• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Standard DRAM is also Static Ram

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Fushyuguru

Member
Joined
Jul 19, 2002
Location
WPAFB, OH
I love school. Every once and a while, you hit a topic between all of the garbled equations and random banter that seems to hit home and make some sense when it comes to what youre ultimate intrests are.

The other day in my "Transport Phenomena in Semiconductor Devices" Class, we were talking about Carrier Densities and intrisic Fermi Levels in Energy Bands ect ect (dont ask). Our professor, Dr. Tewksbury, who had a hand in all that amazing research at Bell Labs back in the day went off on a tangent to help relate what were trying to understand to something we know.

He stated that Dynamic Memory (DRAM's and the like) can also be used as ultra long term static memory. Well how can this be? If you didnt know, Dynamic memory needs to refreshed roghly every 10ms to hold its information. It uses capacitance to store data which naturaly looses its charge due to basic Energy leakage. Remember nothing is perfect like in High School Physics (oh I loved those days).

Now you dont have to understand all of stuff (I do), but I will try and explain these equations at a slightly higher level:

If you take the equation for intrinsic carrier density for an n-type semiconductor,

ni = (Nc*Nv)^0.5 * exp[ - ( Eg/ (2k * T)) ],
and isolate the exponential : e^ -Eg/2kT where Eg = (E-ConductionBand - E-ValanceBand),

then you calculate this value at room temperature which is the normal operation temperature for most DRAM modules:

e^(-Eg/2kT) @ Room Temp : Eg~= 1 electron Volt, kT ~= 0.025 eV

Therefore ~= e^-40 = 4E-18 which gives you a good approximation of the leakage of this room temperature semiconductor (if it doesnt, just trust me way too much physics to explain every var and concept).

Now going from room temp 300k to lets say 77k (temperature of space) this value changes to an amplitude of 3E-70. Do a bit of calculus (which Im sorry but the forum wont let me type in these complex expressions, hehe) and we come out with the ultimate reason for the ppost which is that:

In Space @ 77degrees Kelvin, DRAM modules hold their data, unpowered, for ~10^20 years. Thats 10 with 20 zeros behind it!

Nasa and other space agencies have used dynamic memory as actual static ramdrives for years. By the time they lose their data if never refreshed, will last longer than probably this feeble planet.

Just thought Id share this fun little tidbit of info.
 
I guess the only other questions might be WHAT data is so important as to be stored by nasa, HOW MUCH are they storing, and WHERE. Oh and why they're spying on me, requiring me to wear this stylish tin-foil hat.

in all seriousness though, this is an interesting idea to possibly store some important data about what happened while this planet/species existed for some future generation/explorers,
somewhat like voyager did.. i guess the only limitation is
size/densities.. meh. anyways neat thread.

cheers
 
doesnt the components that make up ICs rip apart on the molecular under a certain temperature because different elements expand and contract at different rates - just something I have read somewhere.
 
ninthebin said:
doesnt the components that make up ICs rip apart on the molecular under a certain temperature because different elements expand and contract at different rates - just something I have read somewhere.

I wouldnt doubt it. But I believe that temperature would have to be incredibly low. With current devices being manufactured with component seperations of 4-10 atoms in width, that expansion and contraction is incredibly miniscule even with a delta of several hundred kelvin below solid state phase. Correct me if Im wrong but the expansion and contraction should be a linear relation when in solid phase. Im not a material engineer, but Im sure there is a point where that would happen, it just solely depends on whats mated with what. We have already seen what strained silicon can do cooled with LN2. Theres no ripping occuring there. I would love to find that article you read though, sounds neat. May have to do some googlage.
 
Back