memristors?

[Jim Barrett]

Good Day Folks,
So, I've been reading about the new device discovered or invented or whatever that is being called a memristor or maybe memistor, depending on who's writing. All I can find are what sounds like press releases and opinions expressed as fact. It would seem to be a device whose resistance changes as a function of the rate of current passing through it.
One article I read indicated that it will be a new basic component added to the resistor, capacitor, and inductor.
After this the info gets, to me, kinda fuzzy. What seems like a lot of double talk...I know each word and term but it don't make sense to me the way they're put together.
Has anybody else been following this? I'd post a link but I can't find one that is really intelligible to me.
This device does appear to be capable of being non-volitile RAM.
I'm a very simple applied technician kinda guy. Can anyone here explain this for me in words of one syllable or less?

[Forrest Mims III]

Analog memory technology bears close watching, for it can revolutionize analog computer technology.

The new methods are much more sophisticated than my ultra primitive high school hybrid analog-digital language translator--but it worked. Briefly, a word was dialed letter-by-letter into a panel of six pots. This assigned each allowable word a unique resistance. The analog "memory" was a panel of 20 tiny trimmer resistors, each being set to the resistance of a dialed in word. After a word was dialed in, a modified electric music box stepped through the memory panel while the input and sample memory resistances were compared with a Wheatstone bridge. When there was a match, a meter needle pointed to zero where the needle blocked the light from a flashlight bulb from reaching a tiny solar cell. This switched off the music box motor and switched on 1-of-20 numbered lamps in an output box.

OK, pretty primitive. But that was back in 1962--and it worked. Might be fun to rebuild this by replacing the electric music box with a 1-of-n sequential decoder. It would certainly be much faster. But there are far more sophisticated applications that could be tackled with high-speed analog memory technology.

Forrest

Forrest M. Mims III
www.forrestmims.org
twitter.com/fmims

[Bob Scott]

It would seem to be a device whose resistance changes as a function of the rate of current passing through it. One article I read indicated that it will be a new basic component added to the resistor, capacitor, and inductor.

There already is a component that matches the requirements: a hot filament. The resistance of a light bulb filament changes with temperature. This characterisitic can be used and has been used by installing a small instrumentation bulb into the feedback of a sine wave oscillator. It automatically varies the gain of a sine wave oscillator to keep it's output at a steady level, without going into clipping or distortion.

This "new" IBM memory story keeps popping up every six months or so in one of various electronics forums like this one, and like an urban legend that won't die, it just keeps going around and around.

[MrAl]

Hi Bob,

Just a quick note...

The bulb filament, although used in the past for various types of gain controls, still has the innate problem that
it is relatively fragile, and subject to change over time. I think the MR, if done right, will get rid of these
problems once and for all.
When we will see these surface for every day use i have no idea, and stuff like this usually takes a long
time to hit market.
Cant wait to have a brand new circuit element to play with and model.

[Bigglez]

Has anybody else been following this? I'd post a link but I can't find one that is really intelligible to me.
This device does appear to be capable of being non-volitile RAM.

I found this link to be helpful.

[Bob Scott]

more here: http://www.electro-tech-online.com/elec ... emory.html

I like this electronic symbol of a memristor:

[Robert Reed]

I like this electronic symbol of a memristor:
Image


No Bob, when resistors sleep, they have dreams that only they are the "Ideal Resistor"

[MrAl]

Hello again,


Actually, this is the official symbol for the Memristor:



This one has a memristance of 47 memries

[haklesup]

Bob Scott's diagram originally came from the IEEE Spectrum Cover artical on May 08

http://www.spectrum.ieee.org/may08/6207

For the least hype ane best description I have seen so far, check it out.
This may have been tha artical that broke the story

Think of it as a section of wire whose resistance can be modified and that this change is both reversable and persistant when the power is turned off. Not bad for an accidental result from an experiment. Not since Pennicillin has there been such a possibility for revolution. (Pn was also an accidental discovery)

Digital circuits are made up of analog components. The memristor promises to replace a 2 or 4 transistor memory cell with just one component. Higher density and non-volitility make this a natural solution for memory if they can make repeatable and reliable devices compatible with todays predominant CMOS Fabrication technology.

[MrAl]

Hi again,

Hackle:
Yes, i envision a matrix similar to the old magnetic core memories only on a die.

I suspect the potential for these things is *greatly* underestimated. The possibilities for specialized analog
signal processing are almost endless ! This could cause a second revolution in electronics, no kidding.

For just one out of a billion or more incredible examples, here is one possible memristor in action...
At the moment i dont think i can even grasp how this could change things...




Ok, the unofficial symbol for the memristor again:

[Bob Scott]

[Bob Scott]

[Bob Scott]

Yes, i envision a matrix similar to the old magnetic core memories only on a die.

It looks like both the IEEE and you have an understanding of the capabilities of this "fourth quadrant" device. Is it a negative resistance? That's also an op-amp circuit, a reverse attenuator. Is it a conductance? That's just the reciprocal of a resistance. Please tell me what you think the requirements of the "4th quadrant part" are, and how a device that gains resistance when current flows, and loses resistance when current flows in the oposite direction, fulfills those requirements?

I suspect the potential for these things is *greatly* underestimated. The possibilities for specialized analog signal processing are almost endless ! This could cause a second revolution in electronics, no kidding.

Then someone will come up with the digital version in a DSP program in a 16FXX PIC and it'll be obsolete the next day. Come to think of it, if DSP programs can simulate anything analog, it would have already been done. Some digithead would be screaming "AND YOU CAN"T DO THIS WITH REGULAR ANALOG!!"

For just one out of a billion or more incredible examples, here is one possible memristor in action...
At the moment i dont think i can even grasp how this could change things...

That's my problem too: envisioning what the heck a "4th quadrant" device actually does. IEEE has made a big assumption here. I think maybe the article in the May issue may have just missed its intended April 1st issue publication date.



In this diagram, the magical memresistor has the capability of passing a fundamental sine wave frequency but it completely blocks the odd harmonics. How does it discriminate between a fundamental sine wave at any frequency and the odd hamonic sine waves? What does it do with a rectangle wave containing even harmonics? (including the DC offset)

Or are you just pulling our collective legs?

[haklesup]

I/m still trying to fathom its operation and possible applications myself so I'll hold onto that until it comes to focus but as to it revolutionizing electronics I think it will come on at more evolutionary speeds (and we know how fast electronics has evolved). First, don't expect leaded or SMT discrete memristors. They will likely only be available to IC designers working with certain processes (and certain licensed fabs). Second, their transfer function is already available from active multi component curcuits so it won't enable any new functions or applications all by itself. It will (should) however reduce size, complexity and power requirements for a number of existing applications allowing Moore's law to move forward a little more. Our fun with this component will likely be limited to simulation.

Unfortunately, we may not see its full potential until its patents and licence agreements expire. But I hope not.

[Jim Barrett]

Whew, you guys make me feel alot better already. It would appear that most everyone has confusion about this & that makes my own confusion more understandable. Excellent!
hakelsup is most likely correct, I understand this is nano technology & unlikely to be a discrete component.
The general consensus though appears to be that the possibilities for memory applications could be exciting. Booting my computer in seconds (or faster).