly except for entry to BIOS.  The production batch could be an issue but the revision does not seem to be for me.  The low profile fan runs about 600 for Q8300 or 1100 for P4 3.0GHz and older fan at 1100 with Q6600.  I see that these do not relate to C states listed in my BIOS.

                            • 71. Re: DG45ID Irregular fan speed
                              Flying_Kiwi

                              joew wrote:

                               

                              the fan speed is too low, ie. zero.

                              What happens when you reinstall the original Intel CPU fan (assuming it's a retail boxed CPU that came with a fan)? My CPU fan normally runs at around 1125 RPMs and my chassis fans are ~565 RPMs for the 140 mm fan and ~625 RPMs for the 120 mm fan - never had a problem with booting or fan speed warnings (even though both chassis fans are right down near the bottom of the green zone in Intel Desktop Utilities). I suspect that you have a faulty fan causing this problem, try an original Intel one - they're cheap (and quiet) enough for most people.

                              • 72. Re: DG45ID Irregular fan speed
                                Curious592

                                Flying Kiwi

                                Someone should followup elsewhere with the TV tuner card issues.  Several Sleep and Hibernate cycles succeeded outside of the 6 hour Media Center schedule update period.  Following overnight hibernation Aero failed to restart prior to update period.  After a couple minutes I restarted and CHKDSK automatically ran with no problems and produced normal desktop.  This is with 240GT.  System with 9600GT and BIOS 129 had tuner removed and schedule updates stop without discontinuing in setup.  No problem with power there but unknown high definition audio quality probably due to display port in monitor (but connected DVI).  Will be moving the Hauppauge 1250 from 307 board to 310 w BIOS 131 since this is not acceptable for usage on the 307.  All slots on DG45ID found to work very well for fan card or TV tuner and no significant fan issues found.  Your board may have hub problems.  I don't know if a quad core would distribute tasks more for you.

                                • 73. Re: DG45ID Irregular fan speed
                                  Flying_Kiwi

                                  I'm not experiencing TV Card related probs at all and I'm quite happy with the E8500 in terms of performance (recording two TV channels simultaneously and playing back a HD-DVD or Blu-Ray still has plenty of CPU cycles in reserve). It also means my system runs nice and cool with slow, quiet fans.

                                  It'd be worth contacting your TV card manufacturer to see if they have newer firmware to address these issues you're experiencing (especially if there are no problems without the TV card installed). Good luck.

                                  • 74. Re: DG45ID Irregular fan speed
                                    Curious592

                                    It was not clear if you are able to use sleep/hibernate with the card.  I see WMC as the problem.  It needs to be closed (at least with my card) or resources are not fully released.  Scheduled recording (or manual) works fine but the computer probably needs to be shutdown or restarted.  Win7TV might work differently.  Following CHKDSK the icon for media center was in the taskbar so it was not released.  I'll look at instructions for WMC but I doubt they fully cover such aspects.  Windows forums might help me since I'm pretty much done here.  Thanks for the input.

                                    1 3 4 5 6 7 Previous Next