gkar (4P G34) is down

P

plext0r

[H]ard DCOTM x3
Joined
Dec 1, 2009
Messages
780
Darn it, gkar died over the weekend. When I found it, all the fans were running, no video, no keyboard. Power cycled at the P/S (hold power button for 10 seconds did nothing). Tries to come on, no video output.

Re-enabled IPMI to see if it would tell me anything. IPMI thinks the m/b is powered off (only offers the power on option under Power Control).

Plan to reset CMOS and/or pull the battery when I get a chance. It was running stable at 3.4GHz since I joined team 33 on April 11.

Grandpa is going to kick my butt until I get gkar back up and running. :(
 
I have had that a couple of times when the PSU started to fail. I had to disconnect the power cord from the PSU. Even though the PSU switch was off, I think some type of trickle current was still being maintained to the MB. 99.99% that would fixes things. A couple of times I did have to pull the power cord and then reset the CMOS. Thankfully at least the Supermicro MBs are pretty tough.
 
No luck reviving it after CMOS reset. Pulled plug, pulled battery, hit contacts.

Hit power button, fans spun for a split second, then turned off. Since then I cannot get them to spin again, etc. Guess I need to round up a spare P/S to try.
 
Not sure how much trouble it would be but you could try reducing the load, say down to 1 CPU and minimum memory.
That might give you an indication whether it's the PSU dying/passed on, or one of the components you remove.
Then if still no boot, swap what you reduced down to for pieces you first removed.
That should eliminate all but MB and PSU.
 
I am already kicking your backside now quit playing around and get that puppy going. :D

Hopefully it is just a Power Suplply issue and you are able to get her going again without allot of trouble. It kind of sounds like that may be the problem to me, I hade the same type of symptoms 1 time when I tried to power up a 4P with a 650 Watt PS, patients is not one of my strong points my PS had not arrived yet so I was trying the 650 needless to say it did not work but had the same symptoms you are describeing.
 
Grandpa_01 said:
I am already kicking your backside now quit playing around and get that puppy going. :D

Hopefully it is just a Power Suplply issue and you are able to get her going again without allot of trouble. It kind of sounds like that may be the problem to me, I hade the same type of symptoms 1 time when I tried to power up a 4P with a 650 Watt PS, patients is not one of my strong points my PS had not arrived yet so I was trying the 650 needless to say it did not work but had the same symptoms you are describeing.
Click to expand...

While gkar is down, I've increased my PPD with some new systems. They should start checking in tomorrow. Eight dual E5-2680 systems. ;)

I need to grab my spare P/S and try it before I start removing CPUs with musky-fied coolers. I don't think I can remove them while the M/B is inside the Blackhawk case, so I'd need to disassemble the thing. I'm hoping it's the P/S, although, if it is, this would be my first Seasonic to bite the dust.
 
brilong said:
While gkar is down, I've increased my PPD with some new systems. They should start checking in tomorrow. Eight dual E5-2680 systems. ;)
Click to expand...

Think, this statement will trigger Grandpa to wake up all sleepers in the house ..... :D
Eight dualies should be good for additional 2m+ ppds , offsetting gkar easily.

You guys are truely in a different league, :)
Andy
 
I tried a new 750W power supply, no dice. I know the system should at least POST with this underpowered PS since that's how I figured out the Corsair PS was broken a few months ago. I tried just the 24-pin connector and one 8-pin at a time as well as the standard all three 8-pins.

Time to take it apart, inspect the VRMs under the copper heatsinks, cycle through the CPUs one-by-one, etc. Uggh.
 
I pulled CPUs off the motherboard in reverse order. Tried booting without CPU4, without 3 and 4, with just CPU 1. No joy. Decided to pull the motherboard out of the case and remove the copper heatsinks. Some of the thermal tape is still stuck to the VRMs. How do I remove that without damaging the VRMs?

When I removed the heatsinks from CPU4 VRMs, I noticed the VRMs were wet and both the heatsinks and VRMs stink. I'm not sure if the thermal tape melted? Take a look at the following images and see what you think. Thanks!

20130530175958.jpg

CPU4 VRMs:
20130530180218.jpg

CPU4:
20130530180226.jpg


CPU3 VRMs:
20130530180353.jpg


Remember this board is damaged underneath CPU4's VRMs:


Any thoughts on next steps?
 
Wow. Is that PCB supposed to be dark brown? :D

At work we use a ball of two-sided tape to take off tape. WD-40 is second choice.
 
Crispy critters are best served with BBQ sauce....usually.

Got a warranty on this board?
 
Qinsp said:
Wow. Is that PCB supposed to be dark brown? :D

At work we use a ball of two-sided tape to take off tape. WD-40 is second choice.
Click to expand...

I assume you're telling me the board is toast due to its funny color. Time to RMA after cleaning it up and putting back the factory aluminum heatsinks?
 
While there is no way to check circuits in a picture, on electronic PCB's of all kinds that have failed, the dark color is often present from excessive heat.

Dead minimum, cooling is not adequate in that area.
 
Qinsp said:
While there is no way to check circuits in a picture, on electronic PCB's of all kinds that have failed, the dark color is often present from excessive heat.

Dead minimum, cooling is not adequate in that area.
Click to expand...

I'm not an EE, but I wonder if the damaged underside (broke off a cap when I first installed the system) caused some of the VRMs for CPU4 to be unused and this overtaxed / overheated the other VRMs. CPU4 is the only one with brown color around the VRM area.

I guess it's time to call SuperMicro.
 
Qinsp said:
At work we use a ball of two-sided tape to take off tape. WD-40 is second choice.
Click to expand...

I got all the tape off the VRMs, but there's still residue left behind. I used my fingernail to get rid of a lot of gunk. Should I use WD-40 or alcohol or something else to get rid of the residue?

Thanks.
 
Alcohol leave little residue and will not eat plastic. DO NOT use acetone. WD-40 is good at removing sticky stuff, but leaves an oily residue and sweet smell. (I like the smell of diesel fuel and gasoline, your results might vary).
 
Def alchol should clean it up. Walmart actually sells alchol wipes in their eye glass section for extremely cheap and they work really well. A lot less exspensive then buying "professional" computer wipes...
 
m33pm33p said:
Def alchol should clean it up. Walmart actually sells alchol wipes in their eye glass section for extremely cheap and they work really well. A lot less exspensive then buying "professional" computer wipes...
Click to expand...

I was thinking of dipping Q-tips in isopropyl alcohol and using them to wipe off the gummy residue. I don't like sticking my fingers in between the components since I'm afraid something will pop off (like those big caps). If that doesn't work, alcohol wipes sound like a good plan.
 
brilong said:
I was thinking of dipping Q-tips in isopropyl alcohol and using them to wipe off the gummy residue. I don't like sticking my fingers in between the components since I'm afraid something will pop off (like those big caps). If that doesn't work, alcohol wipes sound like a good plan.
Click to expand...

Q-Tips with iso-alc has been my method of cleaning electronics for years. Its good for removing WD40 residue too if you have to use that first for glue. Just pour some into the bottle lid and dip the QTip in, pour unused back into bottle - no mess, no cleanup. Hope the RMA goes well and you get it back up and running soon. I've had an ASUS mobo sent out for RMA for over 2 weeks so I feel your pain ...
 
Called SuperMicro today. They shipped my replacement board on 6/12 and I expect it next Wed, 6/19. No charges mentioned.
 
I got everything put back together and it won't POST. SM sent me a new board, I installed all 4 CPUs. Now I get to try 1 CPU at a time. :( I hope one of the CPUs didn't get fried when CPU4 socket got toasted.
 
brilong said:
I got everything put back together and it won't POST. SM sent me a new board, I installed all 4 CPUs. Now I get to try 1 CPU at a time. :( I hope one of the CPUs didn't get fried when CPU4 socket got toasted.
Click to expand...

Uh-oh, hopefully not. I'm pretty sure I don't need to mention it - and I'm no 4P expert (last 2P I had was S370) - but have you checked every BIOS setting? Good to see you got the mobo back and hopefully folding [H]ard soon. I got my ASUS back from RMA and all went without a hitch. Hopefully you get it ironed out soon.
 
Toconator said:
have you checked every BIOS setting?
Click to expand...

Since the board won't POST and outputs no video, I cannot get into the BIOS. I'll try to get a single CPU operational and see if I can get into BIOS, make sure it supports 62xx, etc.
 
Does anyone run BIOS 3.201 with 62xx extra spicy chips? I noticed this new motherboard appears to have a new BIOS. I'm fairly certain my original GL had 3.1.
 
g'kar boots fine with 3 CPUs (1,2,4). No matter which CPU I put in socket 3, it will not boot. This is a "new" board I got back from the RMA and it was supposedly tested with 4 CPUs. I'm escalating with SuperMicro now. :(
 
Any update on this? Did SM put the screws to you or where they pretty cool about it?
 
m33pm33p said:
Any update on this? Did SM put the screws to you or where they pretty cool about it?
Click to expand...

I opened another RMA at the end of my day on Friday. They responded that they wanted tech support to help me diagnose the problem before issuing another RMA. I just re-inspected socket 3 and it appears I have a single bent pin. Now that I think about it, I think it's my fault, so I've asked the RMA dept. for a quote to repair/replace socket 3. g'kar will be down for another one or two weeks. :mad:

I think the problem is that I drank two cups of coffee before assembling the system. I noticed my hands were a little jittery during the assembly process and here's the result.
 
brilong said:
I opened another RMA at the end of my day on Friday. They responded that they wanted tech support to help me diagnose the problem before issuing another RMA. I just re-inspected socket 3 and it appears I have a single bent pin. Now that I think about it, I think it's my fault, so I've asked the RMA dept. for a quote to repair/replace socket 3. g'kar will be down for another one or two weeks. :mad:

I think the problem is that I drank two cups of coffee before assembling the system. I noticed my hands were a little jittery during the assembly process and here's the result.
Click to expand...


N0o0o0o0o0o:eek:
 
Aw man that sucks. You can't get a tiny screw driver or razor blade between the pins to straighten them back out?
 
I am sending one their way as well...again:eek:

Ro remove the schmutz on the vrms, use goo-gone and an old towel.

Anyone wanna buy a board in 4 weeks? Time to upgrade.:cool:
 
brilong said:
Now that I think about it, I think it's my fault, so I've asked the RMA dept. for a quote to repair/replace socket 3. g'kar will be down for another one or two weeks. :mad:

I think the problem is that I drank two cups of coffee before assembling the system.
Click to expand...

Too bad about the pin but at least you're being honest about it. Only two cups of coffee? Switch to espresso shots ( I drink 3 or 4 at a time). Much more effective, "nature's defibrillator" :D
 
On a good note, I got the Intel 4P up and running yesterday. 650K PPD on 8101 so far.
 
You must log in or register to reply here.
Back
Top