Faulty DDRx Memory DIMM with zero hope for RMA? ...bake it! :)

RaptorSA

Executive Member
Joined
Sep 2, 2008
Messages
8,972
Reaction score
10,628
Hi Guys,

Super chuffed about how this worked out for me last night so I thought I'd just share it here.

Long story short: Had a 8GB (2x4GB) Mushkin Silverline DDR3 1333 RAM kit that was part of my Main PC build for a year or two. One of the DIMM's went bad about 1.5 years ago and I never felt like going through all the effort to RMA it so I moved on to a 1600MHz GSkill kit and kept the extra DIMM.

Anyway, so I built an mini-ITX i3 3220 based Media Center machine a few months ago and used the remaining 4GB stick in it, but recently I started doing more and more gaming on the i3 system and it was slowly starting to feel the pinch with that 4GB stick. I looked everywhere on the local Online shops and couldn't find a Mushkin Silverline 4Gb stick ANYWHERE and as we all know a replacement 8GB kit is expensive these days, R1000+... :erm:

So did some thinking last night and against all odds decided "phuuckit!", I'll wrap the faulty stick in some tinfoil and bake it, in case it's a hairline crack (dry joint) soldering issue. Some of you would obviously know about this technique most commonly used to fix GPU's.

Lo and behold!
10 minutes and +-210 Degrees (the soldering melting point of the most commonly used alloys are between 180 and 190 °C) later I let the stick cool and popped it in my PC for some testing.
It passed all windows memory testing and a few other RAM testing applications with flying colours. :D

Here's a link I came across for another dude that apparently did this and it worked:
http://www.maximumpc.com/forums/viewtopic.php?f=18&t=101495

I didn't expect mine to actually work, because unlike the above mentioned scenario, my DIMM wasn't completely dead or causing POST issues, in-depth testing just showed certain address spaces on the chip that couldn't be written and read from properly so my initial thought was that it was most likely the actual cells or a controller IC that went bad.
Apparently it wasn't.
 
Last edited:
in-depth testing just showed certain address spaces on the chip that couldn't be written and read from properly so my initial thought was that it was most likely the actual cells or a controller IC that went bad.
Apparently it wasn't.
Controller couldn't be wrong. There is no controller chip on DDR3. :)
 
Controller couldn't be wrong. There is no controller chip on DDR3. :)

I meant more along the lines of small IC's that's there to abstract some of the routing on the RAM chip and to keep some basic info about the frequency, latencies etc. supported.
 
I meant more along the lines of small IC's that's there to abstract some of the routing on the RAM chip and to keep some basic info about the frequency, latencies etc. supported.

Only problem, and not sure if its applicable at those lowish temperatures is the release of toxic substances into an oven you're later going to bake pies and cakes in. Stuff like mercury, lead, cadmium, cobalt could be released which can be dangerous substances.
 
Only problem, and not sure if its applicable at those lowish temperatures is the release of toxic substances into an oven you're later going to bake pies and cakes in. Stuff like mercury, lead, cadmium, cobalt could be released which can be dangerous substances.

It's ok. I don't think RaptorSA knows that he can use his oven for food
 
Top
Sign up to the MyBroadband newsletter
X