Maybe a bit of topic but I want to tell about my day and how i probably fried half of my ram sticks.
the day started nice I got a new shiny RTX 4090 that I decided to add to the first PCIe lane of my Asus wrx90 motherboard. Everything in my system is water cooled so I took my time made sure there where no leaks (used an external power supply to only run the pumps etc).
After checking everything I decided it was time to boot up the computer. Once I turned it on sparkles came flying of my bottom dims and in complete panic, I fumbled trying to hit to power switch on the PSU, I could not reach or find it so after 3-5 seconds i just pulled the cord as it was the only thing I could think of.
How could this have happened I had checked every tiny little connection in my water loop !?
I turned on the flashlight on my phone and started inspecting the damage, I could not see any water so I started disassembling fearing the worst. had I fried my very expensive CPU, was the motherboard toast, did I just kill my brand new 4090 card?
After removing a few cards it became clear that there was no water anywhere, I also looked for burn marks but could not find any on the motherboard or other add-in cards. It did however smell of burnt electronics. But the smell only came from my ram sticks, I pulled them out and checked my motherboard, I could not see any burn damage or sign of shorts so I decided to put in an older graphic card and try to boot again to see if everything was fried.
the good news is that it booted, and after carefully testing each part, I actually got must of my computer up and running again, its missing 4 dimms but everything else seems to work, I just hope that I didn’t fry the anything else related to the now empty memory slots.
I am still not sure what cause the short, if the ram heat spreaders touched something on the motherboard or if there was components on the memory circuit board that touched the metal heatspreaders, and now that the heatspreaders where in contactwith the metal backplate of the 4090 it could short.
I am probably going to take things apart tomorrow to see if I can see where the short happened, I really hope that it was isolated to only the ram sticks and that the problem is not related to the motherboard.
i should probably be in bed but i got too curious so i started opening the ram heat spreaders.
there’s only 1 dimm that got fried and no sign on the PCB of where the power came from.
i could clearly see the discoloration even before i removed the thermal pads.
now the big question is how did this happen, i don’t know but a friend of mine suggested that the heatspreader might have come in contact with the caps or vrms or whatever it is in the bottom of this picture
Dang that’s carnage!
Did those heatspreaders come with the RDIMMs stock?
I’d be a little concerned about having squishy thermal pads directly contacting some of the exposed component contacts on DDR5 RDIMMs since they could be over compressed and let a conductive headspreader short something.
i was using bitspower DDR5 watercooling kit, as i said no sign of any short anywhere else on the PCB so i am not sure its related to the heatspreaders, but i dont see any situation where the motherboard would just crank up the voltage for one stick just randomly.
it could be just a random cause of a bad component but it seems unlikely to me.
(for anyone who sees these images and did not see the one i posted above, there was a thermal pad on the burnt parts, i just didnt include them in the image)
ahhh the spreaders are anodized so they shouldn’t be able to short the components anyway.
You know how the squishy thermalpads ooze liquid when they are compressed? maybe that got somewhere it shouldn’t have?
I don’t even know if the liquid that comes out of them is conductive though.
you’re right they probably are anodized didn’t think about that. then the liquid might be the problem, the burnt parts are a bit taller so they might compress the pads a bit more. its just strange that it all happened after i replaced a graphics card.
a bit of searching suggest the liquid is non conductive oils
they’re seeping non conductive oils from inside them. It’s technically fine
Most of the “wet” spots you’re seeing is the silicone oil seeping out from the thermal pads and aren’t conductive.
i cant get my phone to focus when too close to the chip so its hard to make any decent photos, but to
me there seems to be both solder and thermal pad residues.
it was not intended to derail the thread but i guess that’s what happened, it was more a comment on all the trouble i’ve had with memory on the wrx90 platform and then when i finally had everything working something else had to break
My guess is you crushed one of these MLCC caps, they are made of ceramic and are fragile. A microscopic hairline crack can cause it to become a dead short.
And yes, that liquid is the thermal transfer silicon in the pads, nothing to worry about.
After reading all this I went back through how I put the heatspreaders on my RDIMMs and I never had any capacitors standing proud enough to contact the heatspreaders, there was a shallow pocket in the heatspreader to account for the capacitors:
My condolences. I just recently killed a motherboard, cpu and 4 RDIMMS from a leak in a gpu waterblock where the port in the acetal terminal block cracked after 2 1/2 years of no issue service. A costly $1700 failure. I got a faster cpu for the repair. It took a month for the gpu RMA to finally make it back and get the system back to full strength.
