[SOLVED] GIGABYTE GA-H81M-DS2 REV: 3.0 Constant Power Cycling

[taxxin]

Board:

• GIGABYTE GA-H81M-DS2 REV: 3.0
• I have a working GA-H81M-S1 REV: 2.0 here for Comparison,
• CPU is a Pentiium G2220 for testing purposes
• Ram is 1600MHZ G skill 4gb pc12800

Symptoms:

• Motherboard cycles its power sequence regardless of whether there is a CPU AND/OR RAM installed. The working board stays on when powered with no CPU and RAM.
• NO POST messages received from plugging in the POST card into each of the PCIE 1x SLOTS
• Tested all the main power rails and voltage responses to the SIO chip reading normal
• Power is reaching the RST + Pin for the reset switch and PWR switch
• BIOS chip has been removed and reprogrammed with the latest GIGABYTE version bios VIA the CH341A china flasher.
• With the CPU 12v connector removed it will power on and shut off without looping.
• Jumping the POWER on at the ATX connector will keep the board in a powered state, nothing seems to be abnormal, PCH heats up normally, CPU Does not heat up. All the main power rails are present with corresponding voltages.

Things ive tried:

• Ive checked the power going into the power controllers and PCH hub and they all seem to be the same, the resistance i thought was low on the 3.3v rail (12.7Ohms) was exactly the same on the working board, the PCH is getting its 3.3v and 1.04V
• I have also swapped the bios chips, desoldered and loaded the latest firmware onto them. but still no luck.
• Tested the resistance on the coils on the cpu Vcore on the good one it climbs right up to 180k ohm and the bad one goes to about 140k but not sure exactly because it keeps climbing slowly.
• Booting with or without CPU+FAN and RAM makes no difference.

Things I havent tried:

• I dont have an oscilloscope so i havent been able to check the Crystal for clock generation.
• Checking the 25mhz clock generation built into the PCH chip.
• PCH gets pretty warm when
• All the VRIN inputs to the SIO chip read around 2.0 to 2.7v which i guess high means present? I havent checked every single resistor and capacitor around the chip for faults, I only did a quick check according to the picture attached with the main 3.3v in and voltage sense values.

According to the Power up sequence if power is getting to the RESET pin, then it would have completed its power sequence. Now the question is, what is causing it to cycle the power? is it a bad read on the PCH or CPU? (none of the pins seem damaged)

[re-atari]

The photo is too fuzzy to say with certainty, but I think I see a defective pin on the CPU socket in the top row of pins below the rectacular 'window'. The 2nd pin from the left immediately below this window looks to be either bent or missing. Check the attached photo, I have encircled the pin I'm referring to in red.

[taxxin]

Originally Posted by re-atari

The photo is too fuzzy to say with certainty, but I think I see a defective pin on the CPU socket in the top row of pins below the rectacular 'window'. The 2nd pin from the left immediately below this window looks to be either bent or missing. Check the attached photo, I have encircled the pin I'm referring to in red.

AMAZING! Great spotting, i checked my other board and that pin is present, the faulty board looks like its completely missing!

I will check it out and report back, thank you so much!!!!

[re-atari]

With a bit of luck this pin carries VCC or GND, in which case it's not a problem if the pin is missing. Plenty other pins carrying VCC and GND available om the socket. If it's a pin carrying an essential IO signal, on the other hand, it might just be the reason for this board not starting up.

If it's an essential pin, there's no alternative other than replacing the socket if you really want to revive the board. That will probably cost more than what the board is worth, though.

A description of the CPU pinout is necessary to determine the signal this specific pin carries. Sadly documentation describing the pinout of Intel CPU's is hard to find these days (speaking from own experience...)

[taxxin]

Originally Posted by re-atari

With a bit of luck this pin carries VCC or GND, in which case it's not a problem if the pin is missing. Plenty other pins carrying VCC and GND available om the socket. If it's a pin carrying an essential IO signal, on the other hand, it might just be the reason for this board not starting up.

If it's an essential pin, there's no alternative other than replacing the socket if you really want to revive the board. That will probably cost more than what the board is worth, though.

A description of the CPU pinout is necessary to determine the signal this specific pin carries. Sadly documentation describing the pinout of Intel CPU's is hard to find these days (speaking from own experience...)

BREAKTHROUGH, After seeing that the pin was completely snapped off, im pretty sure that the dumb kid that i sold it to was the culprit for damaging the pin and claiming RMA with me.

So i thought what the hell, this board is gonna be useless without the pin, so i had a quick look at some pictures on google to see what the back side of the socket looked like. they are not really balls but more like just blobs of solder holding the pins to the board, so i thought why the hell not? lets try an old scrap lenovo board and see if it works.

I doused the pins flux and hit it with my hot air station from china at 500deg F highest heat and lowest fan speed. After about 10-20 seconds i could start picking the pins out one by one, this was looking good. they were coming out still in the correct shape!!

So now the hard part was extracting the old pin, doused the pinhole and surrounding holes in flux and hit it with the hot air, i managed to remove the pin but slightly butchered all the other pins around it as well as the plastic socket hole that it sits in,

lastly i put a dab of SMD solder paste (LEAD MUHAHAH) on the end of the pin and inserted it carefully with the tweezers to as deep as it would accept, then i hit it again with the hot air and just firmly pressed it in with the tweezers. after all this work i realised that i had just put in a different kind of pin, the pin i replaced it with was much fatter in the body and looked like it would definitely make contact with other pints so spent about 20minutes trying to straighten out the remaining pins and used the precision of my asian eyes to try and make sure the pins resembled a grid as best as possible.

Finally i popped in the CPU and connected all the power and ram and POST card and started it up, here goes nothing i thought to myself as i shorted the pins on the power switch. it lit up with a 15 i thought sweet! its not just showing 0.0 it went through a few codes and it shut off, my heart sank for a bit but it cycled and then it went through more codes and went to a Ab! I couldn't believe my own eyes!. (Ab is booted and awaiting user input) so i slapped the heatsink on and vga cable and boom booted right into the bios!.

I am so happy it all worked out in the end it seemed impossible.

Big thanks to Re-Atari for the big breakthrough tip, thank you for going through my pictures and helping get to the end!!!

[dicky96]

Hmmm I am very interested in trying your technique for replacing broken pins....

[RJARRRPCGP]

Originally Posted by taxxin

my heart sank for a bit

While I don't know about remembering any power-cycling, (unlike an MSI during a BIOS update) Gigabyte motherboards are infamous for this, since at least the FX era looking like it's going to outright fail to boot, (nothing but power and fans) but then the BIOS suddenly comes up! (At least true since the AMD 970 era)
While OTOH, I have an early-2010s Gigabyte AM3 motherboard dating before the 970 chipset, where the BIOS boots like a lightning bolt!

But it's not just Gigabyte with other very-recent motherboards. My MSI B450 Tomahawk, also seems to have a slow UEFI BIOS boot! The strange thing is that, configuring the UEFI BIOS to pure-UEFI-boot-mode seems to cause this!

[taxxin]

Originally Posted by RJARRRPCGP

While I don't know about remembering any power-cycling, (unlike an MSI during a BIOS update) Gigabyte motherboards are infamous for this, since at least the FX era looking like it's going to outright fail to boot, (nothing but power and fans) but then the BIOS suddenly comes up! (At least true since the AMD 970 era)
While OTOH, I have an early-2010s Gigabyte AM3 motherboard dating before the 970 chipset, where the BIOS boots like a lightning bolt!

But it's not just Gigabyte with other very-recent motherboards. My MSI B450 Tomahawk, also seems to have a slow UEFI BIOS boot! The strange thing is that, configuring the UEFI BIOS to pure-UEFI-boot-mode seems to cause this!

Very interesting, Its quite surprising to see that the newer ryzen boards do have many things in common with the old 900 series FX chipsets as dicky pointed out. I have picked up a few of these boards mostly in need of repair. it seems that they around the age of failure.

As for bios the uefi ones are a pain to flash they take far too long, I just recently did a 990fx board bios flash its only 8mb but it was too long, the clip wouldnt maintain contact for that long and would corrupt the whole write. ended up having to desolder the chip and program it with chip flashing tool instead.

[re-atari]

Good job, another board saved from the landfill! Don't thank me, you did all the hard work
Looks like this is another viable way of replacing an individual pin, as long as it's placed at the inner or outer ring of the socket grid. I wouldn't like to replace more than a few pins this way though, especially if they're adjacent to each other.

I think the backup BIOS is the reason it took the board a fair amount of time to start up initially. If the main BIOS IC contains an error, the contents of the backup BIOS is copied to the main BIOS. May well have been the case here. AFAIK Gigabyte is the only manufacturer that builds a backup BIOS IC on its mainboards.

I personally only have positive experience with reading/writing flashroms using a clip and programmer (TL866, CH341). It is a bit fiddly to reliably connect the clip to the flashrom pins, but (up to now) I have had no contact problems. Having said that, I may just have been lucky with my clip (good build quality).

Some manufacturers like MSI build an (of course undocumented) ISP connector on their boards for reading/writing BIOS on its flashrom. Makes sense in a production environment or repair facility. Especially there you want to use reliable tools and methods.

[taxxin]

Originally Posted by re-atari

Good job, another board saved from the landfill! Don't thank me, you did all the hard work
Looks like this is another viable way of replacing an individual pin, as long as it's placed at the inner or outer ring of the socket grid. I wouldn't like to replace more than a few pins this way though, especially if they're adjacent to each other.

I think the backup BIOS is the reason it took the board a fair amount of time to start up initially. If the main BIOS IC contains an error, the contents of the backup BIOS is copied to the main BIOS. May well have been the case here. AFAIK Gigabyte is the only manufacturer that builds a backup BIOS IC on its mainboards.

I personally only have positive experience with reading/writing flashroms using a clip and programmer (TL866, CH341). It is a bit fiddly to reliably connect the clip to the flashrom pins, but (up to now) I have had no contact problems. Having said that, I may just have been lucky with my clip (good build quality).

Some manufacturers like MSI build an (of course undocumented) ISP connector on their boards for reading/writing BIOS on its flashrom. Makes sense in a production environment or repair facility. Especially there you want to use reliable tools and methods.

Well, I sure couldnt have done it without you, I would still be chasing my tail on this one if you didnt point out the simple things.

Yes it makes so much sense to have a easy to access way to reflash the bios chip, i hate de soldering them all the time.

Anyways im now gonna start on another motherboard MSI B85-43 GAMING that needs some diagnosing im hoping will also be a simple fix, this generation of chipsets seem to be quite reliable so im hoping its nothing major that needs replacing (BGA) it was being overclocked when it failed so hopefully its just something simple like a mosfet and driver.

[re-atari]

On MSI mainboards BIOS updates via 'Live Update' tend to fail quite often, resulting in a corrupt BIOS and ergo bricked mainboard. The ISP connector is then the only reliable route to flash a correct BIOS back. You don't want to flash write 8MB or even 32MB using a clip. On mainboards with Intel chipsets the downloaded BIOS file has to be edited with UEFI-Tool before flashing.

This doesn't take away the fact that I'm quite a fan of MSI! It's just that, if people decide they need a BIOS update, it should be done using a USB stick, and not via 'Live Update'.

[taxxin]

Originally Posted by re-atari

On MSI mainboards BIOS updates via 'Live Update' tend to fail quite often, resulting in a corrupt BIOS and ergo bricked mainboard. The ISP connector is then the only reliable route to flash a correct BIOS back. You don't want to flash write 8MB or even 32MB using a clip. On mainboards with Intel chipsets the downloaded BIOS file has to be edited with UEFI-Tool before flashing.

This doesn't take away the fact that I'm quite a fan of MSI! It's just that, if people decide they need a BIOS update, it should be done using a USB stick, and not via 'Live Update'.

HAHAHA, yes well ive quite taken a liking to GIGABYTE products as you can probably tell, being the cheapest of the lot haha

[taxxin]

Oh by the way, is there anyway to add [SOLVED] to the title? I cant seem to edit the first post anymore.

[dicky96]

Actually that [SOLVED] would be a nice forum feature. All it would take is s permissions change so that the OP can edit the title

[dicky96]

I see you sorted it. How did you get the [SOLVED] on your thread?

[taxxin]

Originally Posted by dicky96

I see you sorted it. How did you get the [SOLVED] on your thread?

got my pakistani mates over in bangladesh to help me scrape a list of possible administrator emails for this site and run a low level dating phishing attack on them until they found someone who uses a similar password for all their accounts then bruteforce through all the common combinations until i could get into the phpmyadmin access so that i could grant myself post editing powers

[RJARRRPCGP]

Originally Posted by taxxin

got my pakistani mates over in bangladesh to help me scrape a list of possible administrator emails for this site and run a low level dating phishing attack on them until they found someone who uses a similar password for all their accounts then bruteforce through all the common combinations until i could get into the phpmyadmin access so that i could grant myself post editing powers

wtf!

[dicky96]

You have a very dry sense of humour my friend lol.

Yeah you can change any user permissions you like if you can access the MySQL database via phpmyadmin, or even just edit the thread title itself, directly in the database, but it would be so much be easier to ask a moderator with access to the Vbulletin Control panel to do it for you?

And I know Vbulletin, php and mysql very very well

So go on, you made me : A lot this morning, but really how did you do it, or who edited the title for you?

[Per Hansson]

Moderator Log
Date			User	Action
14:57, 9th Jun 2021	piernov	Thread title (original 'GIGABYTE GA-H81M-DS2 REV: 3.0 Constant Power Cycling') changed

[taxxin]

Originally Posted by dicky96

You have a very dry sense of humour my friend lol.

Yeah you can change any user permissions you like if you can access the MySQL database via phpmyadmin, or even just edit the thread title itself, directly in the database, but it would be so much be easier to ask a moderator with access to the Vbulletin Control panel to do it for you?

And I know Vbulletin, php and mysql very very well

So go on, you made me : A lot this morning, but really how did you do it, or who edited the title for you?

Hhahahaahah nice, I have a wordpress site and i went through the phpmyadmin and Mysql to try and see if i could understand how the database works and man i got blown away. far more complex than i thought it would be, things really have changed.