Intel DP55WB motherboard powering on then off repeatedly no POST

[poot36]

I have a Intel DP55WB motherboard that when the power supply is turned on it will turn on then off repeatedly and will not POST (this is without the power switch on the front of the case being pressed, pressing it does nothing).  I have tried unplugging the CPU 12V connector and this behavior will not change.  If I remove the CPU from the socket it will not do this sequence but the front power switch will not do anything either.  I am currently testing with out RAM to see if I can at least get it to beep but it is not doing that either.  When I connect a POST card to the board I quickly see a FF and 00 appear just before the power cycle turns it off.  I found out that the motherboard tells the power supply to shut off with the power good signal by taking that line high.  I can force the board to stay on by holding the power good signal low or the green power supply on signal low and the board will stay running but will still not boot.  Got any ideas as to what is causing this?  My power supply is good.

[poot36]

Thanks for the quick reply.  However I do not think that will solve my problem as I have tried disableing the bios chip by shorting out the data and clock lines and nothing changes.  I could try reading the bios chip and seeing if it matches any of the files online for comparison.  My biggist concern with that are most bios images online are not a complete image but just the parts that need to be changed (eg no boot block or serial number included).  I think what is happening is there is a voltage on the board that is measuring out of spec and the board is instructing the power supply to shut down to protect itself but I have no way of confirming that because it stays on for around a second or so and then turns off again.  Although I would think that the board would still turn on without the cpu but maybe this does not work on the boards that for some reason turn on for a second or two when power is first applyed to them and then they shut off waiting for the power switch to be pressed?  I am assuming that effect is for some sort of motherboard preconfigeration routine related to PXE boot or the UEFI bios?  I have seen this effect on older Intel Core 2 Duo boards as well so maybe this is an Intel thing?

PS The bios chip is a SMD SOT 8 package that I am not afraid of removing.

[poot36]

Good news, I did not have to remove the bios chip to read it.  I used one of those 8 pin SOIC clips and used a parallel port based programmer that uses 4 resistors and the SPIPGM program to read the chip.  I am going to attach the bios dump to this message to see if you can find anything wrong with it.  It will be a rar file, you will have to change the file extension from zip to rar.  Here is a link to what bios version I think I am running: http://downloadmirror.intel.com/18170/eng/WB_0144_ReleaseNotes.pdf

[Skimask]

Who...is going to debug a dump from a BIOS chip?
Tell me that.
Who, in there right mind, would really WANT to spend their time analyzing a dump from a BIOS chip?  Who?
Would you do it for somebody else?
Does what you are asking the masses to do for you make any sense at all?
Hello?
Really?
An identical new motherboard is only ~$100 or so.  A non-Intel motherboard with practically the same specifications (usually better spec's) can be had for around ~$50.
In short, I'll debug your BIOS dump for you for $100.
In return, I'll buy a new motherboard, confirm it works properly, pull the BIOS, dump the hex file, and email it back to you.
Does that sound reasonable?
Does what you are asking sound even remotely reasonable?
Are you numb from the neck up?

[poot36]

Calm down, I never expected someone to disasemble the code, just compare it to an identical bios version dump from a working board.  If I remember correctly there is a Intel bios modification tool out there I could try to load up the bios file with that and if it errors out then it is either A) corrupt or B) expecting the larger bios file from the Intel website.  I got the motherboard + CPU (Core i7) but no RAM for free from the recycling bin at work so I am more trying to learn why this is happening and how I can try to fix it.  Ie. fun learning project.  If I can get it working I am sure it will be way faster then my current desktop that has a Core 2 Duo E8600 in it.  Would the next steps be to test all of the caps, MOSFETs, and 3 pin voltage regulators?  I think this is a power regulation issue because I can see the motherboard attempting to read the bios chip based on the very quick trace I see on my old Tektronix 465 scope.  Would there be a way to capture what is going on between the motherboard and the bios chip?  I only have a Arduino and a PC parallel port to work with.

[ludzinc]

I got the motherboard + CPU (Core i7) but no RAM for free from the recycling bin at work ....

It's dead Jim.

Your time would really be better spent getting a part time job and earnign the cash for a new Motherboard.

Good on you for wanting to learn, I commend you on that, but there comes a point where you need to call it quits.

[niflheimer]

Either the VRMs are fried or the controller for the mosfets is. The fact that it behaves differently with the atx12v disconnected confirms this.

Could you fix it ? maybe. Is it worth fixing ? Hell no. It would take a lot more hours of work than what's it worth. Just go on ebay and find a compatible mobo ...

[wraper]

You can try to just program the bios downloaded from the website. Even if it lacks serial, etc, usually it will work anyway (at least will be able to boot computer). Also, unlike in notebooks, often there is no any serial or other data. Therefore there is a very good chance that what you download is a full image. If downloaded image is not full, it is easy to cook the right image by combining dumped and downloaded images together (replacing blank areas in downloaded image).

[wraper]

Also, did you try to replace the power supply? It can be stupid as that (been there). Even if apparently it does work. Are you sure that you have not bent any pins on the CPU socket? It is very easy to do and hard to notice.

[poot36]

Re: ATX 12V power: it responds the same way with and without that connector plugged in or not.  I have looked at the bios image on the Intel website and it is 3.1Mb and my bios is only 2Mb.  I suspect that there is extra code in there for reflashing from a USB or CD or floppy.  I have tried the motherboard with 3 different power supplies and get the same result.  I am beginning to suspect the voltage regulators/MOSFETS near the center of the board due to there been no voltage out of them when I kept the power supply from turning off.  When I did this the cpu voltage regulators were putting out 0.9V so I think they are working.  I also tested the regulators near the ram and the 24 pin power socket and there was voltage coming out of them to.  Would the northbridge or southbridge chips control all of the rest of the voltage regulators besides the cpu ones?  Or are they controled by individual chips just like the cpu?

[niflheimer]

Did you try that cpu in another board?

[poot36]

Sadly I do not have another motherboard to test the cpu with or another cpu to test this board with.  Is it normal for a modern motherboard to not turn on with the front case power switch if there is no cpu in the board?  I know old P4 boards would still turn on with no cpu installed and some would beep or even talk to you saying no cpu.  Tonight I will try to measure all of the voltages from all of the regulators and VRMs on the board.

[wraper]

I have looked at the bios image on the Intel website and it is 3.1Mb and my bios is only 2Mb.

Don't know particularly about intel but often you can use such file as is. Additional data usually is located after actual bios data. My programmer just cuts off the part which exceed the size. You can just pick some free HEX editor and compare dump and download file. Then cut off the end if necessary.

[wraper]

BTW did this motherboard and CPU work together before? Or you got them separately? It could be that bios is just too old and do not support particular CPU.

[wraper]

I think a 5 series north bridge should support all LGA771/775 processors.

Where did you get that this motherboard has LGA775 socket? Also P55 chipset does not support LGA775 CPUs at all.
CPU support in chipset means nothing if there is no support in Bios.

[poot36]

I assume that this motherboard and cpu were working at one time because they came together and had an equal amount of dust on them.  By the way I found the short, it is on the 1.5V rail.  At least I think it is 1.5V, I got this and all subsequent readings by forcing the power supply on. I removed inductor L10BV and the board stopped power cycling.  However this has caused all other VRM rails to stop producing voltage so I assume what ever this rail powers runs all of the other rails including the cpu.  The rail that measured 1.5V shows 53.5 ohms to ground and on diode mode shows 0.026V.  These readings are with the negative probe on ground and the positive on the output of the inductor.  I have also tried supplying the rail with a 1.5V battery and it goes back to power cycling.  The amperage draw from the battery was around 250 to 350mA.  I do not think that the short is here but I am not sure.  When I would jumper the power supply before I removed the inductor the other rails that I measured were putting out 1.5V for the ram, 0.7V for the cpu, 3.3V and 5V for I assume other logic.

[wraper]

The rail that measured 1.5V shows 53.5 ohms to ground and on diode mode shows 0.026V.  These readings are with the negative probe on ground and the positive on the output of the inductor.

This is not a short. If you measure the resistance between CPU power rail and GND (with CPU installed), likely will be surprised even more.

[wraper]

I have also tried supplying the rail with a 1.5V battery and it goes back to power cycling.  The amperage draw from the battery was around 250 to 350mA.

Considering the voltage, that's a whopping 0.5W power consumption on this rail .

[poot36]

Ok, a small update.  I have made a special harness that allows me to select what voltage rail from the power supply goes to the mobo.  Using this I have found that the 5V rail causes the board to power cycle.  I was powering the 1.5V rail once again from the battery to help isolate the rails.  The 5V rail has a resistance of 12 ohms or so to ground and I have noticed that if there is power on just this rail (no other power supply connections psu jumpered to turn on) the green standby light lights up and there is 3.2V or so on the power button header and the standby power supply pin.  Also the amperage drawn from that rail in this state is close to 500mA.  I have no clue if this is normal or not.  I also have noticed that if the bios jumper is on pins 2,3 (advanced mode) the beeper makes a little click and does not do this in normal mode (1,2) or recovery mode (no jumper).  I have also noticed that the power cycling happens faster with the cpu 12V plugged in.  I do know that the 5V rail drives the switcher chip for the cpu voltage regulator and the 12V rail provides the 5V to the usb ports (if this rail is disconnected the usb ports are around 4.3V).  I think that the 3.3V rail powers the 1.5V regulator that runs the mobo shutdown process.  I have also seen the 1.5V rail when powered by the battery and having the psu connected in various ways with my harness caused current spikes of up to 1 amp or so.

[wraper]

As for me, your experiments with power rails seem to be useless waste of time and potentially harmful. In what way stopping power cycling says anything about the fault? Once you disconnect a power rail, board just stops to live/work properly = no cycling = board just don't try to restart normal operation anymore. 
Don't you understand that power cycling is not the fault by itself? It is the mechanism how motherboards try to resume normal operation.

[Stonent]

If you remove the power from the motherboard and short the green wire to black, does the power supply supply good voltage at the right power levels?
You might try putting as many hard drives and optical drives as you can on the drive connectors to load it down and test the power stability.

[poot36]

I think we can call this one dead.  I was testing the fan controller and power supply chip and my probe slipped and shorted out the 12V and 5V or 2.5V monitoring rails releasing the magic smoke from the chip.  The board still acted the same though so it was not that chip signaling it to shut down.