Help wanted: fixing a Rampage Extreme

[Gigelz]

2 minutes ago, Yoshiaski said:

Yes my friend, I had seen it, I thought it was the best option, it is just for sure. By the way. So it is not a problem to exceed 1000uf? For example put 4 of 330uf? Or 3 of 470uf?

in theory you can bust weak vrm´s or trap ocp by adding too much capacity. i think in case of nb or memory vrm until 5000uF should not be a problem at all. and cpu vrm is even stronger. but you need to take care of the cap quality itself. you need very low esr and the right voltage rating for best results.

[Yoshiaski]

13 minutes ago, Gigelz said:

in theory you can bust weak vrm´s or trap ocp by adding too much capacity. i think in case of nb or memory vrm until 5000uF should not be a problem at all. and cpu vrm is even stronger. but you need to take care of the cap quality itself. you need very low esr and the right voltage rating for best results.

Thank you very much really. Well, I will be asking for the pieces to be welded, for many years now I have not welded, hehe, I don't know why I stopped studying electronics. Let's see how it goes, greetings.

[ground]

1 hour ago, TerraRaptor said:

Outside terminals (long strips) are negative, inside terminals (dots) are positive.

Are you sure? Would recommend probing to confirm. I currently don't have a rex with missing fujitsu on hand to probe, but in boardviews for R2E and R3E the contacts are the other way around...

[Yoshiaski]

Physically the dots are outside and inside the long strips. But. The points are positive?

[ground]

Got a multimeter on hand? in that case just probe gnd against either kind of pad and whichever has the low resistance is gnd.

[Yoshiaski]

2 minutes ago, ground1556 said:

Got a multimeter on hand? in that case just probe gnd against either kind of pad and whichever has the low resistance is gnd.

No, unfortunately I do not have.

[TerraRaptor]

1 hour ago, Yoshiaski said:

Yes my friend, I had seen it, I thought it was the best option, it is just for sure. By the way. So it is not a problem to exceed 1000uf? For example put 4 of 330uf? Or 3 of 470uf?

Make sure to re-read my post - I edited it 'cause I was wrong with cap polarity. No problem to exceed that much at all. However very unlikely you will be able to fit more than 3 caps there.

[Gigelz]

Also doing such electronic stuff without a multimeter is pretty dumb

[Noxinite]

Here is some info from my latest session to try and fix my boards.

Resistances of main voltage rails measured:

At these points:

I then removed all of the Fujitsu capacitors as most of them appeared to be bad, giving me these resistances:

Unfortunately boards 1, 2 & 4 were still having issues, so I checked the voltages as well:

I then checked the post codes and board 1 has d5 and board 2 has d4 - this is with various memory tried in each ram slot.

For board 4 if I remove the EPS 8-pin the board will turn on, so clearly there is an issue there but the general resistances seem okay.

Also, I tried a working BIOS chip from board 3 in each of the other boards, however I still got the same behaviour.

All in all no progress and still 3 dead boards, but maybe someone finds something useful in this info.

[GRIFF]

Hi Noxinite.

Do this test: connect power supply, just 24 pins, turn on power supply and do not turn on motherboard.  leave 10 minutes and touch SB, lan chip, audio chip, multicontroller chip.  They must all be of the room temperature.  if you have infrared thermostat it would be perfect, infrared camera even better.  Pay particular attention to the SB.  If it becomes at least 3-5 degrees warmer than the rest, the SB surely to be changed.

Edited April 5, 2020 by GRIFF

[GRIFF]

Another test to do: compared to the previous one, the card must be turned on and the temperature of the same components of the previous test is checked.  SB temperature I detect with an infrared thermometer pointing it to the center before turning on the board and turning on watching how the temperature changes in 1-2 minutes.  Temperatures above 45-50 degrees indicate the failure of the SB.

[GRIFF]

Another test to do: measure resistance data lines of all USB ports.  all must be almost equal and have at least 1 Mohm.  normally the value is 2-4 Mohm

[GRIFF]

Why do I speak to you about the SB above all, because two REXs that have arrived to me apparently good with all the tensions OK, restarted by changing SB.

[GtiJason]

On 4/5/2020 at 7:46 AM, Noxinite said:

For board 4 if I remove the EPS 8-pin the board will turn on, so clearly there is an issue there but the general resistances seem okay.

Also, I tried a working BIOS chip from board 3 in each of the other boards, however I still got the same behaviour.

Exact problem with my 2nd board

[GRIFF]

Hello guys. Since I have 3 cards still to be repaired and you also have dead REX collections. And also considering that I managed to fix my infrared soldering station, I picked up one of my dead REXs. Result)))))))))

[GRIFF]

Let's go in chronological order:
Before the failure - while I was trying for the max freq I raised the voltage to SB by setting 1.2V. - F10 - Enter - reboot - black screen - bios reset - black screen - attached post card marks NO.
Later - Considering all this I thought that SB died. Today changed SB chip (If you look closely at the photo of the chip you will find that it is different from the original. Since I didn't have ICH9R I soldered ICH10, the one they mount on P45 boards.) and card started to hang on post 2A (better than NO). Checking the heating of the LAN and Audio chips I realized that one LAN chip gets hot and the other remains cold. Then unsoldered chip that warmed and board kept crashing on POST 2A. Unsoldered other LAN chip and Yessssssssss !!!!!!!!!!! It works!!!!!!!!!!!!!!!!!!

Edited April 9, 2020 by GRIFF

[GRIFF]

Considering all this, first try to check these LAN chips, how they heat up and if there is a difference between them. Good chip becomes lukewarm. Bad or cold or too hot chip. I hope this info helps you.

[Gigelz]

Thanks @GRIFF for taking part of this thread very interesting infos about other, than the usual, defects ! You have a workint IR soldering station ?

[GRIFF]

Good morning guys.
This morning I picked up another REX. Symptoms: endless cyclic restarts. It comes on - stays on 3-5 sec - goes off and stays off 2-3 sec - comes on again. First of all I unsoldered two LAN chips, nothing has changed. Replaced bios chip for safety, although I have already done so previously. Nothing. Then decided to change SB. And as you can see, this too has started up again. In addition, this card had resistances on the USB and SATA data lines in the norm.

[Noxinite]

Anyone figured out why the SB keep dying? Anything we can do to stop it like more filtering caps?

[GRIFF]

Let's say that the solution has not found anyone for now.  I read a lot of forums in Russian and there are good people who repair by trade and no one has been able to say why.  However, everyone claims that good cooling improves the situation but does not solve it.  They say the chips are the problem and it changes little if they are ICH7 or 8 or 9 or 10.

[d0minat0r]

Good job GRIFF

[TAGG]

@GRIFF Where do you source the SB from? Aliexpress? or do you salvage them and reball them?

I think i have a board with broken SB here, it used to work when in shorted 2 pads next to SB with my fingers (found out on accident), now it's completely dead...

[GRIFF]

1 hour ago, TAGG said:

Where do you source the SB from? Aliexpress? or do you salvage them and reball them?

Aliexpress. For the moment not used reballed chips. In my opinion it does not make sense to reball for chips that cost 5-10 euros.

Edited April 10, 2020 by GRIFF

[Noxinite]

I tried having a look at my boards again. Checked temps of all of the chips at stand-by and quickly on boot-up, but nothing looks out of the ordinary. Only thing I found was that REX-2 was missing DQ57 in the outer two ram slots and in fact posted with ram in the inner ram slots. Somehow the trace between the outer dimm slots and the NB broke??

Edited July 3, 2021 by Noxinite