Yet another EVGA 680i won’t POST

[ladderman]

My EVGA 680iA1 won’t post and there’s no display through the monitor. All I get is “--”. I have three tiny LEDs in the top right corner of the mobo. One blue, one green, and one amber.

I would be grateful if someone could check the LEDs on their EVGA 680iA1 to see if it should show two green instead of one green and one amber.

The system had been running ok for the past six months. But I’ve just rebuilt after I stripped down my liquid cooled system to clean all the parts and change the tubing.

I left the mobo in the case, so the only damage could have been caused by over tightening the NB block.

I’ve tried doing the following twice, but to no avail.

1. Shut down system and turn off PSU.
2. Press power button to discharge mobo
3. Jumper Clear CMOS jumper
4. Remove battery
5. Remove all memory
6. Let sit for 15-20min
7. Place 1 stick of mem in system in slot farthest from the CPU.
8. Replace Clear CMOS jumper.
9. Repace battery
10. Turn on PSU

Anyone got any ideas, or is it a RMA?

[Talonman]

Re-seat the CPU?

[Isriam]

if you follow the procedure to clear -- and it doesnt work ,then rma. i've had 2 that did this, and neither could i bring back from the dead.

[NaeKuh]

80% of the time, this means you have a grounding problem. Backplate is being pushed up against the top mosfet that its causing a short. You need to pull your block off, and plate off, and redue it so it wont push up against the metal contacts.

If worse come to worsse, you need to take the board out, pull the cmos battery, short cmos for about 15 min, and retest hardware. That usually fixed it 90% of the time for me when i had this board.

If you run into the 10%, a RMA is very likely in your future.

[smee]

Try pressing hard on the edges of the bios chip until you here a click or snap.
The board won't post unless the bios chip is in properly.

[rogard]

Go back to air to test that the board still works.

[Martinm210]

Mine is one amber, one green, one blue and running fine.

A while back I though my 680i was dead when no codes would post, turned out my PSU was dead. But if you're getting all three leds and the PSU is on, it's probably something else.

Might pull the CPU out and clean the contacts on the bottom and make sure no CPU socket pins are bent.

Try it again without anything in the board, and slowly add the ram, then put in the CPU, etc...

[ladderman]

Mine is one amber, one green, one blue and running fine.

A while back I though my 680i was dead when no codes would post, turned out my PSU was dead. But if you're getting all three leds and the PSU is on, it's probably something else.

Might pull the CPU out and clean the contacts on the bottom and make sure no CPU socket pins are bent.

Try it again without anything in the board, and slowly add the ram, then put in the CPU, etc...

Thanks Martinm210

I’ll strip everything down tomorrow, check and reseat the CPU, clear the CMOS again. Then I’ll try the board on air.

[lennox]

it's dead... happened to me twice. Moved to ASUS.

[Isriam]

why would you go to asus? evga has the best replacement policy i've ever come across. lifetime warranty on a1? advanced RMA for 10 dollars?

[rehpyc]

And a nice step-up program on their products :p

[Jedda]

80% of the time, this means you have a grounding problem. Backplate is being pushed up against the top mosfet that its causing a short. You need to pull your block off, and plate off, and redue it so it wont push up against the metal contacts.

If worse come to worsse, you need to take the board out, pull the cmos battery, short cmos for about 15 min, and retest hardware. That usually fixed it 90% of the time for me when i had this board.

If you run into the 10%, a RMA is very likely in your future.

QFA!

What I find highly unusual is the circuit configuration around the CPU. If you take an OHM meter prior to the mounting bracket for the CPU you will find the LGA-775 CPU socket and bracket are in no way grounded. For all intents and purpose those metal parts at first appear to be isolated electrically from the rest of the board.

That is, until you insert the CPU. Now, the bracket, CPU case etc are now making a complete ground to chassis ground. I should rephrase that, most boards seem to make a complete ground, not all. On two boards I have worked on that were acting flakey I found that with a DC voltmeter I had DC potential between the CPU case and chassis ground when the machine turned on. In one case reseating the CPU fixed the issue, in the second case the problem didn’t go away and ultimately customer returned the board for RMA. His new board does not exhibit any DC potential between his CPU case and chassis ground.

It’s just an observation of course, but why would the designers force the socket to ground through the CPU? It would seem to me that a very slightly misaligned CPU or even a piece of grunge on a contact pad, or worse a contact spring being bumped slightly off center could be the cause of some interesting issues.

Mounting a water block obviously is in the same category.

Thanks to BillR.

[civictyper16]

Omg dude we have the same problem I had to take out my whole system apart that took weeks to get assembled I think I might have overtightened the nb I put it back on air and tried everything we should just rma the dam thing and forget it.

[Jedda]

Its not necessarily the same problem, civictyper.
All the -.- read out is indicating is an earthing error. That is the electricity isn't going to the right place the right way.
Pretty general indication.
Some are fixed by a cpu remount others are a psu issue or short to mobo backplate. QA is about the same as all the other mobo made in China.
RMA is your last, not first, option.
RMA is not a diagnostic procedure.

[ladderman]

QFA!

What I find highly unusual is the circuit configuration around the CPU. If you take an OHM meter prior to the mounting bracket for the CPU you will find the LGA-775 CPU socket and bracket are in no way grounded. For all intents and purpose those metal parts at first appear to be isolated electrically from the rest of the board.

That is, until you insert the CPU. Now, the bracket, CPU case etc are now making a complete ground to chassis ground. I should rephrase that, most boards seem to make a complete ground, not all. On two boards I have worked on that were acting flakey I found that with a DC voltmeter I had DC potential between the CPU case and chassis ground when the machine turned on. In one case reseating the CPU fixed the issue, in the second case the problem didn’t go away and ultimately customer returned the board for RMA. His new board does not exhibit any DC potential between his CPU case and chassis ground.

It’s just an observation of course, but why would the designers force the socket to ground through the CPU? It would seem to me that a very slightly misaligned CPU or even a piece of grunge on a contact pad, or worse a contact spring being bumped slightly off center could be the cause of some interesting issues.

Mounting a water block obviously is in the same category.

Thanks to BillR.

Great post Jedda.

I’m stripping everything down today and will pay extra attention to the CPU mounting area when I remount the CPU.

I’ll post the results, lets hope its good

[Talonman]

Good luck buddy! ;)

[Malik]

By me this code " - - " was ... dead cpu :(

[Eddie3dfx]

There is no other board in the history of overpriced crap that has been more defective than the evga/foxconn 680i.
It is the pontiac fiero of mothboards.
Pay the steup fee to the 780i and see if the problems are still there.

[mouawad]

why would you go to asus? evga has the best replacement policy i've ever come across. lifetime warranty on a1? advanced RMA for 10 dollars?

so you can use a board which most likely will never need 3 RMA's per year like the 680i.

evga might have great support but the 680 is a joke.

[ladderman]

I’m sitting here rather depressed, its not the boards fault, it’s mine.:slapass:

As you can see by the photo below I have bent two pins on the CPU socket.

I can’t understand how I’ve done this because I always take great care when I seat a CPU into the CPU socket.:confused:

One thing I have noticed is that the bent pins are directly under were one of the knobs on the plastic socket cover is. I’m wondering if this might of coursed the damage when I used it to cover the CPU while I had my system stripped down?

I managed to bend one pin back, but the other snapped off.

Do you think the CPU will be damaged as well?

Is it worth trying to RMA it or do you think they will tell it’s my fault?

I had added the SF800 and T-Balancer + Sensor Hub. I had also placed loads of temp sensors around the system. So all that hard work has been a waste on time.

Will I ever get this rig back together:(

[NaeKuh]

why would you go to asus? evga has the best replacement policy i've ever come across. lifetime warranty on a1? advanced RMA for 10 dollars?

cuz unless he's talking about a striker, which sucks hardcore. i swapped out from a 680i to a P5K-DLX.

The P35 boards will own any 680i board in stability, and endurance. :up:


680i *sigh* my expensive eyesore i have sitting in a box waiting for stepup.

[Jedda]

Looks like the plastic lug did the damage, bugga!
You think you're being extra careful but it bites you.

'supose you could try sending it in, or the 780 step up option?
My immediate reaction when I screw up is to get a replacement and chalk it up to private education fees at the school of hard knocks, but there's not many evga 680i A1 boards around. Finding a replacement could be difficult atm.

[pika198]

well i bought my 680i se from newegg for a 134.99 so i figured it was a good deal for me :D I shorted mine out too, but I had my advanced placement overnight shipped for 30 bucks and got everything all setup...now if only bfg did that because I'm still waiting for them to ship out my rma card :P

[sanhacker]

@ladderman,
Kudos to you for stepping up and pointing out that you bent the pins on the CPU socket.

I'm not sure how that cover would bend the pins but it was straight up of you to post this out as opposed to blaming the board.

It sounds to me like there is an inherent 775 socket problem as I'm seeing the same thing on eVGA, ASUS and XFX forums.

[ladderman]

@ladderman,
Kudos to you for stepping up and pointing out that you bent the pins on the CPU socket.

I'm not sure how that cover would bend the pins but it was straight up of you to post this out as opposed to blaming the board.

It sounds to me like there is an inherent 775 socket problem as I'm seeing the same thing on eVGA, ASUS and XFX forums.

Thanks mate:up:

Very interesting what you mention about this happening on other boards. It’s defiantly a very vulnerable area on a 775 socket board

I always take great care when seating or removing a CPU, so after this expensive lesson I’ll be taking even more care.

I spoke to the Scan Computers, the company who supplied the board, and explained that I thought the CPU socket cover may of done the damage. They informed me that I could not RMA because when it comes to bent pins “the blame lies at my door“.

Lucky they had some EVGA A1 boards in stock £134 so I’ve got one coming on Monday, and this one will get extra special treatment.

At least I have highlighted an area that may help others pay a little more attention so they don’t make the same expensive mistake as I did.

[taowulf]

I spoke to the Scan Computers, the company who supplied the board, and explained that I thought the CPU socket cover may of done the damage. They informed me that I could not RMA because when it comes to bent pins “the blame lies at my door“.



I can see why they say this, but when I put together my former roomie's EVGA 680i TR together, when we took the protective cover off, there was a bent pin that definitely was not my fault. Knowing how retailers will not RMA for this, I was fortunately able to bend it back (tweezers rock) but had I broken it he would have killed me.

They need to realize that it not outside the realm of possibility that it might have shipped that way (not referring to your experience) and allow for that possibility.

Of course, hammer thumbed people will abuse this.

No deliverance for the innocent in this case.

[ladderman]

I can see why they say this, but when I put together my former roomie's EVGA 680i TR together, when we took the protective cover off, there was a bent pin that definitely was not my fault. Knowing how retailers will not RMA for this, I was fortunately able to bend it back (tweezers rock) but had I broken it he would have killed me.

They need to realize that it not outside the realm of possibility that it might have shipped that way (not referring to your experience) and allow for that possibility.

Of course, hammer thumbed people will abuse this.

No deliverance for the innocent in this case.

Funny, but I was starting think that to day, if the new board has any bent pins how do I prove it came like that!

Maybe the manufactures need to put some form of sealed clear cover that you could view the socket before you use it Then that would prove it came like that.

Another downer, Scan has just sent me an e-mail saying I wont get the board until the 11th Jan. I getting a bit pissed now:mad:

[taowulf]

Funny, but I was starting think that to day, if the new board has any bent pins how do I prove it came like that!

Maybe the manufactures need to put some form of sealed clear cover that you could view the socket before you use it Then that would prove it came like that.

Another downer, Scan has just sent me an e-mail saying I wont get the board until the 11th Jan. I getting a bit pissed now:mad:


it is just as bad as those software license agreements that state that as soon as you opened the packaging you accept the licensing agreement...even though the software license is not seen until you attempt to install the software! What is an innocent consumer to do?

Yeah, clear socket covers would rock, but people wouldn't look until they went through what you have.