Announcement

Collapse
No announcement yet.

Dead Asus P5P800-VM

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

    Dead Asus P5P800-VM

    Hello all

    I'm new to the forums.
    I am trying to troubleshoot an old LGA775 Asus P5P800-VM. This MB comes from an old 35mm Cinema sound processor which I am collecting for nostalgic reasons! As this is a custom system, it may work with another board or it may not. If I cannot fix it I'll get another identical one on the internet but for now I am doing my best to fix it.

    Little history.

    - System failed to boot intermittently. I thought it was one of the expansion boards but in the end it failed to boot altogether even with nothing connected.
    - Inspecting the board on the bench I found that the 12V CPU line is immediately drawing 8A as soon as the PSU is switched on. The VRM mosfets gets ludicrously hot, the CPU gets warm but it takes several seconds. The current is so high that I can hear the chokes whining.




    - If I remove the CPU, there is no activity and the Mosfets stay cool.
    - I have to power on the PSU manually as the Power On pins don't work.
    - I ordered a new CPU (Intel Celeron) but issue is unchanged
    - Pins on the CPU socket seem ok.
    - I have a thermal camera, I do not see anything getting hot besides two of the mosfets - the other 6 (three phases) are still getting fairly hot though.
    There are also four 4Ohm SMD resistors close to another mosfet in the middle of the board which are getting very hot - considering they are tiny. But I read the correct resistance when I test them. I removed the mosfet which is nearby to test but nothing changed.



    - From my understanding the mosfets are not shorted. On the CPU VRM ones I can read approx 1V on each phase. Voltage is not exactly the same on all the phases though.
    - I have tested all other mosfets for shorts and nothing seems to be out of the ordinary, and none of them is getting hot.
    - I do not read a short on any power line. But If I understand correctly the resistance I am reading is a bit low.
    - I have tested a different PSU.
    - I have replaced some of the capacitors on the VRMs. Those are KZG caps and I read around they may go bad without bulging. They tested ok once removed and indeed new ones did not help.
    - I checked the phases with my scope. I do not know much on the subject but I thought they were supposed to be staggered so that each phase would only work 1/3 of the time. On my 4-ch oscilloscope they indeed pulse but they run simultaneously.

    So in short my VRM gets really hot only when the CPU is in. But the CPU is not bad. When the CPU is plugged in, there is something drawing 8A@12V (96W) that's pushing the mosfets to 100+ degrees C in a matter of seconds.

    I am a bit short of ideas here. Would anybody have any suggestions of what to try next? I appreciate any input you may have!

    Thank you!
    Tony
    Attached Files

    #2
    Re: Dead Asus P5P800-VM

    When you say voltage is not quite the same on all VRMs which side of the inductors are you measuring? The CPU side or the Mosfet side?

    Without the CPU fitted what voltage do you see on Vcore? It could be that without the CPU fitted Vcore is not running at all

    I may be going out on a limb here but possibly if there is something wrong with the timing between the gate drive on the high side and low side mosfets then just maybe something like this could happen

    Also on those pics it looks like just one phase VRM that is getting hot, so possibly leaky high side or low side FET on that phase, or some problem with the gate drive to that phase... or a shorted turn in the inductor on that phase?
    Last edited by dicky96; 05-05-2020, 03:52 PM.
    Follow me on YouTube
    ------------------
    Learn Electronics Repair
    https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

    Comment


      #3
      Re: Dead Asus P5P800-VM

      The boardviews might help .

      https://openboardview.org/
      Attached Files
      Last edited by SMDFlea; 05-06-2020, 03:53 AM.
      All donations to badcaps are welcome, click on this link to donate. Thanks to all supporters

      Comment


        #4
        Re: Dead Asus P5P800-VM

        Hello Dicky96 and SMDFlea

        Schematics! I would have never hoped for those! Thank you!

        Dicky69
        To answer your questions
        I measured both. It is a bit difficult as the voltage slightly change over the first few seconds and I cannot keep the board powered for long as the components get really hot. Below are my measurements:

        Phase 1 - CPU side 1.025V - Mosfet side 0.995V
        Phase 2 - CPU side 1.04V - Mosfet side 1.04V
        Phase 3 - CPU side 1.04V - Mosfet side 1.06V

        One phase VRM is indeed the hottest but the others are getting fairly hot too - it's just that the colour scheme does not really show that once one component skyrockets to 120 degrees C!

        I have checked the PWM modulation coming to the drivers - my knowledge on this matter is limited so I'm not sure whether what I found is correct (please note that one phase is half the amplitude as the scale is different: my old Fluke O'Scope has some limitations on channels 3 and 4). EDIT: well actually no, the scale is 0.5V and the pulse should be 4 times higher than channel 1 and 2, not smaller. Interesting. I need to check to which phase is that channel connected to.

        The frequency of those pulses change a few time when I power up - so it changes the whining sound I hear and also the voltage I read.

        Without CPU I do not see the driving PWM and hence I read no voltage.

        It's hard to say as there is a large area getting hot when I power up but I believe the 4V 680uF caps are the first one getting hot when power is applied.

        Would you think that removing the hottest phase mosfets could be a good test to see if the other two phases work ok by themselves? Would I risk to damage the CPU or would it just receive an incorrect - lower- voltage?

        Cheers for your help again!
        Attached Files
        Last edited by tony359; 05-06-2020, 04:10 PM.

        Comment


          #5
          Re: Dead Asus P5P800-VM

          Looking at the thermal images with a better colour scale, I can see that there are mainly three mosfets from two different phases that are getting very hot.

          I still need to find a way to save uncompressed colour pics but the attached should give you a better idea.

          Checking the mosfets with the scope does not show obivious issues but, again, I am a learner

          Any ideas?
          Attached Files

          Comment


            #6
            Re: Dead Asus P5P800-VM

            Hi Tony

            Phase 1 - CPU side 1.025V - Mosfet side 0.995V = this suggests to me there is a lot of current flowing - which we kinda of know already - as the peak pulse voltage is being held down


            Unsolder the MosFETS from phase 1 and swap them with phase 2 or 3

            What happens now?
            Follow me on YouTube
            ------------------
            Learn Electronics Repair
            https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

            Comment


              #7
              Re: Dead Asus P5P800-VM

              @dicky96

              Thank you!
              Which one is phase 1 for you on the last B/W picture? The one on top with both mosfets getting really hot?

              Comment


                #8
                Re: Dead Asus P5P800-VM

                Let's call phase one the hot one - doesnt' really matter what phase it is, though the schematic/boardview would tell you, just swap the hot mosfets with one of the cold phases
                Last edited by dicky96; 05-08-2020, 02:17 AM.
                Follow me on YouTube
                ------------------
                Learn Electronics Repair
                https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                Comment


                  #9
                  Re: Dead Asus P5P800-VM

                  thank you for your input!

                  I've swapped the Mosfets between two phases and the issue follows the location and not the components. Does that mean the issue is with the PCB or with some other component? What puzzles me a bit is that if it depended on the PCB (ground plane shorted) I feel all three phases would behave the same?

                  Comment


                    #10
                    Re: Dead Asus P5P800-VM

                    OK so the MOSfets are good most likely
                    Swap the inductors. What happens now?
                    Follow me on YouTube
                    ------------------
                    Learn Electronics Repair
                    https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                    Comment


                      #11
                      Re: Dead Asus P5P800-VM

                      I'll check tomorrow.

                      What if I removed all three inductors? I believe that would isolate the mosfets from the CPU but I should still read a voltage at each inductor input pin. Is there any negative effect I should be aware of?

                      Comment


                        #12
                        Re: Dead Asus P5P800-VM

                        Just to check, you did swap both the high side and low side FETs of the bad channel yes?

                        Without the inductors (and capacitors) the circuit would not work - its the energy stored in the inductor when the high side Mosfet turns on that generates the output voltage by charging the output capacitors when the high side Mosfet turns off and the inductor discharges. The pulse width controls the amount of energy stored, and therefore the output voltage.

                        This will explain how the circuit works - what you have is three synchronous converters. You don't need to understand the maths, the basic principle is enough to know.
                        https://www.maximintegrated.com/en/d...es/6/6129.html

                        Most likely without them the VRM would not run at all, as the controller would detect a fault - but other results could occur I guess.

                        The reason I asked you to swap the inductors between the bad phase and a good one is because there is a possibility one inductor has a shorted turn. This would ruin Q of the inductor, so in effect it would reach saturation (the point at which it can no store no more magnetic energy) and effectively act like a piece of wire rather than a coil causing a lot of current to pass in that phase. Another way to picture this is the shorted turn becomes a short circuit secondary winding, and the rest of the inductor acts as a primary winding. So it's now acting like a transformer with a dead short on the secondary - and again, a lot of current is going to flow....

                        As all phases drive the same load (Vcore), the problem can't a short on Vcore or they would all get hot.

                        Another possible is you have a situation where both high side and low side gates on hte faulty phase are being turned on simultaneously (basically a faulty controller IC). You would have to use a dual channel scope to check for this, and compare against a good channel.

                        I can't think of another scenario at the moment but there are some clever guys on here more experienced than me who probably can

                        Rich
                        Last edited by dicky96; 05-09-2020, 02:49 AM.
                        Follow me on YouTube
                        ------------------
                        Learn Electronics Repair
                        https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                        Comment


                          #13
                          Re: Dead Asus P5P800-VM

                          Hi Rich,

                          Thank you for taking the time to explain me the situation!

                          I have checked the PWM signal coming from the VRM controller and it seems ok - I am not sure I understand why there is a big gap between sets of pulses though. I posted the pictures earlier on - they're attached again for you!

                          Unfortunately swapping the inductors didn't change anything.

                          I tried making a better thermal picture of what's happening. Indeed all mosfets become quite hot - there is a 20 degrees difference between phase 3 (very hot reaching 120 degrees on both mosfet) and phase 2 (quite hot, reaching about 100 degrees).

                          Phase 1 is weird, one mosfet goes as high as Phase 3 but the other one is 20 degrees behind.

                          When I say 120 degrees C I mean after about 10 seconds. I then power off before the whole thing melts down. I have a powerful fan blowing air on the PCB while I do these experiments.
                          Attached Files

                          Comment


                            #14
                            Re: Dead Asus P5P800-VM

                            Well at least that eliminated the inductor coil

                            In a synchronous converter there must never be a situation where both FETs are turned on at the same time

                            https://www.researchgate.net/figure/...g_fig4_3130479

                            If you look at the above fig.7 you will see there is a short delay after the high FET gate turns off and before the low FET turns gate on, and vice versa. These times are shown as td1 and td2 in the diagram. So at that time both high and low side FETs are off.

                            It is hard to see on your waveforms if that is the case. I may actually have a P5P800-VM at my workshop, I certainly had one at some time or other. If I still have it I will have a look to see what waveforms I get on mine.

                            Other possible causes of your problem - controller running at the wrong frequency or PWM drive pulse too wide so the inductor goes into saturation - at which point a very high current will flow through the MosFETS.

                            By the way, the reason the pulse width duty cycle is low (as you asked) is because each phase is firing in turn, kinda like a Gatling gun - and for much the same reason too. Thermodynamics.
                            Last edited by dicky96; 05-09-2020, 07:37 AM.
                            Follow me on YouTube
                            ------------------
                            Learn Electronics Repair
                            https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                            Comment


                              #15
                              Re: Dead Asus P5P800-VM

                              Hi Rich,

                              The scope pictures I posted before were taken from the output of the controller chip. That should be the PWM signal driving each mosfet pair. If I understand it correctly it seems ok: pulses are happening in sequence.

                              I have done some more investigation following your advice.
                              I have probed the signal arriving on the Gate of both mosfets which is labelled R_LG1 and R_HG1 (High and Low I reckon) on the driver chip. The below is the driver chip with the outputs I mentioned.



                              I do see two opposite square waves but I am not sure it is what is supposed to be as the pulses happen simultaneously.

                              This happens on all three phases - but what I call phase 1 (the one with only one mosfet overheating) looks different.

                              Let me show what I see.

                              This is phase 2-3. I call Phase 3 the one that massively overheats and 2 the one adjacent.



                              I made a video of it. What I do is just to short the POWER_ON line on the PSU.

                              https://youtu.be/OJ_rV7vwGHM

                              This is phase 1. I call phase 1 the one where only one mosfet gets very hot



                              Video for phase 1: https://youtu.be/-q90-hoOMy0

                              Back to the controller output (Phase 1,2,3 on channel 1,2,3) there does not seem to be a timing problem - in my limited understanding.



                              Just to be safe, the other day I tested all the 680ohm/4V caps on the VCore line and they all tested fine - both capacitance and ESR.

                              Any other ideas? Am I missing something obvious here?
                              Huge thank you for your help so far!
                              Attached Files

                              Comment


                                #16
                                Re: Dead Asus P5P800-VM

                                It looks to me like the three phase PWM controller does not like something that is going on with the MosFET driver on phase 1.

                                I currently have something similar going on with an Asrock Z87, though in that case the PWM controller apparently detects some problem and continually shuts down and restarts Vcore

                                I ordered some MosFET drivers and MosFETS for that one but they are not here yet. I think you have reached the point where you have to swap the driver chip from Phase 1 with another phase to see what happens, or order some MosFET drivers (and while you are at it get a PWM controller), and see if changing either of these could fix the problem.

                                Possible it is not the driver chip but some SMD resistor/capacitor connected to it. Maybe a quick resistance check from all pins of the MosFET driver to ground and to 12V on a good channel compare with the bad one may help.

                                Which MosFET drivers do you have?

                                I did check if I have one of the P5P800-VM here now but I don't
                                Last edited by dicky96; 05-11-2020, 06:39 AM.
                                Follow me on YouTube
                                ------------------
                                Learn Electronics Repair
                                https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                                Comment


                                  #17
                                  Re: Dead Asus P5P800-VM

                                  Hey Rich,

                                  I swapped phase 1 Mosfets with phase 3 on my previous attempt and the issue followed the location - that is, phase 3 mosfets installed in phase 1 location were behaving like original phase 1 mosfets. So as you say it may be something else.

                                  Just so that I understand: Am I reading my own readings the wrong way? I thought my scope was telling me that High and Low Mosfets were switched simultaneously - which is not good - but now I am looking at those pictures again, maybe I am just being silly and reading them the wrong way. See below: assuming Red is one Mosfet and Green is the other, is my interpretation correct? I initially thought that one of the Mosfets was getting a negative signal and so I thought they were switching on simultaneously...



                                  I have checked the six Mosfets as you recommend. I do not observe any difference to ground or to 12V. The resistance value they read is the same.

                                  My Mosfets are 04N03LA and 60T03GH. Those are long discontinues. Would you be able to help me finding a suitable replacement?

                                  For the 60T03GH I found this

                                  https://uk.rs-online.com/web/p/mosfe...gU3849ERaBpI-o

                                  For the 04N03La I found this

                                  https://uk.rs-online.com/web/p/mosfe...Yb6EygMdYB1UdI

                                  They sell in packs of 10 so it would be £26 of components and I'd like to make sure they are suitable - any cheaper alternative is very welcome!

                                  Controller chip is ADP3180 https://cdn.badcaps-static.com/pdfs/...5a7e781990.pdf
                                  Drivers are ADP3110 https://cdn.badcaps-static.com/pdfs/...7c6912a1a3.pdf

                                  Those are discontinued too. I can find the Driver - again, a bag of 25!

                                  I've noticed that the controller is getting a feedback signal back from all the three phases. I'll probe it later and see what happens. But as you say Phase 1 and its weird switching graph is probably where I should look into.

                                  Thanks for having a look if you had the same MoBo - appreciated!

                                  Have a great day!
                                  Tony
                                  Attached Files

                                  Comment


                                    #18
                                    Re: Dead Asus P5P800-VM

                                    Update.

                                    As the signal arriving the drivers seem correct, I swapped driver 1 with driver 2. The issue followed the DRIVER - now phase 2 has a boiling mosfet and a cool one.

                                    I shall start with replacing the drivers then. However, I can only find ADP3110A available. I cannot really say whether the 3110 and 3110A are interchangeable or whether I should replace all three? Can anybody help me?

                                    https://cdn.badcaps-static.com/pdfs/...7c6912a1a3.pdf
                                    https://cdn.badcaps-static.com/pdfs/...3ebe29a6a2.pdf

                                    I'll keep you posted. Please feel free to share any other course of action!

                                    Thanks!
                                    Last edited by tony359; 05-13-2020, 10:00 AM.

                                    Comment


                                      #19
                                      Re: Dead Asus P5P800-VM

                                      Update

                                      The driver has arrived. I have replaced it and the good news is that the Mosfets are now properly driven - the pulses seem correct on all three pairs.

                                      The bad news is that the motherboard still doesn't boot up. I still hear that whining noise when I power up. Mosfets are not the hottest components now, inductors are around 90 degrees C. Mosfets stay at 70-80 degrees. I read 5-6A drawn at boot, not 8A anymore.

                                      Being a custom box, the board comes with a special little daughterboard which connects to the Front Panel terminals. The purpose of that little board is to have the system turning on and off with a rocker switch at the front of the unit (ON-OFF).

                                      The manufacturer of the box claims the MB has a custom bios on it - I doubt it as they were not producing many of those boxes. But I wonder whether there is any sort of modification anywhere that changes the way the board switches on and off. I could not test this when the board was working so I am assuming lots of things here.

                                      Still, this is a step forward. I will now do more tests and see what happens.

                                      I have noticed that when I attempt to boot up, the CPU gets warm-ish then it cools down and then again (I keep my finger on it to monitor). This cycle takes approx 15 seconds. It's like something is happening but I am not sure what.

                                      Comment


                                        #20
                                        Re: Dead Asus P5P800-VM

                                        Originally posted by tony359 View Post
                                        The bad news is that the motherboard still doesn't boot up. I still hear that whining noise when I power up. Mosfets are not the hottest components now, inductors are around 90 degrees C. Mosfets stay at 70-80 degrees. I read 5-6A drawn at boot, not 8A anymore.
                                        That's still very hot.

                                        I suggest replacing all three MOSFET drivers at once and all of the MOSFETs too. While at it, also check all of the MOSFET Gate resistors and caps, if any.

                                        I have an Intel DG33TL motherboard that for the longest time kept blowing MOSFETs on me. I'd replace one, run the board for a few minutes, only to have another one die. Doing resistance measurements between the Gates of the MOSFETs on each phase revealed some differences in the readings, but nothing that was obviously bad. So in the end, I changed all of the MOSFET drivers and upper MOSFETs. This got the motherboard working finally.

                                        That may or may not be the solution to your motherboard, of course, but it's something still worth trying IMO, given how hot your VRM runs.

                                        Also, is that discoloration on the MOSFET(s) near the AGP port or just bad lightning? I suspect there may be something going on there as well.

                                        Last but not least, you may want to replace those Panasonic FL 4V, 680 uF caps, even if they read OK. After a while (may power-on hours), I find that any liquid electrolytic cap on LGA775 mobos with the caps under the heatsink like that will eventually fail or develop high ESR or just get flaky. Using solid polymer caps like Sanyo SEPC or Nichicon LF or similar should fit under there and work great. I typically go with 2.5V, 820 uF.

                                        Oh, and do you have a PCI POST card? That should help us see if at least the mobo is trying to boot or not.

                                        Comment

                                        Working...
                                        X