The dawn of the memristor...

Thread Starter

cmartinez

Joined Jan 17, 2007
8,218
"The overwhelming majority of people that are ejected from a vehicle in a crash do not survive," and someone "disproves" that statement by citing some instance in which someone did survive being ejected from a vehicle.
That happens to me a lot... so either I'm a bad communicator, or I hang around with bad listeners... anyway, we all have egos, and I try to always keep mine in check... it's always counterproductive to have a big ego... no... let me rephrase that... it's almost always counterproductive to have a big ego... ;)
 

WBahn

Joined Mar 31, 2012
29,976
Good video -- I passed along the link to some folks so that we can show it to our young cadets, many of whom are just at the age to start driving.
 

MachineHum

Joined Nov 3, 2014
74
Awesome thread... OP was interesting as well.

"According to the characterizing mathematical relations, the memristor's electrical resistance is not constant but depends on the history of current that had previously flowed through the device,i.e., its present resistance depends on how much electric charge has flowed in what direction through it in the past." - Wiki

If we're altering the resistance (analog) of memristors with a current (analog), a few things come to mind quickly, regarding the analog portion...

-1970's
-Temperature drift
-Data conversion, lots of it (D/A & A/D)
-Electrical noise messing everything up
-Settling time, how long do you need to run a current through to get desired resistance?

Another point, how do you measure resistance without running a current through a device? I'm sure there is a way but the only way I know is to run constant current (usually small) then measure the voltage... so unless you want to erase the memory every time you read from the device...

There was talk about "nanoamps" being able to effect the devices, this is where the "low power" argument comes from ... Last time I checked it's not that hard to induce "nanoamps" of current in a wire... So if we want our data to be secure, we're just going have to keep our HD's in a 3'' lead case.

The "LIKE A BRAIN" header on the news article was the stupidest thing I've read all day...

I'll watch it... as it's very interesting, and if you can keep a 64bit number in one of these devices and dodge all the stuff above... it has potential, but for now I'll stick with the transistor...
 

Thread Starter

cmartinez

Joined Jan 17, 2007
8,218
I'll watch it... as it's very interesting, and if you can keep a 64bit number in one of these devices and dodge all the stuff above... it has potential, but for now I'll stick with the transistor...
I'd settle for an 8 bit number... that would be far more than enough for it to be useful somehow.

Another point, how do you measure resistance without running a current through a device?
Maybe there's a minimum amount of current required for the device to store information?
 

WBahn

Joined Mar 31, 2012
29,976
Another point, how do you measure resistance without running a current through a device? I'm sure there is a way but the only way I know is to run constant current (usually small) then measure the voltage... so unless you want to erase the memory every time you read from the device...
How do you measure the voltage on a capacitor without changing the voltage of the capacitor? In practice, you generally can't. But you can read the voltage and not change the amount of the charge stored on the capacitor by very much. If you read it over and over and over then, at some point, you will have changed the charge stored enough to make a difference. But all you have to do is just periodically ask what the voltage is and then recharge the capacitor back up. That is basically how memory refresh on DRAM chips works today. So you could do the same thing with a memristor and use a small probe current to measure the resistance and, if it was getting close to the boundary where its state would be indeterminate, simply refresh it.
 
Top