http://www.embedded.com/products/integratedcircuits/207403718?_requestid=485214


Mark

On May 1, 6:22 am, Mark <makol...@yahoo.com> wrote:
> http://www.embedded.com/products/integratedcircuits/207403718?_reques...
>
> Mark


It is a new type of diode.  Since a diode is any active part with two
legs and this part can't really be called passive I suggest that the
term diode be applied to it.

The article is full of the usual wild claims that nontechnical
reporters tend to put in articles about technical subjects, but
reading around those claims, it is an interesting and perhaps useful
device.

MooseFET wrote:

> On May 1, 6:22 am, Mark <makol...@yahoo.com> wrote:
> > http://www.embedded.com/products/integratedcircuits/207403718?_reques...
> >
> > Mark
>
> It is a new type of diode.  Since a diode is any active part with two
> legs and this part can't really be called passive I suggest that the
> term diode be applied to it.
>
> The article is full of the usual wild claims that nontechnical
> reporters tend to put in articles about technical subjects, but
> reading around those claims, it is an interesting and perhaps useful
> device.

Indeed. I could hardly believe what I was reading.

Graham

Well, it seems the guy writing that article doesn't have much of a EE 
background...

"Hysteresis has been explained away by current circuit theory as an 
anomaly..."  Really?  I kinda doubt the folks making ferroresonant 
transformers or mag amps or core memory considered hysteresis an "anomaly..."

Some of the discussion about passivity looks completely wrong too.  "changes 
in voltage, or flux" -- huh?  Maybe they mean electric field flux or 
something?

On Thu, 1 May 2008 06:22:01 -0700 (PDT), Mark <m...@yahoo.com>
wrote:

>
>
>http://www.embedded.com/products/integratedcircuits/207403718?_requestid=485214
>
>
Doesn't sound like a two-terminal device, in the intended application.
The change of state is delivered by a second current path.

RL

On Fri, 02 May 2008 08:39:03 -0400, legg <l...@nospam.magma.ca> wrote:

>On Thu, 1 May 2008 06:22:01 -0700 (PDT), Mark <m...@yahoo.com>
>wrote:
>
>>
>>
>>http://www.embedded.com/products/integratedcircuits/207403718?_requestid=485214
>>
>>
>Doesn't sound like a two-terminal device, in the intended application.
>The change of state is delivered by a second current path.

That is not how I read it.  Where is the second current path in a
crossbar switch?  There is only the upper and lower crossbar wires.
With the lower layer nonconductive the switch looks like a capacitor
with a high value resistor in parallel, when the lower layer is made
consuctive with suitable di/dt into the capacitor (programming pulse)
then it becomes a capacitor in parallel with a low resistance.  State
of the resistor then read with a pulse with too low a di/dt to change
the state.  Sounds like a two terminal device to me.  The low power
100 Gbit memory device they claim to have already produced sounds like
something I hope goes into production, it will be interesting to see
if it does.  Interesting article any way you look at it, thanks for
the post Mark.

Regards,
Glen

On Fri, 02 May 2008 14:17:45 GMT, Glen Walpert <n...@null.void>
wrote:

>On Fri, 02 May 2008 08:39:03 -0400, legg <l...@nospam.magma.ca> wrote:
>
>>On Thu, 1 May 2008 06:22:01 -0700 (PDT), Mark <m...@yahoo.com>
>>wrote:
>>
>>>
>>>
>>>http://www.embedded.com/products/integratedcircuits/207403718?_requestid=485214
>>>
>>>
>>Doesn't sound like a two-terminal device, in the intended application.
>>The change of state is delivered by a second current path.
>
>That is not how I read it.  Where is the second current path in a
>crossbar switch?  There is only the upper and lower crossbar wires.

Two wires = 4 terminals. Look at the picture.

http://www.physorg.com/news128786808.html

The changing resistance does not occur within the control wire, but
within/between the layers of titanium dioxide. I'm puzzled by the
apparent lack of reversibility in the process of creating and erasing
'oxygen vacancies', unless simple reversal of current direction in the
'control wire' suffices.

>With the lower layer nonconductive the switch looks like a capacitor
>with a high value resistor in parallel, when the lower layer is made
>consuctive with suitable di/dt into the capacitor (programming pulse)
>then it becomes a capacitor in parallel with a low resistance.  State
>of the resistor then read with a pulse with too low a di/dt to change
>the state.  Sounds like a two terminal device to me.  The low power
>100 Gbit memory device they claim to have already produced sounds like
>something I hope goes into production, it will be interesting to see
>if it does.  Interesting article any way you look at it, thanks for
>the post Mark.

A two terminal network requiring current flow to block would be a
one-shot fuse, to open. Getting a low Z branch to store charge is a
contradiction of terms. An infinitely low impedance could store
energy, in charge-flow-generated flux, but inductors are not used for
this purpose in any lengthy time frame for just that impractical
reason. 

A two terminal device requiring current flow or charge storage to
avalanche-conduct would also be a one-shot, to short.

A third terminal (minimum) is still required for a control, to remove
or introduce the charge.

No doubt the aim or the basic acheivment may involve a two wire
structure, but the present application does not. Two wire elements
with memory (like tunnel diodes and other dynamic elements that serve
the purpose) are tricky to actually use, without complex refresh.

"When Chua wrote his seminal paper, he used mathematics to deduce the
existence of a fourth circuit element type after resistors, capacitors
and inductors" 

A resistor isn't a circuit element - it's just a degree of conductor,
as 'opposed' to being a degree of insulator. There was never a
triangle that needed to be squared. These 'elements' already operate
at 90 and 180 degrees to each other. Chua was merely indicating that
possible new elements with memory 'memory' of being either one or the
other, should be considered. 

"a misconception that has pervaded electronic circuit theory(...)is
that the fundamental relationship in passive circuitry is between
voltage and charge. What the researchers contend is that the
fundamental relationship is actually between changes-in-voltage, or
flux, and charge." 

Thats voltage with stored charge as potential energy, in joules 
( C x V^2 / 2 ), 
and current with stored flux as potential energy, also in joules 
( L x i^2 /2 ).

It would be nice to see sensitive linear flux sensors coming out of
this. A switch is nice, and efficient switches are essential in power
conversion and logic, but a linear effect would show real mastery of
the phenomena.

RL