The Memristor’s Fundamental Secrets Revealed(spectrum.ieee.org)
spectrum.ieee.org
The Memristor’s Fundamental Secrets Revealed
http://spectrum.ieee.org/nanoclast/semiconductors/nanotechnology/the-memristors-fundamental-secrets-revealed
3 comments
There is a real physical device with a real application and that is called Resistive RAM (RRAM). Then there is a phenomenological mathematical model of a two terminal device.
In 2005 some guy from HP wrote a highly publicized paper where he tried to assign the model to the device and called it memristor.
There are a couple of issues with this:
1) The model is not physics based and is therefore not suitable to accurately describe the device operation.
2) The device existed even before the crude memristor model was attached to it. The memristor model did not help the application of RRAM in any way, it only publicized it in a way that made many people misunderstand its origins and application.
3) Research and development of RRAM is at an all-time high right now. Most semiconductor conferences have several sessions dedicated to this device, not the memristor. Actually there has been tremendous progress in RRAM. Toshiba/Sandisk made a 32Gbit prototype device and there are demonstrators of microcontrollers using RRAM. The only thing preventing market introduction right now is the momentum that flash has.
edit: HP is all talk. They have not shown anything in term of practical application yet. I don't understand how they are supposed to have a product any time.
In 2005 some guy from HP wrote a highly publicized paper where he tried to assign the model to the device and called it memristor.
There are a couple of issues with this:
1) The model is not physics based and is therefore not suitable to accurately describe the device operation.
2) The device existed even before the crude memristor model was attached to it. The memristor model did not help the application of RRAM in any way, it only publicized it in a way that made many people misunderstand its origins and application.
3) Research and development of RRAM is at an all-time high right now. Most semiconductor conferences have several sessions dedicated to this device, not the memristor. Actually there has been tremendous progress in RRAM. Toshiba/Sandisk made a 32Gbit prototype device and there are demonstrators of microcontrollers using RRAM. The only thing preventing market introduction right now is the momentum that flash has.
edit: HP is all talk. They have not shown anything in term of practical application yet. I don't understand how they are supposed to have a product any time.
You should help projects in kickstarter or do ANY R&D on a company or on your own to learn something:
In real life, the time it takes to do something compared with what you planned it is at least triple.
I had experience as the responsible person of a very talented R&D team for years. My parents were also researchers all their lives.
One of the amazing things is that in the end if you persist, you get it, but it is extremely easy to abandon as progress slows down(and it always happens with anything if it is new and nobody has done it before). The main problem with researchers is that they have the tendency to start new projects and finish none.
Steve Jobs learned this early from his mentor Robert Noyce.
In real life, the time it takes to do something compared with what you planned it is at least triple.
I had experience as the responsible person of a very talented R&D team for years. My parents were also researchers all their lives.
One of the amazing things is that in the end if you persist, you get it, but it is extremely easy to abandon as progress slows down(and it always happens with anything if it is new and nobody has done it before). The main problem with researchers is that they have the tendency to start new projects and finish none.
Steve Jobs learned this early from his mentor Robert Noyce.
The planning fallacy was named in 1979, so it's not exactly news that one's inside view estimate will be optimistic. I believe the poster you are are replying to is implicitly criticizing HP for publicly claiming commercialization earlier than was ever realistic.
The people at HP who are responsible for providing timeline estimates either should know about the planning fallacy and correct for it, or they are hopelessly incompetent planners in 2013. A more realistic explanation is that they knew they were being optimistic and lied about it for the short term press benefits that they realized when they made those statements. Companies lying about their development timelines is a valid object of complaint.
The people at HP who are responsible for providing timeline estimates either should know about the planning fallacy and correct for it, or they are hopelessly incompetent planners in 2013. A more realistic explanation is that they knew they were being optimistic and lied about it for the short term press benefits that they realized when they made those statements. Companies lying about their development timelines is a valid object of complaint.
Shouldn't you wait until, at least, 2014 before you mock them?
You don't go from zero to mass production in a year. Without signs of life it seems very unlikely the deadline will be met.
Hmmm, I left that article with the impression that the author was Dr. Gandhi was seeing what he wanted to see, not some fundamental new thing (or old thing).
When I read Dr. Chua's paper originally I also thought his reasoning for a third element was weak. In particular while it is well studied how the magnetic field and electric field interact, and essentially are analogs for the other, it is entirely unclear that voltage and current are in fact in a similar relationship. Now if he had proposed a device in which its permeability to magnetic flux varied based on the current passing through it? That I could see as a resistor analog in the magnetic domain.
I agree with kken who commented that Resistive RAM is a thing, but its more of a materials properties thing than a fundamental circuit element thing.
When I read Dr. Chua's paper originally I also thought his reasoning for a third element was weak. In particular while it is well studied how the magnetic field and electric field interact, and essentially are analogs for the other, it is entirely unclear that voltage and current are in fact in a similar relationship. Now if he had proposed a device in which its permeability to magnetic flux varied based on the current passing through it? That I could see as a resistor analog in the magnetic domain.
I agree with kken who commented that Resistive RAM is a thing, but its more of a materials properties thing than a fundamental circuit element thing.
I believe what you meant to say is that the reasoning for the fourth element is weak, as the memristor is claimed to be the fourth electrical engineering component discovered, after: resistor, capacitor and inductor. Also, as HP and others have observed and further, intentionally manifested the pinches hysteresis loop indicative of memristives, I'm curious to hear why you think that his science is off?
Populistic bullshit meant to grab attention, not science.
A coherer is not a metal-metal contact.
I am too lazy to elaborate, but some people in the comment section of the article already pointed out some of the more obvious issues.
A coherer is not a metal-metal contact.
I am too lazy to elaborate, but some people in the comment section of the article already pointed out some of the more obvious issues.
Wasn't it suppose to COMPLETELY revolutionise the way we do computing? Why haven't we heard more about it?