Announcement

Collapse
No announcement yet.

MSI K8 Neo4

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    MSI K8 Neo4

    Earlier this week my pc died and I'm a bit lost in my diagnosis of the fault. Hopefully someone will read all this long post and can give me some advice.

    The board is a MSI K8 Neo4 Platnium (MS-7125) and it has served me very well since I got it second hand. I do want to build a new pc but I'd like to do that gradually so I can research hardware more carefully so I'd like it if the K8 could serve me a bit longer.

    As I'm a bit lost I'm wondering if bad caps could be the problem or if something else has gone bad. I've done lot's of googling and reading these forums without much luck on id'ing the cause.

    What happened is I booted the pc and logged in and then left the room. Returning a few minutes later the pc was dead and would not turn on again. Nil nada zip nothing.

    After some reseach I tested the power supply and made a discovery that many apparently have made. With the power supply not connected to the mobo in any way shorting the ps-on pin to ground turned the power supply on. When just the 24pin connecter is hooked up to the mobo the fans work but when the atx12 4 pin connecter is hooked up then nothing.

    I specifically measured the ps-on pin when the atx12 4 pin was hooked up to the mobo.
    - When the wall power is first turned on the ps-on pin is at 5v like it should be. When shorting the ps_on pin to ground (fpc button or just shorting the pins on the mobo) the ps_on pin goes initally towards zero but then goes to 14v! I've never seen reports with that in them.

    The power supply is a Usicase, which I suspect is just another low end PS branded for a system builder. I opened it up and there was one slightly bulging cap which I replaced. Still got the same result. I also tried with another PS on hand, also not good quality (Shaw) with the same results.

    So that brings me to the mobo. I've inspected it very carefully and can not see anything obviously wrong, no bulging caps or blackened mosfets, no scorch marks anwhere. There is one smd ceramic cap (I believe, one of those smd brown with silver ends things) right next to the 4 pin connector which has a bit of the brown chipped off but I think it has been like that for ever.

    On the board near the CPU is 5 Rubycon 6.3v 3300uf caps, 4 brown (marked KZG) 16v 1500uf caps and a black w/ gold writing 6.3v 1800uf cap that has a M inside a square on it. Around the board in various spots are various small caps (green marked KZE, brown marked KMG and some black with the M logo). Again, none of the caps show any sign of bulging or pushing the bung out. I do know that this isn't the only failure mode.

    So the big question before I try recapping it is could failed caps cause the problem I'm having (like causing the ps_on pin to go to 14v when shorting it to ground) or should I be looking for something else which I'm not sure I'll have the skill to tackle.

    Thanks,

    - Tim

    #2
    Re: MSI K8 Neo4

    Originally posted by ml2 View Post
    the ps_on pin goes initally towards zero but then goes to 14v!
    i hate to rain on your parade but dont be too optimistic about reviving this mobo. when the mobo gets exposed to this high of a voltage spike, its most likely fried.
    Originally posted by ml2 View Post
    The power supply is a Usicase, which I suspect is just another low end PS branded for a system builder. I opened it up and there was one slightly bulging cap which I replaced. Still got the same result. I also tried with another PS on hand, also not good quality (Shaw) with the same results.
    looks like the psu failed and took the mobo along with it. had an fsp crapxon psu do that to one of my amd k8 agp mobos, an asus a8v. if u wanna fix this system, u may need to fix/replace both the psu and mobo. u might wanna post some pictures of the inside of the psu so we can see what kind of a psu it is but most likely from what happened, its a junk psu and not worth saving.

    u'll have to replace the psu with a good quality japcap psu or sumthing if u want the system to last. no point fixing the mobo and not changing/improving the psu because the psu will just kill the caps (or something else unfixable) on the mobo *again* and u'll end up having to recap/fix the mobo *again* after a year or so. guess u found out the hard way NOT to use a junk power supply...
    Originally posted by ml2 View Post
    There is one smd ceramic cap (I believe, one of those smd brown with silver ends things) right next to the 4 pin connector which has a bit of the brown chipped off.
    those smd mlcc caps go short if they are damaged or chipped in any way. u'll have to pull them with either a soldering iron with a fork tip or a hot air gun. soldering a replacement one back on is very tricky due to its tiny size. u'll have to be very dextrous with your hands.
    Originally posted by ml2 View Post
    a black w/ gold writing 6.3v 1800uf cap that has a M inside a square on it. Around the board in various spots are some black with the M logo.
    those should be panasonic FJ caps. u should be able to see FJ written on one side of the caps. your mobo appears to use all jap caps from what u posted so that is good. however, the ripple from the psu can kill even good caps like on my asus a8v.
    Originally posted by ml2 View Post
    4 brown (marked KZG) 16v 1500uf caps
    this japcap series is known to be problematic. they'll have to be replaced with equivalent series from other manufacturers.
    Originally posted by ml2 View Post
    So the big question before I try recapping it is could failed caps cause the problem I'm having (like causing the ps_on pin to go to 14v when shorting it to ground)
    its not suffering only from failed caps. when the psu failed, it most likely caused a series of cascading failures down the chain. when computers stop working, they often trigger a series of multiple failed components. therefore, u'll end up having to deal with more than just bad caps.

    since this is a mobo subforum, create a thread and post pictures of the psu in the power supply subforum instead. ask if the psu is worth saving/recapping or "improving" or not. once the psu is either properly "fixed" or replaced, post back here about fixing the mobo.

    Comment


      #3
      Re: MSI K8 Neo4

      can you test your PSU ? PS_on the ground and check if you get correct 12V,5V,3.3V,-5V and -12V correctly ?
      Have a look :
      https://en.wikipedia.org/wiki/ATX

      Comment


        #4
        Re: MSI K8 Neo4

        Yes, I can test the PSU. Everything looks in range, except for the strange behavior of the -12v line. I've posted the results in my PSU thread.

        https://www.badcaps.net/forum/showthread.php?t=56973

        Comment


          #5
          Re: MSI K8 Neo4

          The voltages look normal for me..

          Comment


            #6
            Re: MSI K8 Neo4

            How about Vcore voltage ? are they around 1.2-1.3V ?

            Comment


              #7
              Re: MSI K8 Neo4

              hmm, I not exactly sure where to measure vcore from. I'm still very new to this side of things. I did read the measuring vrm thread but that didn't penetrate the skull too well.

              I've attached a pic of the vrm area of the K8 board. From what I understand after stepping down through the mosfets there I would find vcore somewhere.

              I did try measuring but I've most likely done it all wrong. With the ps-on pin shorted to ground all I got trying to measure the first fet closest to the 4 pin plug was about 3.2mV.

              I have decided to build a new system, starting with a good PSU but I'd still like to see if this one can be trouble shooted.
              Attached Files

              Comment


                #8
                Re: MSI K8 Neo4

                i circled the area in red where u can measure the vcore from. please enlarge the picture if u cant see the red ink.

                normally, measuring the vcore is done from the legs of the inductors or secondly, from the positive terminal of the caps. however, the inductors on your board are covered up and the only way to measure them is from the underside of the board like the caps. so the only way left to measure the voltage right side up is to measure from the body pad of the mosfets.

                be careful not to touch any other solder component with the multimeter probe that is not within the red area as it might short something out tho that is unlikely. i recommend measuring either the mosfets circled on top or the ones at the bottom. as they are the easier ones to reach. the middle one is obsured by caps and may be hard to reach.
                Attached Files

                Comment


                  #9
                  Re: MSI K8 Neo4

                  measuring Vrm....easy way....have a look on inductor (L) find the pin which is closest to your CPU....in there you can see VRM....how much do you get there ?

                  1.2V - 1.3V is still acceptable...depends on what CPU you have there
                  ChaosLegionnaire is right...
                  MS-7125 ===> on AMD CPU......could be Athlon 64 x2 or Phenom on your board there...
                  On my experience....I have athlon 64 x2 3600+ now under test....I can see it...
                  1.25V normal....

                  Comment


                    #10
                    Re: MSI K8 Neo4

                    Have a look on this list and see Vcore
                    https://en.wikipedia.org/wiki/List_o...icroprocessors
                    Is it the same with this one ?
                    http://www.evertek.com/largePic_All....-PB-R-unit.jpg
                    Last edited by bianchi77; 09-21-2016, 05:08 AM.

                    Comment


                      #11
                      Re: MSI K8 Neo4

                      Thanks.

                      Each of those 6 fets measured 2.8v, without the PSU even being on. I know the PSU I'm using is not very good at all but even I know that can't be a good sign.

                      It really looks like the board is trashed, and with 2.8v on those fets I guess the CPU would have been trashed as well.

                      I should be getting a new PSU this week for my new build (probably Corsair RM750X if I can find one). Any danger of damaging the new PSU by hooking it up to the K8 board? With a new good PSU I'll try getting measurements again.

                      It was an Athlon 64FX. Can't remember which one and don't have the box with me but I seem to remember the bios being set for 1.49v for vcore.

                      The board is almost exactly the same, mine has 4 extra sata headers that are down by the last PCI slot. You can see where they would be in the pic you posted.

                      Comment


                        #12
                        Re: MSI K8 Neo4

                        Code:
                        It was an Athlon 64FX.
                        To be exact, it's simple, could you please open the cooler fan, and clean up abit with isopropyl alcolhol the chip mark will be seen...
                        Normally you will see AMD Athlon xxxxxx sometimes you can see the FSB speed and voltage required there...but 2.8V is way above it's needed....about two times....

                        I reckon you have toasted your FET and your buck controller....the filter, inductors and capacitors ( passive ) mostly survive but I'm not sure for FET and buck controller....
                        There's a breakdown on buck controller, depends on your motherboard designer...it could be a FET driver after PWM generator for issuing command for voltage level....

                        From your PSU there above the voltage looked normal....
                        Can you see from those pins on your FET equal to 5V or 12V, high side FET...
                        can you check it....??

                        Have a look as well on Northbridge, Southbridge, DDR FET....normally about 1.8V on each....and Northbridge sometimes uses the same Vcore with CPU....

                        Comment


                          #13
                          Re: MSI K8 Neo4

                          Abit of PCB footprint that I've drawn, hope it will give you abit of image of voltage map on your motherboard, more or less
                          Attached Files

                          Comment


                            #14
                            Re: MSI K8 Neo4

                            There's a possibility that your motherboard suffered a voltage spike from your main electricity....
                            Perhaps there was a motor starting around you....and it's going through your main PSU for few milliseconds, but it did matter for your motherboard, because it's working on nano second, millisecond is a giant spike for it unfortunenately.....

                            Hope it will give you abit of picture....cheers...

                            Comment


                              #15
                              Re: MSI K8 Neo4

                              A simple test for your current PSU....pull it out from your motherboard and connect PS_ON, pin 16 on your PSU molex connector to GND....
                              Have a look on the picture.....you can see if its cooler fan rotating ( behind ).....it's doing the job a already.....
                              You can try with another PSU if you have, but basically they're doing the same job....different power it's only matter of different transformer and rectifier diode inside...if you see 500W....mostly about 80% to 90% is the real max output

                              500W === max about 400 to 450 VA...
                              your motherboard and CPU will not ask that much....even an old 250W ATX PSU will be sufficient enough to test mainboard and CPU only...
                              Attached Files
                              Last edited by bianchi77; 09-21-2016, 05:36 PM.

                              Comment


                                #16
                                Re: MSI K8 Neo4

                                GND = Ground....to be clear

                                Comment


                                  #17
                                  Re: MSI K8 Neo4

                                  Originally posted by ml2 View Post
                                  What happened is I booted the pc and logged in and then left the room. Returning a few minutes later the pc was dead and would not turn on again. Nil nada zip nothing.

                                  After some reseach I tested the power supply and made a discovery that many apparently have made. With the power supply not connected to the mobo in any way shorting the ps-on pin to ground turned the power supply on. When just the 24pin connecter is hooked up to the mobo the fans work but when the atx12 4 pin connecter is hooked up then nothing.
                                  That points to a blown MOSFET on the CPU buck regulator (what we often call the "CPU VRM" here on BCN). It may have been caused by bad power from the power supply or it could have been caused by those brown 16V 1500 uF KZG caps going dry and failing without bulging.

                                  Either way, just post some pictures of your motherboard, particularly of the components around the CPU area. Please make sure the pictures have good focus so we can see them well. Then we can go from there. To continue, you will also need a basic digital multimeter that can measure resistance and test for continuity/diode forward voltage drop (these last two are typically incorporated in the same function on cheap multimeters)

                                  Originally posted by ml2 View Post
                                  the ps_on pin goes initally towards zero but then goes to 14v!
                                  I doubt a PSU can show 14V on the PS_ON pin and still have the PSU power up / supposedly working. More likely there is something wrong with your multimeter or the point you are using for reference/ground... either that or perhaps using the wrong setting? Make sure you are measuring DCV (DC Voltage). And make sure the batteries in the multimeter are fresh. Better yet, try measuring the voltage of a new 9V square. You should get 9 to 9.6V. If not, your multimeter may need new batteries.

                                  Originally posted by ml2
                                  Each of those 6 fets measured 2.8v, without the PSU even being on.
                                  This further proves that something is wrong with your multimeter or the way you are using it. There's no way those FETs can have 2.8V on them with the PSU turned OFF.

                                  Originally posted by ml2 View Post
                                  The power supply is a Usicase, which I suspect is just another low end PS branded for a system builder.
                                  Sounds like it for sure.
                                  Post pictures of its insides if you'd like more feedback on it.

                                  Originally posted by ml2 View Post
                                  I opened it up and there was one slightly bulging cap which I replaced. Still got the same result. I also tried with another PS on hand, also not good quality (Shaw) with the same results.
                                  And that confirms the problem is CPU VRM -related and not a PSU a problem... though the fact that you are using cheap PSUs isn't helping either. Once we get the board running, consider getting something better for the PSU.

                                  Originally posted by ml2 View Post
                                  On the board near the CPU is 5 Rubycon 6.3v 3300uf caps, 4 brown (marked KZG) 16v 1500uf caps and a black w/ gold writing 6.3v 1800uf cap that has a M inside a square on it. Around the board in various spots are various small caps (green marked KZE, brown marked KMG and some black with the M logo). Again, none of the caps show any sign of bulging or pushing the bung out. I do know that this isn't the only failure mode.
                                  The 5 Rubycon are probably MBZ or MCZ series. MCZ is a little heat-sensitive, but overall they are good caps. MBZ series are reliable.

                                  The four brown 16V 1500 uF "KZG" are made by United Chemicon (aka just Chemicon). KZG is a series from them that fails very often, along with KZJ and TMV. However, their other series like KZE and KY are excellent - very reliable caps. The brown KMG is a general purpose series, but they are also fine. As for the black caps with an "M" logo: those are probably Matsushita (Panasonic) - again, extremely reliable caps as well. On motherboards, the most likely series you will see from Panasonic are FJ, FJS, or FL.

                                  Originally posted by ml2 View Post
                                  So the big question before I try recapping it is could failed caps cause the problem I'm having (like causing the ps_on pin to go to 14v when shorting it to ground) or should I be looking for something else which I'm not sure I'll have the skill to tackle.
                                  We first need to find what component is causing the short-circuit on CPU VRM. Good chances it's a shorted MOSFET, since your computer won't even power ON with the 4-pin 12V CPU power connector connected.

                                  Comment


                                    #18
                                    Re: MSI K8 Neo4

                                    Sounds like it for sure.
                                    Post pictures of its insides if you'd like more feedback on it.
                                    Something like this one :
                                    http://www.the-computer-problems-gur....yBhnIDyS.dpbs

                                    Sometimes there are some bulging capacitors inside if it's an old PSU....easy fix....

                                    Comment


                                      #19
                                      Re: MSI K8 Neo4

                                      Actually, I just found the OP did post pictures here:
                                      https://www.badcaps.net/forum/showthread.php?t=56973
                                      ... and it is indeed a crappy power supply.

                                      Originally posted by bianchi77
                                      Sometimes there are some bulging capacitors inside if it's an old PSU....easy fix....
                                      Not always. If it's a crappy PSU like the one above, new caps won't really help much. With only a single 1000 uF capacitor per rail, your motherboard and HDDs are bound to be drowned in ripple and noise. Might as well plug in the motherboard in your wall socket and be done with it.

                                      PSUs like the one above are hopeless. And they were designed and built that way from the factory. They only work enough for a little bit to make the end user think everything is okay and they got a great deal on a really cheap PSU.

                                      Comment


                                        #20
                                        Re: MSI K8 Neo4

                                        Gents, I must apologise, I am such a tool this week I was reading mV and not even realising it.

                                        So to recap with correct values, with the ATX12 4 pin plug connected to the board nothing worked; no life from the CPU and the PSU would not turn on, (ps-on pin would go down to about 14 mV when shorted to GND, not go up to 14V as I had said).

                                        When measuring the FET's as instructed by ChaosLeginonnaire I would get a reading of 2.8mV, not 2.8V so the PSU still wouldn't turn on.

                                        Now we get the interesting part for today. Today I bought my new PSU (EVGA SupaNova G2 750W) for my new build and hooked it up to the K8 board to take the measurements again. This is what happened. C11 decided to go out in a nice blaze of glory. This is the cap that had a chip of the brown covering missing that I mentioned in my first post but again I'm pretty sure it had been like that for years. It's right next to the atx12 4 pin plug.

                                        I'll be gutted if that little smd cap is all that was originally wrong, I probably could have replaced it but getting the value of it I would have not had the faintest idea. Could that little cap have stopped the whole vrm from working full stop?

                                        PSUs like the one above are hopeless. And they were designed and built that way from the factory. They only work enough for a little bit to make the end user think everything is okay and they got a great deal on a really cheap PSU.
                                        momaka, that "little bit" was at least 10 years. I got the thing free, and in the perhaps 6 years I've had it it hasn't faltered once, until maybe now if it was indeed responsible for killing the motherboard. I'm not saying it isn't crap (now that I know a bit more about PSU's), it is ,but that crap lasted a lot longer than I would have thought.

                                        bianchi77, I found the CPU box today, it's an Athlon 64 FX (FX-55) from 2004 so the vcore would be 1.50V, not that that matters now.
                                        Attached Files

                                        Comment

                                        Working...
                                        X