Announcement

Collapse
No announcement yet.

Epox 9NDA3+ with exsessive coil noise

Collapse
X
Collapse
 
  • Page of 4
  • Filter
  • Time
  • Show
Clear All
new posts
Previous 1 2 3 4 template Next
    #1

    Epox 9NDA3+ with exsessive coil noise

    That 9NDA3+ mobo I bought from tiresias on this board, it has some really amazing coil noise, it's so high it drives me crazy... I am sure he never experienced it because then he would have killed it with a sledgehammer

    Anyway, gonna try swapping out all caps on it except the Sanyos... Because I could not make the coil noise stop with hot glue on the offending coil! I can not even get it to stop by putting my finger against it; only make it change frequency...

    That app "hrping" trodas recommended to me made it whine it's coil in a different, albeit more friendly tone... Very nice progg I am sure... But on this system just playing Winamp makes it sing like crazy The thing that makes it sing the most is downloading a file at high speed; which makes me think the problems tiresias had with it's memory are related to caps afterall, because during a file transfer memory is used to save the file before it is copied to disk...

    This will be a very fun problem to diagnose, I love these kind of problems, they make you so satisified when you find what it was that caused it
    "The one who says it cannot be done should never interrupt the one who is doing it."
    Tags: None
    #2
    Re: Epox 9NDA3+ with exsessive coil noise

    Originally posted by Per Hansson
    That 9NDA3+ mobo I bought from tiresias on this board, it has some really amazing coil noise, it's so high it drives me crazy... I am sure he never experienced it because then he would have killed it with a sledgehammer
    Oh no Per, you're kidding me! It's not something I had ever experienced myself, and the board was operating in a relatively quiet room - doing video encoding/audio editing even!

    There is one thing that comes to mind - you mention that downloading/hrping seems to "affect" the coil whine in some way. Are you using the onboard Gigabit NV Ethernet by any chance? If you are, try, for a moment, turning it off, and seeing if that makes a difference. That was the one function I had never used, since I had initially had some 'issues' with the Nvidia LAN driver, and had simply installed an old 10/100 PCI card instead.

    Let me know of your findings - now I'm really curious!
    Last edited by tiresias; 04-06-2006, 12:59 PM.

    Comment

      #3
      Re: Epox 9NDA3+ with exsessive coil noise

      Yea, I think it is the CPU that affects it or something else, i.e. a voltage regulator is running at a different voltage and thereby frequency than it was when you was using it... (I'm using a 3500+ Clawhammer)

      I'm using the onboard NIC but disabling it did not affect it, for which I am thankful because that is a really good NIC, not connected via the PCI bus but directly to the Northbridge... (hrping to localhost still caused the coilnoise with the LAN cable unplugged BTW)
      "The one who says it cannot be done should never interrupt the one who is doing it."

      Comment

        #4
        Re: Epox 9NDA3+ with exsessive coil noise

        Originally posted by Per Hansson
        Yea, I think it is the CPU that affects it or something else, i.e. a voltage regulator is running at a different voltage and thereby frequency than it was when you was using it... (I'm using a 3500+ Clawhammer)
        That could well be the case, since the Clawhammer will by default 'demand' a higher Vcore than the 90nm Winchester part I was using. Has the mysterious "C1 error" been giving you any grief yet by the way?

        Comment

          #5
          Re: Epox 9NDA3+ with exsessive coil noise

          Originally posted by tiresias
          That could well be the case, since the Clawhammer will by default 'demand' a higher Vcore than the 90nm Winchester part I was using. Has the mysterious "C1 error" been giving you any grief yet by the way?
          Just put the system together two days ago... Ran it with one stick untill last night

          Now running 2 1GB sticks and Prime 95, atleast it has passed about 6 hours now...
          "The one who says it cannot be done should never interrupt the one who is doing it."

          Comment

            #6
            Re: Epox 9NDA3+ with exsessive coil noise

            Passed over 18 hours of Prime 95 now...

            So I took out the soldering iron and replaced all Hermei LZ 6.3v 3300uF caps with Rubycon MBZ 6.3v 3300uF caps...

            Thinking this would fix the coil whine issue I also removed the hot glue from the middle coil, however to my disappointment the coil noise was not affected at all, it did not even change frequency

            Also I notice that this coils windings are not as neat as the others; can that affect it? (the two coils on it's sides are silent...)

            So, I will ask google but do not have big hopes, i.e. I want to know what causes coil noise/whine and how to really fix it, not just supress it with hot glue or nail polish... (Okay, I know it is ripple current on the voltage lines that causes the ferrite core to move which causes the noise, but really, why only this coil, why why why? etc etc etc)




            Attached Files
            Last edited by Per Hansson; 05-11-2014, 08:05 AM. Reason: Offsite images uploaded due to problems with host
            "The one who says it cannot be done should never interrupt the one who is doing it."

            Comment

              #7
              Re: Epox 9NDA3+ with exsessive coil noise

              Hmm, interesting

              Overclocking the processor affects the coil whine... Gonna see if I can find a noise free frequency...

              But; running f@h effectivley kills it, quite odd because for example that tool "hrping" also caused 100% CPU load but that only made the coil noise worse...
              "The one who says it cannot be done should never interrupt the one who is doing it."

              Comment

                #8
                Re: Epox 9NDA3+ with exsessive coil noise

                If you carefully put your finger on a coil which is not hot, and the noise changes, this is the coil causing some of the noise (usually 'fizzing'). Don't do this test on high-voltage transformers! I found the component making a 'fizzing' sound in a satellite receiver: the DC-DC converter for the tuner voltage.

                Usually, a little bit of high-temperature silicone (non-aceatic so it does not attract ants) on the coils and board silences the 'fizzing' sound.
                My first choice in quality Japanese electrolytics is Nippon Chemi-Con, which has been in business since 1931... the quality of electronics is dependent on the quality of the electrolytics.

                Comment

                  #9
                  Re: Epox 9NDA3+ with exsessive coil noise

                  japlytic; as I said I already know which coil it is, the middle one in my pics...

                  If I put my finger on it and push it a little it changes frequency, but it does not stop...

                  And I put a very big ammount of hot glue over it (now removed) which also did not help...

                  It is not your averge coil noise, I estimate in excess of 45dB!!! I.e. it is so high that it causes my ears to hurt within minutes...
                  "The one who says it cannot be done should never interrupt the one who is doing it."

                  Comment

                    #10
                    Re: Epox 9NDA3+ with exsessive coil noise

                    I made a recording of the noise, it is of course a bit disorted but not too bad

                    It starts with the computer playing Winamp muted, then I type something on the keyboard which you can hear; I start hrping, when the sound changes frequency that is hrping... Then I stop hrping so it's back to winamp, and the last few seconds I resume work on f@h and the coil goes almost 100% silent...

                    To set your speakers to the right volume you should be clearly able to hear when I type on the keyboard, i.e. probably allot higher than you might first think...
                    Attached Files
                    Last edited by Per Hansson; 05-11-2014, 08:06 AM. Reason: Offsite MP3 file uploaded as ZIP due to problems with host
                    "The one who says it cannot be done should never interrupt the one who is doing it."

                    Comment

                      #11
                      Re: Epox 9NDA3+ with exsessive coil noise

                      Good morning Per

                      Thanks for the recording - I can feel your pain now! Or rather, hear the motherboard's atrocious ultrasoprano voice (for those who can't hear the recording properly, it's a loud hissing-grating sound at about 7.5 kHz.)

                      Interesting that running "folding" (implying increased current draw) should decrease the perceived noise. Would Prime95 also make it "go silent"? What about an even heavier load like BurnK7?

                      As for the fact that only a single coil in the three-phase VRM is doing this - could this mean there's something 'strange' with either of the two MOSFETs in that particular phase?

                      I'm even more intrigued by the fact that the MB hasn't given problems yet with your 2 DIMM configuration - particularly since I had put it through some pretty gruelling troubleshooting (as specified in the "for sale" thread ), but without success. And, by extension, during some of this troubleshooting, I had my face surprisingly close to the motherboard itself, but never even heard the smallest hint of coil whine!

                      Comment

                        #12
                        Re: Epox 9NDA3+ with exsessive coil noise

                        Haha, yea, it is interesting...
                        I have to run the memory at 2T though, else it will not even post... But thus far it has been fine, not a single lockup... And I'm not exactly stock any longer either...

                        My idea is also that there is something wrong with that phase in the VRM, but how to diagnose it...

                        The Distributed.net client also makes it quiet, as does Prime95 if I test with small fft's, not if I also test memory (lower cpu load...) BurnK7 also made it quiet... The thing is that that hrping app also (according to taskmgr atleast) loads the CPU max but it mades to coils whine... even if I am running any of the previous apps in the background (at lower priority of course)

                        Attached Files
                        Last edited by Per Hansson; 05-11-2014, 06:53 AM. Reason: Offsite images uploaded due to problems with host
                        "The one who says it cannot be done should never interrupt the one who is doing it."

                        Comment

                          #13
                          Re: Epox 9NDA3+ with exsessive coil noise

                          Looking good there - a K8 at 2.65GHz will give you some excellent performance for gaming or any other single-threaded application.

                          Congratulations, by the way, on the very professional-looking recap with the MBZ - very smart of you to have done the DIMM VRM ones too, since, for all we know, that could have been the source of my mysterious 2-DIMM problems.

                          As for that loud singing... who knows? The scandinavian air perhaps? The fact that it's in the same country as the most beautiful women in the world?

                          .

                          PS - A word about this particular board's overclocking abilities - I've found that it could, without trouble, reach 267FSB without requiring any kind of chipset voltage increase. With the memory divider at /166, it would, very stably, run the Crucial BallistiX at 2-2-2-5-1T at over 220MHz! They were single-sided DIMMs though, so that could explain your necessity to use 2T with the 1GB parts.
                          Last edited by tiresias; 04-08-2006, 07:09 AM.

                          Comment

                            #14
                            Re: Epox 9NDA3+ with exsessive coil noise

                            Hmm, just did some test with a multimeter on the 3 VRM chips that control vcore:

                            idle 1,569v
                            load 1,640v (Burn7K)

                            That is a delta of 0,071v seems maybe a tad much but still ok to me for a 3 phase VRM layout... I notice that the "hrping" tool only increases the vcore to 1,609v so I guess it is not 100% CPU afterall... (distributed.net=1,628v and f@h=1.618v) On the "output" side of the coil the voltage is 0,02v lower then on the "input"

                            By default this mobo overvolts the CPU by 0.05v (the lowest vcore setting in the BIOS)
                            "The one who says it cannot be done should never interrupt the one who is doing it."

                            Comment

                              #15
                              Re: Epox 9NDA3+ with exsessive coil noise

                              Originally posted by Per Hansson
                              Hmm, just did some test with a multimeter on the 3 VRM chips that control vcore:

                              idle 1,569v
                              load 1,640v (Burn7K)

                              That is a delta of 0,071v seems maybe a tad much but still ok to me for a 3 phase VRM layout... I notice that the "hrping" tool only increases the vcore to 1,609v so I guess it is not 100% CPU afterall... (distributed.net=1,628v and f@h=1.618v) On the "output" side of the coil the voltage is 0,02v lower then on the "input"

                              By default this mobo overvolts the CPU by 0.05v (the lowest vcore setting in the BIOS)
                              The idle / load drop on this board is incredibly common (I think only rev 2.1 boards are affected from my reading at the time).

                              There's a mod which you can do that helps ALOT with this strange flucuation. My vcore only swings 0.02v either way, and very rarely, now

                              http://bone.servebeer.com/vdroop.jpg if you're interested.

                              Comment

                                #16
                                Re: Epox 9NDA3+ with exsessive coil noise

                                Hmm, interesting, but that picture does not look to me like a 9NDA3+?

                                I guess this is a vdroop mod?
                                "The one who says it cannot be done should never interrupt the one who is doing it."

                                Comment

                                  #17
                                  Re: Epox 9NDA3+ with exsessive coil noise

                                  Originally posted by Per Hansson
                                  Hmm, interesting, but that picture does not look to me like a 9NDA3+?

                                  I guess this is a vdroop mod?

                                  Well, I assure you it is..... because that's an actual picture of my board :P

                                  I guess so :P On my board the original difference between idle and load vcore was around 0.08v from what I remember.... This wasn't constant fluctuation either, it was *only* when going from an idle -> load state, or load -> idle state that it would drop / rise. All other times the vcore remained pretty stable.


                                  EDIT: Thought I might just state again, that mod is *only* for a v2.1 board. I think the layout is slightly different on lower revision boards.

                                  Comment

                                    #18
                                    Re: Epox 9NDA3+ with exsessive coil noise

                                    It's not a 9NDA3+, doesn't seem like Rev2.1 either:

                                    http://www.geocities.co.jp/SiliconVa...A3/9NDA3R2.htm


                                    [Edit: I just read Markie's post, where he says it is! I guess Epox has different PCB designs for the various revisions of this board ]

                                    ...

                                    Increasing Vcore with increasing Icore is something I've measured before, albeit only in particular boards. My kid's Asus A7V600-X has a severe case of this (in the range of +0.07-8V), while the Tyan S2865 I'm using remains, as far as I can measure with this multimeter, entirely unchanging.

                                    What is actually, electronically speaking, happening to cause the VRM's output voltage to increase with greater current draw?
                                    Last edited by tiresias; 04-09-2006, 07:39 AM.

                                    Comment

                                      #19
                                      Re: Epox 9NDA3+ with exsessive coil noise

                                      Originally posted by tiresias
                                      It's not a 9NDA3, not Rev2.1 either:

                                      http://www.geocities.co.jp/SiliconVa...A3/9NDA3R2.htm


                                      [Edit: I just read Markie's post, where he says it is! I guess Epox has different designs for particular revisions ]
                                      Yes, it is :P I'll get a bigger picture of my board if you want. But I assure you, it is a 9NDA3+ v2.1 :P

                                      Infact, I can even circle the section my picture is showing from a picture on the link you provided :P
                                      Attached Files
                                      Last edited by markiemrboo; 04-09-2006, 07:42 AM.

                                      Comment

                                        #20
                                        Re: Epox 9NDA3+ with exsessive coil noise

                                        Originally posted by markiemrboo
                                        Yes, it is :P I'll get a bigger picture of my board if you want. But I assure you, it is a 9NDA3+ v2.1 :P
                                        Hi Markie,

                                        We certainly believe you!

                                        It just suprises me somewhat to see that - since I had, on the 2nd February 2005, shortly after buying a Rev1 9NDA, contacted the UK "representative" at epox.org (now dead, just a redirect in its place, it seems), and he had "assured me" (in salesman's terms, now I realise) that the only difference between Rev1 and 2 was a different CPU Voltage regulator chip.

                                        PS. All Sanyos it seems too - is that your own recap or was Epox having a good hair day?

                                        PPS. They even gave you bigger, chunkier, FETs! That's just not fair, is it?
                                        Last edited by tiresias; 04-09-2006, 07:48 AM.

                                        Comment

                                        Previous 1 2 3 4 template Next
                                        Working...
                                        X