No, there's no OCing results this time

Just a rant about AMD's cold-bug!!!
Let's start then...
Chapter One: Preparation.
1. Chilly1's DI/LN2 container that I got from Conrad to play with. Thanks a lot, bro
2. Besides seal string that I use on my phase change I decided to apply some Liquid Electrical Tape to the areas not covered with seal tape. Looks pretty messy but it's very effective
3. This is how I attached thermal probe to the bottom of the block...
4. Slowly getting ready to roll...
5. This what I saw while sitting in BIOS @ 3730Mhz 1.80vcore BIOS LOAD, meaning just normally loaded system, nothing extreme. Lowest I've seen was -68C while in BIOS... I did not takle a pic as I was kind of pissed then already
6. This is why it's great I decided to use Liguid electrical Tape

Very slight condensation due to alck of airflow in that area. I forgot to direct fan in there

After that I put 120mm fan in that area and problem was gone...
7. Quick cool shot of Acetone+DI boiling
8. Battlefielld after 3.5hrs of trying to get some decent fun with no success
9. Container without insulation - just a cool pic
...
and nothing else to show
...
*** My observations about cold-bug ***
1. On my San Diego, it kicks in past -44C on a block. Any lower and I cannot get to Windows - it hangs past BIOS BACK-UP.
2. I can sit in BIOS at -64C on the block and at least 3730Mhz on CPU. Once I leave it, what above happens.
3. Once I get lucky to get into Windows I need to maintain
constant LOAD - I used looped Sandra CPU Arithmetic Bench @ LOW Priority to do it. But even with that running, past -50C systems hangs
CONLUSION: On-die memory controller + 90nm process is the worst nightmare combination. I wish I was a bit more lucky and got some "healthy" chip
Anyone, feel free to post yur experiences with a DI/LN2 cold-bug and maybe any walkarounds or ideas
Thanks for looking and take care,
Michal