Announcement

Collapse
No announcement yet.

Asus P5S800-VM no POST

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

    Asus P5S800-VM no POST

    Hi guys, here is the latest motherboard I'm trying to fix just for the hell of it

    ASUS P5S800-VM

    It turns on/off with the ATX power switch, but it doesn't boot (POST 00 on my diagnostic card)

    I tried another P4 CPU, no difference

    I tried with or without DDR. No difference

    I have Vcore 1.4V, Vram 2.6V, 3.3V to the BOIS, 5V, 12V etc. I'm using a known good PSU

    I reflashed the BIOS - no difference

    From looking around with my scope I think i have no system clock. I'm looking at the RTM363-655 which I am sure is the clock generator. The crystal doesn't seem to be oscillating (just some high frequency noise on one pin probably from the PSU). I have 3.3V supply to this chip. Nothing seems to be happening on the outputs.

    I couldn't find the exact datasheet. I found this one https://datasheetspdf.com/pdf-file/7...K/RTM363-210/1 for the RTM363-210

    Does anyone know if the -210 and -655 version of RTM63 have the same pin out & functions?

    Are these clock generator ICs known to fail?

    best regards
    Richard
    Attached Files
    Last edited by dicky96; 10-02-2019, 07:00 AM.
    Follow me on YouTube
    ------------------
    Learn Electronics Repair
    https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

    #2
    Re: Asus P5S800-VM no POST

    Here`s some boardviews ,it might help with your repair.
    Attached Files
    All donations to badcaps are welcome, click on this link to donate. Thanks to all supporters

    Comment


      #3
      Re: Asus P5S800-VM no POST

      Thank you. That's much appreciated
      Follow me on YouTube
      ------------------
      Learn Electronics Repair
      https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

      Comment


        #4
        Re: Asus P5S800-VM no POST

        Sorry to be a pain, I have openboardview but when I open one of those files i just see a large square with a few capacitors (or similar) on it. How do I get the boardview with all the components?


        I had a look at the RTM363-210 datasheet again - link in post #1

        Assuming that is basically the same as the RTM363-533 then i have #PD is high 3.3V so according to the datasheet the IC should be active.

        #RESET is low but the datasheet says that is an open collector output

        According to the datasheet Block Diagram 4, there should be control inputs for #PCI STOP and #CPU STOP but those don't appear to relate to any pins on the IC.

        On the Xtal I have 1V on one pin and 1.5V on the other pin but it isn't oscillating.
        Clock Outputs such as pin 21 and 22 are low. So it isn't working.

        Should I now order an Xtal and RTM363-533, or is there some other possibilities I am missing here?

        Also i notice there are some control registers and a serial interface to program the RTM363 output frequencies. But how does that happen unless there is a clock to drive the CPU in the first place, unless it is supposed to oscillate at some frequency like 100MHz by default at power on?
        Last edited by dicky96; 10-03-2019, 03:35 AM.
        Follow me on YouTube
        ------------------
        Learn Electronics Repair
        https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

        Comment


          #5
          Re: Asus P5S800-VM no POST

          BoardViewer will open .asc files .
          Attached Files
          All donations to badcaps are welcome, click on this link to donate. Thanks to all supporters

          Comment


            #6
            Re: Asus P5S800-VM no POST

            OpenBoardView works fine with them too. You may need to flip the board pressing the space bar.
            OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

            Comment


              #7
              Re: Asus P5S800-VM no POST

              I've come back to look at this again. Thanks for the board view files - I got them to open in Boardviewer and they helped me.

              As far as I can see there is no system clock, I've looked around the clock generator IC RTM363-655 using board view as a guide and it is enabled (reset is high) it has power on all the pins it should have according to board view, the crystal is oscillating, but nothing comes out of the IC. All the pins (other than xtal) are static either 0V or 3.3V

              Unfortunately I don't have a datasheet for RTM363-655. I do have datasheet for RTM363-210 but it is clearly not the same chip, even some voltage supply pins do not match the board view. Also I can't find a supplier for this 655 chip on aliexpress. I found one on alibaba who would sell me five at a reasonable price but the postage is killing the deal, and this repair is only for educational purposes.

              So I could possibly assume that chip is faulty. However I apparently have another problem. I've noticed that a power FET on the board gets quite hot and from board view it is clearly feeding VTT which I measured at approx 1.2V.

              I measured the resistance from VTT to ground and it is 2.5 ohms ,with no CPU fitted.

              I then got a couple of other LGA 775 boards I have lying around and on those ones VTT to ground is 125ohms on one and 187 ohms on the other with no CPU. Having said that, one of those has an intel P45 chipset (125ohm) and the other has intel G31 (187ohm). This motherboard is SIS and I don't have another one lying around to compare with.

              As VTT also supplies the north bridge and south bridge I am wondering if the low resistance reading means that one of those is faulty? The south bridge is getting warm but not hot. Or is it as feature of the SIS661X/SIS964 chipset on this board?

              I also can't understand how a faulty north bridge or south bridge would stop the system clock from running, and it seems odd to have two different faults.

              Can anyone advise what resistance to ground I should see on VTT? Did I find a useful diagnosis test on boards what will power on but not boot (0000 on post card)?


              best regards
              Last edited by dicky96; 02-11-2020, 05:33 AM.
              Follow me on YouTube
              ------------------
              Learn Electronics Repair
              https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

              Comment


                #8
                Re: Asus P5S800-VM no POST

                You'd need a oscilloscope with a decently high bandwidth to be able to check those clock signals (they'll most likely be 24MHz, 33MHz, 48MHz, 66MHz, 200MHz and maybe others). Clock generator can fail as any component can fail, but not really the first thing I'd suspect.

                Can you post a picture of the board identifying the VTT MOSFET you're talking about?

                I'm not sure I have a board with those specific northbridge/southbridge but I'll try to take a look. I don't really like SiS chipsets…

                Unfortunately there can be a lot of reasons for a board to power on but not to start POSTing. Usually in this case I check CPU VCore, CPU reset signal if I can find it easily, I flash the BIOS, and I try to feel around the board for anything getting abnormally hot. You already checked CPU VCore and tried to flash the BIOS though.
                OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

                Comment


                  #9
                  Re: Asus P5S800-VM no POST

                  I also suggest taking off the stock Intel pushpin heatsink and testing with the heatsink just resting on the CPU. These stock Intel pushpin heatsinks are known to warp boards badly and can sometimes damage the BGA on the CPU socket. So once you take off the push pins, try booting the board. Also try applying some pressure on the CPU heatsink as you try to boot the system. Hopefully this might show if you have a bad socket BGA or not. It's rare, but it can happen.

                  As for the very hot MOSFET that regulates Vtt... if it's outputting the proper voltage (which 1.2V sounds like it is) and ripple is in spec (will need scope to measure that), then it's probably regulating properly and not getting overloaded.

                  Comment


                    #10
                    Re: Asus P5S800-VM no POST

                    Actually I was using my 20MHz LG scope but I will look again tomorrow with my 100MHz Tektronix scope.

                    Vcore and VRAM are OK. I will have a look to see where the CPU reset signal comes from. Probably the SIO?

                    I've added a pic as requested

                    I agree the hot MOSFET is likely outputting the correct voltage. a 2.7 ohm load would be about 500mA at 1.2V so it is not excessive - it just seems very low resistance compared to the LGA775 Intel chipsets I compared with

                    The FET doesn't get very hot but it's still 65C according to my temp probe and that is with the board doing nothing.

                    I'll try the suggestions with the heat sink
                    Attached Files
                    Last edited by dicky96; 02-12-2020, 09:57 AM.
                    Follow me on YouTube
                    ------------------
                    Learn Electronics Repair
                    https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                    Comment


                      #11
                      Re: Asus P5S800-VM no POST

                      Ok so I found a 661GX-M7 rev 1.1 board and the lowest resistance to ground I could find on northbridge/southbridge is 58 ohms. Board dead cold and no power at all since I didn't plug anything in, no CPU/RAM either.

                      I also found 2x 649-M2 rev 1.0B boards and one has 35 ohms the other 30 ohms as the lowest resistance to ground around northbridge/southbridge.

                      CPU reset signal is typically issued by northbridge with Intel chipsets. I don't know SiS architecture well enough to tell if it's the same, I think there are some noticeable differences in the power/reset sequence.
                      OpenBoardView — https://github.com/OpenBoardView/OpenBoardView

                      Comment


                        #12
                        Re: Asus P5S800-VM no POST

                        Thanks momaka, piernov - you helped me some more

                        I tried running the board without the cpu heatsink. It isn't a problem as the CPU only goes up to 45C even after the board has been powered up for 20 mins

                        I tried pushing on the CPU when starting the board but it makes no difference

                        I noticed on my POST test card that the CLK LED lights extremely dimly

                        I don't have any clock, checked on my 100MHz scope. Using boardview to guide me I checked some clock signals. CPU Clock, SATA Clock, NBH Clock - all missing. The 14MHz crystal to the Clock Gen IC is running

                        I also checked some reset signals to the CPU (with CPU fitted)

                        CPURST# (LGA pin 532)
                        DBRESET# (pin 63)
                        HTRESET# (pin 154)

                        all low

                        I noticed that I can turn the ATX PSU on with the power on switch but can't turn it off again by holding the switch down.

                        Without the heat sink on the North bridge it gets quite warm 47C. The South bridge is cool.

                        And given the low 2.7R resistance to ground on VTT with no CPU fitted, the hot VTT MOSFET and what Piernov said about the CPU reset coming from the north bridge, and the fact you also found much higher resistance on VTT with similar SiS boards....

                        All this to me is pointing towards a faulty north bridge BGA. And maybe that is somehow also stopping the RTM363-655 clock gen from running (even though CLKRST# on pin 4 is high)? The lack of a datasheet for this IC is not helping me here :-(

                        I'm thinking if I remove the north bridge and the resistance on VTT goes up a lot this may also indicate a problem there

                        I can get a couple of new SiS 661FX BGA at reasonable price so I would like to try fix this to prove if I have managed to diagnose it correctly or not

                        Either that or replace the RTM363-655

                        Is my diagnosis well thought out?

                        cheers
                        rich
                        Last edited by dicky96; 02-13-2020, 06:31 AM.
                        Follow me on YouTube
                        ------------------
                        Learn Electronics Repair
                        https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                        Comment


                          #13
                          Re: Asus P5S800-VM no POST

                          Originally posted by dicky96 View Post
                          All this to me is pointing towards a faulty north bridge BGA.
                          ...
                          Either that or replace the RTM363-655

                          Is my diagnosis well thought out?
                          Hard to say for me. I haven't dealt with too many SiS boards. Actually, I think I only have one, and it's a finicky piece of turd socket 462 that I never built anything with it.

                          So I can't say how likely that SiS NB is likely to fail. If it was an nVidia nForce 4 or later, I would be all my money on that. ATi/AMD 770/780 chipsets also fail somewhat often.

                          Also, it's not really the BGA that fails on these chipsets above - at least not the BGA between the board and the chip PCB. Rather, (and especially for nVidia nForce 6150 and newer) it's the chip PCB to chip core BGA that fails. Some say the older generation of chips with wire interconnects don't have this issue. But I have seen nVidia GeForce 4 and ALI chipsets fail, despite being built on this older technology.

                          All that said... if you are very comfortable with replacing the NB and can easily get a replacement, then go for it.
                          I personally would instead just "sleep on it" (the issue), until I get more ideas what to try. I find this to work quite well in the past with boards I'm just trying to repair for fun. Usually I wait long enough so that I forget everything I've tried/tested. Then I do every single test I can think of all over again. And not surprisingly, that's how I've often caught simple fixes that somehow I overlooked the first time. Thus, if you don't feel comfortable replacing the NB, then perhaps consider going this route with the repair then.
                          Last edited by momaka; 02-23-2020, 08:07 PM.

                          Comment


                            #14
                            Re: Asus P5S800-VM no POST

                            I got a bit further with this but still haven't found where the problem is

                            I desoldered the north bridge and afterwards I had a direct short on VTT. I used my bench PSU to put 1.2V onto VTT and using some freezer spray found the short was a 1uF SMD capacitor near to the north bridge (C3243). I removed that thinking I had found the problem but then I was back to 2.5 ohms on VTT to ground. I do wonder if that capacitor went short due to heat while removing the north bridge

                            I then decided to remove the south bridge as the board view also showed that was powered from VTT. But it made no difference

                            I then used my bench supply to put 1.2V on VTT again. I have 0.41A going somewhere but there are no VTT 'load' devices left on the board. Unfortunately 0.41A isn't enough to get anything to warm up and reveal itself.

                            I looked on the board view and removed the electrolytic caps on VTT but it makes no difference

                            I then had a look at every Resistor and Resistor network in parallel on VTT and calculated that the resistance from Vtt to ground should be 35R - not 2.7R! In fact one of the resistor networks is not fitted so it would be higher that 35R

                            I then removed everyy smd capacitor between VTT to ground. And still I have 2.7 ohms and 0.41A going somewhere.

                            There were two inductors from VTT to other voltage rails but both of those read slightly lower resistance on the VTT to ground side - so the 'short' is on VTT not the other rails.

                            It can't be the PowerFET that supplies VTT from 3V because no pin on that FET go to ground

                            So now I am either down to a resistor or resistor network which has failed low ohms which seems very unlikely, or a near short between internal PCB layers/tracks which from my experience working at ICL all those years ago does happen

                            Either way this repair was supposed to be for fun/education only and it's getting to not be much fun now so it's back on one side. It looks like I am so it could end up in the minus some useful looking PowerFETS and a BIOS chip
                            Follow me on YouTube
                            ------------------
                            Learn Electronics Repair
                            https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                            Comment


                              #15
                              Re: Asus P5S800-VM no POST

                              Originally posted by dicky96 View Post
                              I then removed everyy smd capacitor between VTT to ground. And still I have 2.7 ohms and 0.41A going somewhere.

                              There were two inductors from VTT to other voltage rails but both of those read slightly lower resistance on the VTT to ground side - so the 'short' is on VTT not the other rails.

                              It can't be the PowerFET that supplies VTT from 3V because no pin on that FET go to ground

                              So now I am either down to a resistor or resistor network which has failed low ohms which seems very unlikely, or a near short between internal PCB layers/tracks which from my experience working at ICL all those years ago does happen
                              Have you checked the controller that drives the FET that generates the VTT... or any other controllers/shunts that use VTT to produce other rails? It's unlikely for that to be the cause, but one of the last few things that hasn't been checked... besides possibly having more ceramic caps on the board somewhere and one of them being shorted.

                              Originally posted by dicky96 View Post
                              Either way this repair was supposed to be for fun/education only and it's getting to not be much fun now so it's back on one side. It looks like I am so it could end up in the minus some useful looking PowerFETS and a BIOS chip
                              I'd save all the power/PWM controllers too and maybe some of the coils.

                              Also, if you want to have fun and can't find the short... take several very thick wires (14 AWG at least, and 3-4 strands minimum), then connect the CPU V_core rail to VTT and fire up the motherboard. Hopefully the CPU V_core should have enough "oomph" to clear/smoke out whatever is causing the short circuit. And who knows - if the CPU V_core doesn't have good protection, maybe you will get a nice smoke show. The few times I've tried this, I've just had my PSU shut down or a MOSFET pop - some shorted caps seem to be able to take infinite current and still not care to heat up or smoke.

                              Comment


                                #16
                                Re: Asus P5S800-VM no POST

                                Originally posted by momaka View Post
                                Have you checked the controller that drives the FET that generates the VTT... or any other controllers/shunts that use VTT to produce other rails? It's unlikely for that to be the cause, but one of the last few things that hasn't been checked... besides possibly having more ceramic caps on the board somewhere and one of them being shorted.
                                VTT comes from a FET that is basically acting as a series voltage regulator. Not a Buck regulator circuit. So there isn't a controller. Nor any ground connected to the FET.

                                I've checked everywhere on the boardview that VTT goes. The only 'loads' appear to be the CPU. NB & SB which I removed. There are resistors and resistor networks connected directly from VTT to ground and I don't understand why they need to put a load on VTT that way, but when I identified them all on the boardview and then used ohms law to work out the parallel resistance it should be around 35 ohms

                                There are two rails produced from VTT via little smd inductors, but both read slightly higher resistance on the other side of the inductor so the problem is on VTT


                                Also, if you want to have fun and can't find the short... take several very thick wires (14 AWG at least, and 3-4 strands minimum), then connect the CPU V_core rail to VTT and fire up the motherboard. Hopefully the CPU V_core should have enough "oomph" to clear/smoke out whatever is causing the short circuit. And who knows - if the CPU V_core doesn't have good protection, maybe you will get a nice smoke show. The few times I've tried this, I've just had my PSU shut down or a MOSFET pop - some shorted caps seem to be able to take infinite current and still not care to heat up or smoke.
                                Well not quite that momaka, though it sounds fun.... but it did occur to me that I could use my bench supply to crank VTT up to several volts and maybe the 'short' would reveal itself that way. Because it's actually a low resistance 2.5 ohms rather than full on short I can't get enough current in at 1.2V to heat anything up. It just sits there draining 0.41A which for sure ain't right.

                                Seeing as there is no NB, SB or CPU fitted now and VTT doesn't go anywhere else according to the boardview.... I am thinking that doing this wouldn't actually damage anything either, as long as I stayed below the rated value of the electrolytic caps on VTT.

                                Sounds like a plan to me I'll report back with the results tomorrow
                                Follow me on YouTube
                                ------------------
                                Learn Electronics Repair
                                https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                                Comment


                                  #17
                                  Re: Asus P5S800-VM no POST

                                  This motherboard is no more. it's gone to meet it's maker. But do have a nice stash of MosFETs and other useful bits from it
                                  Follow me on YouTube
                                  ------------------
                                  Learn Electronics Repair
                                  https://www.youtube.com/channel/UCFX...R8UZ2vg/videos

                                  Comment


                                    #18
                                    Re: Asus P5S800-VM no POST

                                    R.I.P. ASUS P5S800-VM

                                    Ah well... I'm sure over time those scavenged parts will help many "ill" boards work fine again. I did the same with a few older nVidia video cards (GeForce 7000 series) that either displayed artifacts or no video and just couldn't get fixed no matter what. Now the parts from one of them has helped me repair no less than 10 other video cards.

                                    Comment

                                    Working...
                                    X