This week on my bench laid a pretty interesting motherboard – Gigabyte X48-DS4, which finally gave a challenge in terms of my repairman skills. Without further ado, let’s get right into it.
Do You see what’s wrong with it?
These two things – broken plastic and bent pins on the IDE connector as well as missing pin on the NB heatsink, which in reality should be a screwed in bracket. Both of these can be easily fixable, well, maybe not the bracket part, but correct length plastic pins still will hold it together well enough.
That’s all I saw for now.
BUT IT GOT WORSE.
So far so good, socket is fine.
What’s more, it even works.
There’s nothing wrong with it, right?
RIGHT?
Wait, what?
WHAT IS THAT?
AND THERE’S MORE OF IT?
After my initial shock wore down I began investigating.
These ICs were next to the MOSFETs, so likely they were DrMOSes.
Surely enough they were exactly that, an ISL6609 VRM controllers.
Why were they removed though?
I didn’t need to look for answer to that for too long.
Look at the DQ9 – it’s burned. There is also some damage to the package of top left transistor, but it measured fine, so I left it alone.
When it comes to failures like this one I usually had to swap every IC on that line, so now I don’t even measure them separately – if one MOSFET is dead, whole U/L phase gets replacement transistors.
So, out with the old…
…and in with the new.
As You can see I also repaired the controllers, as I found a dead 990FX-UD3 motherboard, which thankfully used same VRM design – I stole the MOSFETs, controllers and supporting capacitors and used them on this motherboard.
While replacing one of the DrMOSes I had to repair one crucial pad (PWM – signal from the VRM controller), which proved to be nothing for me.
“Blah, blah, blah, stupid repair guy talks too much, I want to see it fully working”
I also do. Let’s see If I did a good job.
It does turn on and it even POSTs!
There is also another differece – previously VRM lights were always lit up, but now they power down after POST!
Also, just to be sure I measured the signals on all DrMOSes on my scope, they were all the same. Sorry for no pictures, but holding the probe steady and without shorting anything while also trying to focus my phone on the CRT proved to be too difficult for me.
Sorry.
We’re not out the woods yet though.
Seems like this motherboard has had quite a history. Bad VRM and a failed delid attempt, which thankfully didn’t kill it. Let’s delid it properly.
Delid went exceptionally well, no broken capacitors or further damage to the substrate.
Last thing to do – new, a bit softer thermopad for the MOSFET array to both improve thermal transfer, as well as mitigate height differences on newly soldered replacements. I obviously held the ICs down when I was soldering them, but there for sure are some differences in dimesions.
BONUS: While I was looking online for the NB heatsink bracket I saw that these boards had passive SB cooling. I tried looking around for a fitting heatsink, but that’s the best I could find in my cave.
It doesn’t fit, has an ASRock logo on it, but it helps with overheating.
Probably.
Sorry x2.
Lastly, I updated the BIOS.
I’ve been testing this boad for a few days on and off and didn’t find any more faults with it, so I declare it fully working.
Thanks for reading!