View Full Version : How low can you go before problems occur?
Smalltimer
10-18-2002, 11:27 AM
I have an opportunity to obtain a compact cryo unit from a lab
at college they are updating.
I examined the unit and most of the space is used for
the initial refrigeration shelves etc.
under optimal conditions it can produce temps of
-150c +
so, I am wondering if it would be to low to use for a
extreme cooling project. Also I am looking at a complete
system submersion (excluding drives). with a direct
focus on the cpu and gpu units.
I only have 3 days to think this over then on monday
the unit is gone to another lab.
thanks in advance for you suggestions
-ST
|PuNiSh3R|
10-18-2002, 06:21 PM
It's not the problem of going to low.. it's the problem of coming back up from what I understand.. it's really just the rapid cooling and heating that will screw you up.. but let's say you kept the system on all the time.. or something and say.. had an average of -100C water.. your CPU would probably be something like -80C full load.. so that's not a problem at all.. insulate really good though.. the whole problem is with direct die systems.. cuz of course.. when you turn the system on u get really cold temps.. turn it off your temps come up fairly quick.. that is when you get rapid expansion of parts.. which causes death.. but you are doing water chilling right? So you won't have that problem.
Smalltimer
10-18-2002, 06:34 PM
Actually I have not decided on a definate approach
I dont know if there would be any way to maintain
viscosity at -80c and bellow there are other fluids
though.
Well if the only problem is material expansion then
I guess it worth getting it for a project. I am trying
to obtain documentation from the manufacturing
company, hopefully it will contain some valuable
technical data.
All i know atm is that this is a cascade system
it would be kinda kool to see anything -80c + on
a die :)
I will post more info here as this project unfolds
ronnie0738
10-21-2002, 10:09 AM
But if you kept it running 24/7, it shouldn't be a problem then right?
|PuNiSh3R|
10-21-2002, 10:22 AM
right
The better thing to look into before wasting your time is
can the cryo. unit handle the heat load of such a computer
system?
It's not something I have delved into but you may find that
even if it can get sub -100 temps thats with something
producing little or zero heat output.
I have seen large cryo systems but they only handle 15w of heat
but do get lovely cold temp but no way they will handle 60-80w
of heat.
So I'd check that aspect out first.
|PuNiSh3R|
10-27-2002, 11:09 AM
Well his computer in a water-chiller setup would generate well over 60-80w unless he was just doing the CPU and has like a P4 or something.. He'd be looking more like 100w.
aenigma
10-27-2002, 02:25 PM
Well what kind of cryogenic unit is it?If it is what I am thinking, it should be able to handle it easy.
Smalltimer
10-31-2002, 06:32 AM
I backed out of the deal.
I visited the unit with my uncle who is a refrigeration
and heating technician, and this was his first question
also, how much heat can it remove ...
the cryo unit is a dual stage cascade system that
can.... reach as low as -150c under ideal conditions
but content has a play in this.
the initial storage area is only 1 cubic foot aprox.
very small space designed to handle small sized
samples etc...
after some discussion I decided to look into a
custom phase change project instead, I have
no idea where to start though.
Any suggestions are greatly appreciated:
I also have a few questions I'm hoping some
of you might have some experience and insight
on.
1) Does a chiller system have any benefit over
a direct die evap system? ex heat absorbtion etc.
2) Will sub 0 temps only yeild a small fraction of
clock gains? only valuable for benching etc.
and finally, if I was to use liquid/ chiller combo
do I have to drain and clean that *(!@#&&*($
thing every month !n (ouch)
thanks all for your help on this matter
-ST
aenigma
10-31-2002, 11:02 PM
:eek:
It was a 2 stage cascade system?I built one of those, I didnt think they were made commercially.Strange...
Ok water chillers dont work as good as direct die, if your going to DIY, direct die is the best.Unless you want to cool your video nb and all that, then water chiller might be better/easier.
Sub zero temps do help in overclocking, but I am not exactly sure how much....