Ok so i had random lockups, esp when the computer was left to idle. When i ramped up memtest86 it found some errors that i could narrow down to just one stick. Pretty much the same adress range kept throwing errors even at a stock 2133Mhz on a Ryzen 2700x. I was like shit, this is todays hardware that usually doesnt last longer then 1.5 year on avg. I peeked on the heatspreader, and figured out it could be removed quite easily.
Since i baked motherboards, GPU's in the past with succes, why not try it with ram.... So i turned on the oven (pre-set 200 degrees) and i quickly made something to hold the stick into place. Once the desired temperature was there i inserted it, timed it for 5 minutes (1 minue to heat up) and after the timer went off shut it down, open the lid and let it cool down (do not ever move components that are in 200 degree mark).
So i assembled the thing back again, was quite easy with the heatspreader still intact includig it's glue type of strip, stick it back into my computer and fired it off the bat with 3133Mhz settings and a CL of 14 at 1.35v. Guess what ... The thing is passing every memtest!
I have quite tested that stick on forehand in various slots. It would repeatedly throw the same error at the same memory adress using different (lower) speed and loser timings. Sometimes by simply running a bit looser could solve it, but it seemed like a hardware based error or something. Not quite sure what was going on. I find it extremely remarkable that a 200 degree enviroment did a wonder for this 16GB stick. It works, flawless and without any issues now.
I know when memory errors exist its usually within the chip and thats something you just cant fix with baking. Perhaps the quality of the solder over time got worse. I have no idea. The computer always ran cool. So anyone dealing with the same as i had... Try a baking proces! ... Pics will follow.
Since i baked motherboards, GPU's in the past with succes, why not try it with ram.... So i turned on the oven (pre-set 200 degrees) and i quickly made something to hold the stick into place. Once the desired temperature was there i inserted it, timed it for 5 minutes (1 minue to heat up) and after the timer went off shut it down, open the lid and let it cool down (do not ever move components that are in 200 degree mark).
So i assembled the thing back again, was quite easy with the heatspreader still intact includig it's glue type of strip, stick it back into my computer and fired it off the bat with 3133Mhz settings and a CL of 14 at 1.35v. Guess what ... The thing is passing every memtest!
I have quite tested that stick on forehand in various slots. It would repeatedly throw the same error at the same memory adress using different (lower) speed and loser timings. Sometimes by simply running a bit looser could solve it, but it seemed like a hardware based error or something. Not quite sure what was going on. I find it extremely remarkable that a 200 degree enviroment did a wonder for this 16GB stick. It works, flawless and without any issues now.
I know when memory errors exist its usually within the chip and thats something you just cant fix with baking. Perhaps the quality of the solder over time got worse. I have no idea. The computer always ran cool. So anyone dealing with the same as i had... Try a baking proces! ... Pics will follow.