Help wanted: fixing a Rampage Extreme

[TaPaKaH]

Over the last few years, the board has been on LN2 about a dozen times and every 2-3 sessions it would arrive in a state where it would shut down and no longer react to the power button. The only way to get it back to work was putting it in a 120c oven for 10 minutes or so. However, after its most recent breakdown + ovening it now exhibits the following behavior:

- With Seasonic Titanium 1000W: no reacton to power button at all

- With Corsair AX1500i: it powers up, but shuts down before it completes POST

- With Silverstone OP1000: it powers up and works perfectly fine. But, if I set the Vmch to any value other than [Auto] in the BIOS, the board shuts down during POST with a OCP/OVP-like click coming from somewhere in the chipset PWM area. If I boot into OS at [Auto] value and raise the Vmch using the TweakIT thingy - system works fine, even handles LinX load with a quadcore. If I adjust the BIOS settings so that the [Auto] value no longer induces the default ~1.25V, I get the same shutdown on POST.
Also, the same happens when I set the "EPU Phase Control" option in the BIOS from [Auto] to [Extreme]. But this might be due to [Auto] option for Vmch changing.

I tried ovening it some more. I tried flashing other BIOS version. I tried a BIOS chip from another board. I tried different kinds of CPUs. Nothing seems to help and I'm out of ideas. Any advice / help is welcome.

EDIT: pictures of the chipset PWM area:
https://abload.de/img/img_0525ezivo.jpg
https://abload.de/img/img_0527akc2u.jpg
https://abload.de/img/img_0528q6fho.jpg

Edited December 21, 2018 by TaPaKaH

[ground]

Are the Capacitors raised? Might be that those are messed from the repeated oven sessions... If so recapping the board (at least the relevant ones) could be a potential solution. I would wait for more ideas what could be wrong though.

[TaPaKaH]

The only capacitor the board has in chipset PWM is the infamous Fujitsu that tend to self-destruct over time. I don't have any soldering equipment capable of tackling one so if anyone would be willing / able to give me a hand, I'd be more than happy to ship the board over.

[Noxinite]

You can dremel it off and replace it was other caps if you have a really steady hand, haha. The board I did that on did randomly die some time later though.

Edited December 21, 2018 by Noxinite

[ground]

I was mainly thinking about the input filtering Caps for vcore and other rails and possibly even 12V. Check the caps before the NB VRM, those could be an issue too I guess?

Had a Rampage IV Formula that had been baked and pretty much all the caps were toast due to that. It could very well be something else, but that's all I can think off right now.

Edited December 21, 2018 by ground1556

[Noxinite]

I has pics.

[TaPaKaH]

I see that in your case the Fujitsu cap looks damaged. In my case it looks perfectly fine. Is there any way to "test" it so that the soldering work is not done for nothing?

[Fasttrack]

Sam,

Of course I have no idea or solution to provide.

The reason I am posting is that I have an Asus Apex VI ( X299 ) - retired for 3 months now - that behaves exactly the same.

Lights turn on, other times it boots, other times it boots and shuts down in 5-10 minutes, and other times it does not even start.

Tested with 10 different PSU's. Took it to a local electronics lab ( best there is locally ). The guy who owns the lab went nuts for 3 days. Result - NO CLUE.

So, I decided ( as Michael did ) to hang it to my wife's Christmas cactus, to remind me that we pay an arm and a leg for HW, that may decide on their own WHEN AND HOW

to function.

Merry Christmas to all

 

 

[Noxinite]

16 hours ago, TaPaKaH said:

I see that in your case the Fujitsu cap looks damaged. In my case it looks perfectly fine. Is there any way to "test" it so that the soldering work is not done for nothing?

Yeah, I popped it somehow - maybe moisture??

The orange area is ground and the raised areas have the same resistance as the outputs of the mem chokes (in my case).

[suzuki]

I have 3 “dead” REX ,i am keeping them untill technology will evolve enough to be fixed :)).

[subaruwrc]

don't want to play clever or something but the circled area looks oxidized to me.

[xpower]

seems thats the problem and is difficult to fix with normal soldring if traces or ic are blown

[TAGG]

On 12/22/2018 at 12:50 PM, subaruwrc said:

don't want to play clever or something but the circled area looks oxidized to me.

That was the problem indeed, reflowed the corroded looking VRM controller and the board appears to work fine, just have to put back the Fusitsu cap, after checking if it is good

[subaruwrc]

1 hour ago, TAGG said:

That was the problem indeed, reflowed the corroded looking VRM controller and the board appears to work fine, just have to put back the Fusitsu cap, after checking if it is good

I could help once to names like you 2  

achievement unlocked!

nice resurrection !

[TaPaKaH]

Does it also work with many PSUs and allow to set any Vmch value in the BIOS without crashing now?

[TAGG]

1 hour ago, TaPaKaH said:

Does it also work with many PSUs and allow to set any Vmch value in the BIOS without crashing now?

Have only 2 PSUs works on both, from 1,25 to 1,45 it is no problem, will test more once i have that heatsink on

BTW, might have killed the 2nd one, had it working perfectly, but then i tried to reinstall the original Fujitsu cap, which worked after the 3rd try, but now it doesn't recognize power button, i think the heatcycles might have killed the NB... Initial problem with the 2nd one was shorted capacitor on Vnb if it interrests you

EDIT, Board one tested to 1,65V seems back in bussiness, but nb filtering looks like this now

EDIT2: The board i mentioned i killed lives again, but VRM is allso dead again

Edited March 20, 2019 by TAGG

[TaPaKaH]

Wouldn't additional wire length add noise at filtering? But I guess you don't have much choice due to height restrictions

I just bought yet another Rampage Extreme locally, and guess what greets me in memory VRM area? Mr about-to-die Fujitsu, long time no see...
https://abload.de/img/img_12120mj9k.jpg

[ObscureParadox]

I'd assume that running the board without any capacitor is better than with that long wire due to noise? 

[TAGG]

57 minutes ago, ObscureParadox said:

I'd assume that running the board without any capacitor is better than with that long wire due to noise? 

There are 6 10uF MLCCs under there to cope with that issue

Edited March 20, 2019 by TAGG

[Masterchief79]

Hey guys, my REX also randomly died after laying in the closet for a few months. It appears to have a short in the CPU VRM. With only the ATX Plug connected, it powers up, but no POST of course. With the 8 Pin CPU connected, it powers up for a few milliseconds before it immediately shuts down again and won't power up until I reset the PSU. 

I checked all mosfets and drivers, no shorts anywhere. Could it be one of the Fujitsu caps (like the one on the back of the socket maybe)? Can I test that somehow? Can you desolder them with a heatgun and not damage other critical components? Would really like to get that going again so help and experience is much appreciated.

Edit: I should maybe add: The board has had some reoccurring NB VRM issues with the voltage dropping to 0.74V on bootup. The NB coils have weak soldering points because when I first received the board, he coils had broken off during transport. I kind of bodged it together and it worked for SS, Dice and LN2 for some years. A few times when the board wouldn't boot, I resoldered the coils again and it was fine. So I don't think there's an issue there, but since the NB VRM also has a Fujitsu cap...

Edited April 16, 2019 by Masterchief79

[d0minat0r]

HI guys, I have two REX and have same problem on both sample. MB pass 32M 630FSB and after restart stuck when showing ROG logo. DO you have an idea what is a problem?

[suzuki]

All the leds are working but no image on screen ?

[d0minat0r]

It gives the half logo and freeze here

[TerraRaptor]

That is most probably unsuccessful memory training during reboot. I also get hangs after reboot when skews are a bit off although cold boot works fine. If you disable logo, you should most likely be stuck on usb initialisation.

Did you try different memory for those 630mhz runs to see what happens after reboot?

[Noxinite]

I figured I should add some testing I was doing today.

I have 4 REXs, but 2 have been dead for a while.

>>Board "A":

>Powers on and hangs forever.

>PCI post card gives D5 and then 0E errors.

>D5 is normal start of postcode cycle (so I assume no hardware damage).

>0E is no microcode detected.

>Tried switching BIOS chips in from other boards still gives D5, 0E.

>No BIOS chip installed = instant 00 and hang.

>"BIOS_flashback" jumper on bottom of board in the far left position gives D0 post code and then shut down. I had this error with one of the BIOS chips before reseating it and now with the jumper in the second and third positions it reaches D5, 0E.

>Out of the four BIOS chips I tested one gave 00 and hang (so maybe that BIOS chip is dead?).

 

>>Board "B":

>Powers on and shuts down immediately.

>One missing pin in the bottom left of the socket.

>POST codes are 00, 98 then shutdown.

>Either not detecting CPU in socket (due to the socket damage), or some issue with the SB maybe? As 98 is "Console input devices connect".

Not sure what to test next.